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 578, Tue Mar 14 05:16:29 2000 UTC revision 1047, Tue Jan 29 21:31:50 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
17    Date: 2002/01/29
18    Tag: blume-20020129-INSTALL
19    Description:
20    
21    - Added somewhat detailed installation instructions (file INSTALL).
22    - Fixed curl-detection bug in config/install.sh.
23    - It is now possible to select the URL getter using the URLGETTER
24      environment variable:
25    
26          not set / "unknown"      --> automatic detection (script tries wget,
27                                       curl, and lynx)
28          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
29                                       how to properly invoke them)
30          other                    --> use $URLGETTER directly, it must take
31                                       precisely two command-line arguments
32                                       (source URL and destination file name)
33    
34    ----------------------------------------------------------------------
35    Name: Matthias Blume
36    Date: 2002/01/28
37    Tag: blume-20020128-sparc-ccalls
38    Description:
39    
40    - Fixed problem with calculation of "used" registers in sparc-c-calls.
41    - Make use of the allocParam argument in sparc-c-calls.
42    
43    ----------------------------------------------------------------------
44    Name: Matthias Blume
45    Date: 2002/01/28
46    Tag: blume-20020128-allocParam
47    Description:
48    
49    John Reppy:  Changes c-calls API to accept client-callback for
50    allocating extra stack space.
51    me: Corresponding changes to mlriscGen (using a dummy argument that
52        does not change the current behavior).
53    
54    ----------------------------------------------------------------------
55    Name: Matthias Blume
56    Date: 2002/01/28 12:00:00
57    Tag: Release_110_38
58    Description:
59    
60    This time for real!!!
61    
62    ----------------------------------------------------------------------
63    Name: Matthias Blume
64    Date: 2002/01/28 10:56:00 EST
65    Tag: blume-20020128-retraction
66    Description:
67    
68    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
69       with blume-Release_110_38-retracted.)
70    
71    1. Fixed a problem with incorrect rounding modes in real64.sml.
72       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
73    
74    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
75       and enums fixed.  The naming of corresponding ML identifiers should
76       now be consistent again.
77    
78    ----------------------------------------------------------------------
79    Name: Allen Leung
80    Date: 2002/01/27
81    Tag: leunga-20020127-nowhere
82    Description:
83    
84       Added a target called nowhere in the configuration scripts.
85       Enabling this will build the MLRISC 'nowhere' tool (for translating
86       programs with where-clauses into legal SML code) during installation.
87    
88    ----------------------------------------------------------------------
89    Name: Matthias Blume
90    Date: 2002/01/25 21:27:00 EST
91    Tag: blume-Release_110_38-retracted
92    Description:
93    
94    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
95    
96    README will be added later.
97    
98    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
99    (Release_110_38) removed.  Reason: Last-minute bug fixes.
100    
101    ----------------------------------------------------------------------
102    Name: Matthias Blume
103    Date: 2002/01/25
104    Tag: blume-20020125-ffi
105    Description:
106    
107    A large number of tweaks and improvements to ml-nlffi-lib and
108    ml-nlffigen:
109    
110       - ML represenation types have been streamlined
111       - getter and setter functions work with concrete values, not abstract
112         ones where possible
113       - ml-nlffigen command line more flexible (see README file there)
114       - some bugs have been fixed (hopefully)
115    
116    ----------------------------------------------------------------------
117    Name: Lal George
118    Date: 2002/01/24
119    Tag: george-20020124-risc-ra-interface
120    Description:
121    
122       There is a dramatic simplification in the interface to the
123       register allocator for RISC architectures as a result of making
124       parallel copy instructions explicit.
125    
126    ----------------------------------------------------------------------
127    Name: Matthias Blume
128    Date: 2002/01/22
129    Tag: blume-20020122-x86-ccalls
130    Description:
131    
132    Bug fix for c-calls on x86 (having to do with how char- and
133    short-arguments are being handled).
134    
135    ----------------------------------------------------------------------
136    Name: Matthias Blume
137    Date: 2002/01/21
138    Tag: blume-20020121-ff
139    Description:
140    
141    Another day of fiddling with the FFI...
142    
143    1. Bug fix/workaround:  CKIT does not complain about negative array
144       dimensions, so ml-nlffigen has to guard itself against this possibility.
145       (Otherwise a negative dimension would send it into an infinite loop.)
146    
147    2. Some of the abstract types (light objects, light pointers, most "base"
148       types) in structure C are now eqtypes.
149    
150    3. Added constructors and test functions for NULL function pointers.
151    
152    ----------------------------------------------------------------------
153    Name: Matthias Blume
154    Date: 2002/01/18
155    Tag: blume-20020118-ready-for-new-release
156    Description:
157    
158    Made config/srcarchiveurl point to a new place.  (Will provide boot
159    files shortly.)
160    
161    Maybe we christen this to be 110.38?
162    
163    ----------------------------------------------------------------------
164    Name: Matthias Blume
165    Date: 2002/01/18
166    Tag: blume-20020118-more-ffifiddle
167    Description:
168    
169    Today's FFI fiddling:
170    
171      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
172        C.Set.* that use concrete (MLRep.*) arguments and results instead
173        of abstract ones.
174    
175      - Provided word-style bit operations etc. for "int" representation
176        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
177        and Long.
178    
179    ----------------------------------------------------------------------
180    Name: Matthias Blume
181    Date: 2002/01/18
182    Tag: blume-20020118-use-x86-fp
183    Description:
184    
185    Now that x86-fast-fp seems to be working, I turned it back on again
186    by default.  (Seems to work fine now, even with the FFI.)
187    
188    Other than that, I added some documentation about the FFI to
189    src/ml-nlffigen/README and updated the FFI test examples in
190    src/ml-nlffi-lib/Tests/*.
191    
192    ----------------------------------------------------------------------
193    Name: Allen Leung
194    Date: 2002/01/17
195    Tag: leunga-20020117-x86-fast-fp-call
196    Description:
197    
198       1. Fixed a problem with handling return fp values when x86's fast fp
199          mode is turned on.
200    
201       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
202          of %f32.
203    
204       3. Added a constructor INT32lit to the ast of MLRISC tools.
205    
206    ----------------------------------------------------------------------
207    Name: Matthias Blume
208    Date: 2002/01/16
209    Tag: blume-20020116-ffifiddle
210    Description:
211    
212    More fiddling with the FFI interface:
213    
214     - Make constness 'c instead of rw wherever possible.  This eliminates
215       the need for certain explicit coercions.  (However, due to ML's
216       value polymorphism, there will still be many cases where explicit
217       coercions are necessary.  Phantom types are not the whole answer
218       to modeling a subtyping relationship in ML.)
219    
220     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
221    
222     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
223       clumsy workaround for varargs problem.)
224    
225    ----------------------------------------------------------------------
226    Name: Lal George
227    Date: 2002/01/15
228    Tag: <none>
229    Description:
230    
231    1. Since COPY instructions are no longer native to the architecture,
232       a generic functor can be used to implement the expandCopies function.
233    
234    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
235       TEXT segment.
236    
237    ----------------------------------------------------------------------
238    Name: Matthias Blume
239    Date: 2002/01/15
240    Tag: blume-20020115-ffiupdates
241    Description:
242    
243    1. Fix for bug resulting in single-precision float values being returned
244       incorrectly from FFI calls.
245    
246    2. Small modifications to C FFI API:
247    
248        - memory-allocation routines return straight objects (no options)
249          and raise an exception in out-of-memory situations
250        - unsafe extensions to cast between function pointers and pointers
251          from/to ints
252        - added structure C_Debug as an alternative to structure C where
253          pointer-dereferencing (|*| and |*!) always check for null-pointers
254        - added open_lib' to DynLinkage;  open_lib' works like open_lib
255          but also takes a (possibly empty) list of existing library handles
256          that the current library depends on
257    
258    ----------------------------------------------------------------------
259    Name: Matthias Blume
260    Date: 2002/01/10
261    Tag: blume-20020110-newffigen
262    Description:
263    
264    1. Updates to portable graph code.
265    
266    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
267       more scalable now so that even huge interfaces such as the one
268       for GTK compile in finite time and space. :-)
269       See src/ml-nlffigen/README for details on what's new.
270    
271    ----------------------------------------------------------------------
272    Name: Lal George
273    Date: 2001/01/09 14:31:35 EST 2002
274    Tag: george-20011206-rm-native-copy
275    Description:
276    
277            Removed the native COPY and FCOPY instructions
278            from all the architectures and replaced it with the
279            explicit COPY instruction from the previous commit.
280    
281            It is now possible to simplify many of the optimizations
282            modules that manipulate copies. This has not been
283            done in this change.
284    
285    ----------------------------------------------------------------------
286    Name: Lal George
287    Date: 2001/12/06 16:50:13 EST 2001
288    Tag: george-20011206-mlrisc-instruction
289    Description:
290    
291    Changed the representation of instructions from being fully abstract
292    to being partially concrete. That is to say:
293    
294      from
295            type instruction
296    
297      to
298            type instr                              (* machine instruction *)
299    
300            datatype instruction =
301                LIVE of {regs: C.cellset, spilled: C.cellset}
302              | KILL of {regs: C.cellset, spilled: C.cellset}
303              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
304              | ANNOTATION of {i: instruction, a: Annotations.annotation}
305              | INSTR of instr
306    
307    This makes the handling of certain special instructions that appear on
308    all architectures easier and uniform.
309    
310    LIVE and KILL say that a list of registers are live or killed at the
311    program point where they appear. No spill code is generated when an
312    element of the 'regs' field is spilled, but the register is moved to
313    the 'spilled' (which is present, more for debugging than anything else).
314    
315    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
316    We used to generate:
317    
318            DEFFREG f1
319            f1 := f2 + f3
320            trapb
321    
322    but now generate:
323    
324            f1 := f2 + f3
325            trapb
326            LIVE {regs=[f1,f2,f3], spilled=[]}
327    
328    Furthermore, the DEFFREG (hack) required that all floating point instruction
329    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
330    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
331    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
332    this architecture).
333    
334    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
335    all the architectures. This will result in further simplification of the
336    register allocator that must be aware of them for coalescing purposes, and
337    will also simplify certain aspects of the machine description that provides
338    callbacks related to parallel copies.
339    
340    ANNOTATION should be obvious, and now INSTR represents the honest to God
341    machine instruction set!
342    
343    The <arch>/instructions/<arch>Instr.sml files define certain utility
344    functions for making porting easier -- essentially converting upper case
345    to lower case. All machine instructions (of type instr) are in upper case,
346    and the lower case form generates an MLRISC instruction. For example on
347    the alpha we have:
348    
349      datatype instr =
350         LDA of {r:cell, b:cell, d:operand}
351       | ...
352    
353      val lda : {r:cell, b:cell, d:operand} -> instruction
354        ...
355    
356    where lda is just (INSTR o LDA), etc.
357    
358    ----------------------------------------------------------------------
359    Name: Matthias Blume
360    Date: 2001/11/22 21:40:00 EST
361    Tag: Release_110_37
362    Description:
363    
364    Release 110.37.  This time for real.
365    
366    ----------------------------------------------------------------------
367    Name: Matthias Blume
368    Date: 2001/11/21 16:35:00 EST
369    Tag: blume-20011121-foot-in-mouth
370    Description:
371    
372    Removed the "Release_110_37" tag because of a serious bug.
373    This will be re-tagged once the bug is fixed.
374    
375    ----------------------------------------------------------------------
376    Name: Matthias Blume
377    Date: 2001/11/21 16:14:00 EST
378    Tag: blume-20011121-forgottenfile
379    Description:
380    
381    Forgot to add a file.  (Just a .tex-file -- part of
382    the CM manual source.)
383    
384    ----------------------------------------------------------------------
385    Name: Matthias Blume
386    Date: 2001/11/21 16:10:00 EST
387    Tag: blume-20011121-invalid_110_37
388    Description:
389    
390    Note: I removed the original tag "Release_110_37" from this commit
391          because we found a serious bug in all non-x86 backends.
392          - Matthias
393    
394    1. Modifications to the SML/NJ code generator and to the runtime system
395       so that code object name strings are directly inserted into code
396       objects at code generation time.  The only business the runtime system
397       has with this is now to read the name strings on occasions.
398       (The encoding of the name string has also changed somewhat.)
399    
400    2. CM now implements a simple "set calculus" for specifying export lists.
401       In particular, it is now possible to refer to the export lists of
402       other libraries/groups/sources and form unions as well as differences.
403       See the latest CM manual for details.
404    
405    3. An separate notion of "proxy" libraries has again be eliminated from
406       CM's model.  (Proxy libraries are now simply a special case of using
407       the export list calculus.)
408    
409    4. Some of the existing libraries now take advantage of the new set
410       calculus.
411       (Notice that not all libraries have been converted because some
412       of the existing .cm-files are supposed to be backward compatible
413       with 110.0.x.)
414    
415    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
416       "General.exnMessage"!  The former relies on a certain hook to be
417       initialized, and that often does not happen in the stand-alone case.)
418    
419    ----------------------------------------------------------------------
420    Name: Lal George
421    Date: 2001/11/21  13:56:18 EST
422    Tag: george-2001121-pseudo-ops
423    Description:
424    
425      Implemented a complete redesign of MLRISC pseudo-ops. Now there
426      ought to never be any question of incompatabilities with
427      pseudo-op syntax expected by host assemblers.
428    
429      For now, only modules supporting GAS syntax are implemented
430      but more should follow, such as MASM, and vendor assembler
431      syntax, e.g. IBM as, Sun as, etc.
432    
433    ----------------------------------------------------------------------
434    Name: Matthias Blume
435    Date: 2001/11/14 11:52:00 EST
436    Tag: blume-20011114-srcname
437    Description:
438    
439    1. Routed the name of the current source file to mlriscgen where it
440       should be directly emitted into the code object.  (This last part
441       is yet to be done.)
442    
443    2. Some cleanup of the pgraph code to make it match the proposal that
444       I put out the other day.  (The proposal notwithstanding, things are
445       still in flux here.)
446    
447    ----------------------------------------------------------------------
448    Name: Lal George
449    Date: 2001/11/14 09:44:04 EST
450    Tag:
451    Description:
452    
453      Fix for a backpatching bug reported by Allen.
454    
455      Because the boundary between short and long span-dependent
456      instructions is +/- 128, there are an astounding number of
457      span-dependent instructions whose size is over estimated.
458    
459      Allen came up with the idea of letting the size of span
460      dependent instructions be non-monotonic, for a maxIter
461      number of times, after which the size must be monotonically
462      increasing.
463    
464      This table shows the number of span-dependent instructions
465      whose size was over-estimated as a function of maxIter, for the
466      file Parse/parse/ml.grm.sml:
467    
468         maxIter            # of instructions:
469            10                      687
470            20                      438
471            30                      198
472            40                        0
473    
474      In compiling the compiler, there is no significant difference in
475      compilation speed between maxIter=10 and maxIter=40. Actually,
476      my measurements showed that maxIter=40 was a tad faster than
477      maxIter=10! Also 96% of the  files in the compiler reach a fix
478      point within 13 iterations, so fixing maxIter at 40, while high,
479      is okay.
480    
481    ----------------------------------------------------------------------
482    Name: Matthias Blume
483    Date: 2001/10/31 15:25:00 EST
484    Tag: blume-20011031-pgraph
485    Description:
486    
487    CKIT:
488    * Changed the "Function" constructor of type Ast.ctype to carry optional
489      argument identifiers.
490    * Changed the return type of TypeUtil.getFunction accordingly.
491    * Type equality ignores the argument names.
492    * TypeUtil.composite tries to preserve argument names but gives up quickly
493      if there is a mismatch.
494    
495    installation script:
496    * attempts to use "curl" if available (unless "wget" is available as well)
497    
498    CM:
499    * has an experimental implementation of "portable graphs" which I will
500      soon propose as an implementation-independent library format
501    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
502    
503    NLFFI-LIB:
504    * some cleanup (all cosmetic)
505    
506    NLFFIGEN:
507    * temporarily disabled the mechanism that suppresses ML output for
508      C definitions whose identifiers start with an underscore character
509    * generate val bindings for enum constants
510    * user can request that only one style (light or heavy) is being used;
511      default is to use both (command-line arguments: -heavy and -light)
512    * fixed bug in handling of function types involving incomplete pointers
513    * generate ML entry points that take record arguments (i.e., using
514      named arguments) for C functions that have a prototype with named
515      arguments
516      (see changes to CKIT)
517    
518    ----------------------------------------------------------------------
519    Name: Allen Leung
520    Date: 2001/10/27 20:34:00 EDT
521    Tag: leunga-20011027-x86-fast-fp-call
522    Description:
523    
524       Fixed the bug described in blume-20010920-slowfp.
525    
526       The fix involves
527          1. generating FCOPYs in FSTP in ia32-svid
528          2. marking a CALL with the appropriate annotation
529    
530    ----------------------------------------------------------------------
531    Name: Matthias Blume
532    Date: 2001/10/16 11:32:00 EDT
533    Tag: blume-20011016-netbsd
534    Description:
535    
536    Underscore patch from Chris Richards (fixing problem with compiling
537    runtime system under recent NetBSD).
538    
539    ----------------------------------------------------------------------
540    Name: Allen Leung
541    Date: 2001/10/12 17:18:32 EDT 2001
542    Tag: leung-20011012-x86-printflowgraph
543    Description:
544    
545    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
546    
547    ----------------------------------------------------------------------
548    Name: Lal George
549    Date: 2001/10/11 23:51:34 EDT
550    Tag: george-20011011-too-many-instrs
551    Description:
552    
553    The representation of a program point never expected to see more
554    than 65536 instructions in a basic block!
555    
556    ----------------------------------------------------------------------
557    Name: Lal George
558    Date: 2001/10/09 09:41:37 EDT
559    Tag: george-20011008-mlrisc-labels
560    Description:
561    
562    Changed the machine description files to support printing of
563    local and global labels in assembly code, based on host assembler
564    conventions.
565    
566    ----------------------------------------------------------------------
567    Name: Matthias Blume
568    Date: 2001/09/25 15:25:00 EDT
569    Tag: blume-20010925-exninfo
570    Description:
571    
572    I provided a non-hook implementation of exnName (at the toplevel) and
573    made the "dummy" implementation of exnMessage (at the toplevel) more
574    useful: if nothing gets "hooked in", then at least you are going to
575    see the exception name and a message indicating why you don't see more.
576    
577    [For the time being, programs that need exnMessage and want to use
578    ml-build should either use General.exnMessage (strongly recommended) or
579    refer to structure General at some other point so that CM sees a
580    static dependency.]
581    
582    [Similar remarks go for "print" and "use":  If you want to use their
583    functionality in stand-alone programs generated by ml-build, then use
584    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
585    
586    ----------------------------------------------------------------------
587    Name: Matthias Blume
588    Date: 2001/09/20 17:28:00 EDT
589    Tag: blume-20010920-slowfp
590    Description:
591    
592    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
593    
594    ----------------------------------------------------------------------
595    Name: Matthias Blume
596    Date: 2001/09/20 17:20:00 EDT
597    Tag: blume-20010920-canonicalpaths
598    Description:
599    
600    0. Updated the BOOT file (something that I forgot to do earlier).
601    
602    1. Small internal change to CM so that it avoids "/../" in filenames
603       as much as possible (but only where it is safe).
604    
605    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
606       that arguments that contain delimiters are passed through correctly.
607       This change also means that all "special" arguments of the form
608       @SMLxxx... must come first.
609    
610    3. Changed install script to put relative anchor names for tool commands
611       into pathconfig.
612    
613    ----------------------------------------------------------------------
614    Name: Matthias Blume
615    >>>>>>> 1.169
616    Date: 2001/09/18 15:35:00 EDT
617    Tag: blume-20010918-readme11036
618    Description:
619    
620    Added README files.
621    
622    ----------------------------------------------------------------------
623    Name: Matthias Blume
624    Date: 2001/09/18 11:45:00 EDT
625    Tag: Release_110_36 (retag)
626    Description:
627    
628    Fixed mistake in config/preloads. Retagged as 110.36.
629    
630    ----------------------------------------------------------------------
631    Name: Matthias Blume
632    Date: 2001/09/18 09:40:00 EDT
633    Tag: Release_110_36_orig (tag changed)
634    Description:
635    
636    New version (110.36).  New bootfiles.
637    
638    ----------------------------------------------------------------------
639    Name: Matthias Blume
640    Date: 2001/09/14 16:15:00 EDT
641    Tag: blume-20010914-x86fastfp
642    Description:
643    
644    John committed some changes that Allen made, in particular a (hopefully)
645    correctly working version of the x86-fp module.
646    
647    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
648    flag to "true".  Everything seems to compile to a fixpoint ok, and
649    "mandelbrot" speeds up by about 15%.
650    
651    ----------------------------------------------------------------------
652    Name: Matthias Blume
653    Date: 2001/09/13 11:20:00 EDT
654    Tag: blume-20010913-minimal
655    Description:
656    
657    1. Stefan Monnier's patch to fix a miscompilation problem that
658       was brought to light by John Reppy's work on Moby.
659    
660    2. Implemented a minimal "structure Compiler" that contains just
661       "version" and "architecture".  The minimal version will be
662       available when the full version is not.  This is for backward-
663       compatibility with code that wants to test Compiler.version.
664    
665    ----------------------------------------------------------------------
666    Name: Matthias Blume
667    Date: 2001/08/28 14:03:00 EDT
668    Tag: blume-20010828-ml-lex
669    Description:
670    
671    Fix for bug 1581, received from Neophytos Michael.
672    
673    ----------------------------------------------------------------------
674    Name: Matthias Blume
675    Date: 2001/08/27 11:20:00 EDT
676    Tag: blume-20010827-readme11035
677    Description:
678    
679    Fleshed out the README file for 110.35.
680    
681    ----------------------------------------------------------------------
682    Name: Matthias Blume
683    Date: 2001/08/24 17:10:00 EDT
684    Tag: Release_110_35
685    Description:
686    
687    New version number (110.35).  New bootfiles.
688    
689    ----------------------------------------------------------------------
690    Name: Lal George
691    Date: 2001/08/24 13:47:18 EDT 2001
692    Tag: george-20010824-MLRISC-graphs
693    Description:
694    
695     removed clusters from MLRISC completely and replaced with graphs.
696    
697    ----------------------------------------------------------------------
698    Name: Matthias Blume
699    Date: 2001/08/23 17:50:00 EDT
700    Tag: blume-20010823-toplevel
701    Description:
702    
703    - some reorganization of the code that implements various kinds of
704      environments in the compiler (static, dynamic, symbolic, combined)
705    - re-implemented the EnvRef module so that evalStream works properly
706      (if the stream contains references to "use", "CM.make", etc.)
707    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
708    
709    ----------------------------------------------------------------------
710    Name: Matthias Blume
711    Date: 2001/08/20 15:50 EDT
712    Tag: blume20010820-slipup
713    Description:
714    
715    I forgot to commit a few files.  Here they are...
716    
717    ----------------------------------------------------------------------
718    Name: Matthias Blume
719    Date: 2001/08/20 15:35:00 EDT
720    Tag: blume-20010820-debugprof
721    Description:
722    
723    !!!! NEW BOOTFILES !!!!
724    
725    This is another round of reorganizing the compiler sources.  This
726    time the main goal was to factor out all the "instrumentation"
727    passes (for profiling and backtracing) into their own library.
728    The difficulty was to do it in such a way that it does not depend
729    on elaborate.cm but only on elabdata.cm.
730    
731    Therefore there have been further changes to both elaborate.cm and
732    elabdata.cm -- more "generic" things have been moved from the former
733    to the latter.  As a result, I was forced to split the assignment
734    of numbers indicating "primtyc"s into two portions: SML-generic and
735    SML/NJ-specific.  Since it would have been awkward to maintain,
736    I bit the bullet and actually _changed_ the mapping between these
737    numbers and primtycs.  The bottom line of this is that you need
738    a new set of bin- and bootfiles.
739    
740    I have built new bootfiles for all architectures, so doing a fresh
741    checkout and config/install.sh should be all you need.
742    
743    The newly created library's name is
744    
745        $smlnj/viscomp/debugprof.cm
746    
747    and its sources live under
748    
749        src/compiler/DebugProf
750    
751    ----------------------------------------------------------------------
752    Name: Matthias Blume
753    Date: 2001/08/15 17:15:00 EDT
754    Tag: blume-20010815-compreorg
755    Description:
756    
757    This is a first cut at reorganizing the CM libraries that make up the
758    core of the compiler.  The idea is to separate out pieces that could
759    be used independently by tools, e.g., the parser, the typechecker, etc.
760    
761    The current status is a step in this direction, but it is not quite
762    satisfactory yet.  Expect more changes in the future.
763    
764    Here is the current (new) organization...
765    
766        What used to be $smlnj/viscomp/core.cm is now divided into
767        six CM libraries:
768    
769             $smlnj/viscomp/basics.cm
770                           /parser.cm
771                           /elabdata.cm
772                           /elaborate.cm
773                           /execute.cm
774                           /core.cm
775    
776        The CM files for these libraries live under src/system/smlnj/viscomp.
777        All these libraries are proxy libraries that contain precisely
778        one CM library component.  Here are the locations of the components
779        (all within the src/compiler tree):
780    
781             Basics/basics.cm
782             Parse/parser.cm
783             ElabData/elabdata.cm
784             Elaborator/elaborate.cm
785             Execution/execute.cm
786             core.cm
787    
788         [This organization is the same that has been used already
789         for a while for the architecture-specific parts of the visible
790         compiler and for the old version of core.cm.]
791    
792         As you will notice, many source files have been moved from their
793         respective original locations to a new home in one of the above
794         subtrees.
795    
796         The division of labor between the new libraries is the following:
797    
798             basics.cm:
799                - Simple, basic definitions that pertain to many (or all) of
800                  the other libraries.
801             parser.cm:
802                - The SML parser, producing output of type Ast.dec.
803                - The type family for Ast is also defined and exported here.
804             elabdata.cm:
805                - The datatypes that describe input and output of the elaborator.
806                  This includes types, absyn, and static environments.
807             elaborator.cm:
808                - The SML/NJ type checker and elaborator.
809                  This maps an Ast.dec (with a given static environment) to
810                  an Absyn.dec (with a new static environment).
811                - This libraries implements certain modules that used to be
812                  structures as functors (to remove dependencies on FLINT).
813             execute.cm:
814                - Everything having to do with executing binary code objects.
815                - Dynamic environments.
816             core.cm:
817                - SML/NJ-specific instantiations of the elaborator and MLRISC.
818                - Top-level modules.
819                - FLINT (this should eventually become its own library)
820    
821    Notes:
822    
823    I am not 100% happy with the way I separated the elaborator (and its
824    data structures) from FLINT.  Two instances of the same problem:
825    
826        1. Data structures contain certain fields that carry FLINT-specific
827           information.  I hacked around this using exn and the property list
828           module from smlnj-lib.  But the fact that there are middle-end
829           specific fields around at all is a bit annoying.
830    
831        2. The elaborator calculates certain FLINT-related information.  I tried
832           to make this as abstract as I could using functorization, but, again,
833           the fact that the elaborator has to perform calculations on behalf
834           of the middle-end at all is not nice.
835    
836        3. Having to used exn and property lists is unfortunate because it
837           weakens type checking.  The other alternative (parameterizing
838           nearly *everything*) is not appealing, though.
839    
840    I removed the "rebinding =" warning hack because due to the new organization
841    it was awkward to maintain it.  As a result, the compiler now issues some of
842    these warnings when compiling init.cmi during bootstrap compilation. On
843    the plus side, you also get a warning when you do, for example:
844       val op = = Int32.+
845    which was not the case up to now.
846    
847    I placed "assign" and "deref" into the _Core structure so that the
848    code that deals with the "lazy" keyword can find them there.  This
849    removes the need for having access to the primitive environment
850    during elaboration.
851    
852    ----------------------------------------------------------------------
853    Name: Matthias Blume
854    Date: 2001/08/13
855    Tag: blume-20010813-closures
856    Description:
857    
858    This fix was sent to us by Zhong Shao.  It is supposed to improve the
859    performance of certain loops by avoiding needless closure allocation.
860    
861    ----------------------------------------------------------------------
862    Name: Lal George
863    Date: 2001/07/31 10:03:23 EDT 2001
864    Tag: george-20010731-x86-fmalloc
865    Description: Fixed bug in x86 calls
866    
867        There was a bug where call instructions would mysteriously
868        vanish. The call instruction had to be one that returned
869        a floating point value.
870    
871    ----------------------------------------------------------------------
872    Name: Lal George
873    Date: 2001/07/19 16:36:29 EDT 2001
874    Tag: george-20010719-simple-cells
875    Description:
876    
877    I have dramatically simplified the interface for CELLS in MLRISC.
878    
879    In summary, the cells interface is broken up into three parts:
880    
881      1. CellsBasis : CELLS_BASIS
882    
883            CellsBasis is a top level structure and common for all
884            architectures.  it contains the definitions of basic datatypes
885            and utility  functions over these types.
886    
887      2. functor Cells() : CELLS
888    
889            Cells generates an interface for CELLS that incorporates the
890            specific resources on the target architecture, such as the
891            presence of special register classes, their number and size,
892            and various useful substructures.
893    
894      3. <ARCH>CELLS
895    
896            e.g. SparcCells: SPARCCELLS
897    
898            <ARCH>CELLS usually contains additional bindings for special
899            registers  on the architecture, such as:
900    
901                    val r0 : cell           (* register zero *)
902                    val y : cell            (* Y register *)
903                    val psr : cell          (* processor status register *)
904                    ...
905    
906            The structure returned by applying the Cells functor is opened
907            in this interface.
908    
909    The main implication of all this is that the datatypes for cells is
910    split between CellsBasis and CELLS -- a fairly simple change for user
911    code.
912    
913    In the old scheme the CELLS interface had a definitional binding of
914    the form:
915    
916            signature CELLS = sig
917    
918               structure CellsBasis = CellsBasis
919    
920               ...
921    
922            end
923    
924    With all the sharing constraints that goes on in MLRISC, this old
925    design  quickly leads to errors such as:
926    
927            "structure definition spec inside of sharing ... "
928    
929    
930    and appears to require an unacceptable amount of sharing and where
931    constraint hackery.
932    
933    I think this error message (the interaction of definitional specs and
934    sharing) requires more explanation on our web page.
935    
936    ----------------------------------------------------------------------
937    Name: Matthias Blume
938    Date: 2001/07/19 15:00:00 EDT
939    Tag: blume-20010719-libreorg
940    Description:
941    
942    This update puts together a fairly extensive but straightforward change
943    to the way the libraries that implement the interactive system are
944    organized:
945    
946       The biggest change is the elimination of structure Compiler.  As a
947       replacement for this structure, there is now a CM library
948       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
949       that exports all the substructures of the original structure Compiler
950       directly.  So instead of saying Compiler.Foo.bar one now simply
951       says Foo.bar.  (The CM libraries actually export a collection of
952       structures that is richer than the collection of substructures of
953       structure Compiler.)
954    
955       To make the transition smooth, there is a separate library called
956       $smlnj/compiler/compiler.cm which puts together and exports the
957       original structure Compiler (or at least something very close to it).
958    
959       There are five members of the original structure Compiler
960       that are not exported directly but which instead became members
961       of a new structure Backend (described by signature BACKEND).  These are:
962       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
963       Interact (: INTERACT), structure Machine (: MACHINE), and val
964       architecture (: string).
965    
966       Structure Compiler.Version has become structure CompilerVersion.
967    
968       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
969       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
970       or x86, respectively.
971       Each of these exports the same frontend structures that
972       $smlnj/compiler.cm exports.  But they do not have a structure Backend
973       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
974       Hppa, PPC, Sparc, or X86, respectively.
975    
976       Library $smlnj/compiler/all.cm exports the union of the exports of
977       $smlnj/compiler/<arch>.cm
978    
979       There are no structures <Arch>Compiler anymore, use
980       $smlnj/compiler/<arch>.cm instead.
981    
982       Library host-compiler-0.cm is gone.  Instead, the internal library
983       that instantiates CM is now called cm0.cm.  Selection of the host
984       compiler (backend) is no longer done here but. (Responsibility for it
985       now lies with $smlnj/compiler/current.cm.  This seems to be more
986       logical.)
987    
988       Many individual files have been moved or renamed.  Some files have
989       been split into multiple files, and some "dead" files have been deleted.
990    
991    Aside from these changes to library organization, there are also changes
992    to the way the code itself is organized:
993    
994       Structure Binfile has been re-implemented in such a way that it no
995       longer needs any knowledge of the compiler.  It exclusively deals
996       with the details of binfile layout.  It no longer invokes the
997       compiler (for the purpose of creating new prospective binfile
998       content), and it no longer has any knowledge of how to interpret
999       pickles.
1000    
1001       Structure Compile (: COMPILE) has been stripped down to the bare
1002       essentials of compilation.  It no longer deals with linking/execution.
1003       The interface has been cleaned up considerably.
1004    
1005       Utility routines for dealing with linking and execution have been
1006       moved into their own substructures.
1007    
1008       (The ultimate goal of these changes is to provide a light-weight
1009       binfile loader/linker (at least for, e.g., stable libraries) that
1010       does not require CM or the compiler to be present.)
1011    
1012    CM documentation has been updated to reflect the changes to library
1013    organization.
1014    
1015    ----------------------------------------------------------------------
1016    Name: Matthias Blume
1017    Date: 2001/07/10 17:30:00 EDT
1018    Tag: Release_110_34
1019    Description:
1020    
1021    Minor tweak to 110.34 (re-tagged):
1022    
1023      - README.html file added to CVS repository
1024      - runtime compiles properly under FreeBSD 3.X and 4.X
1025    
1026    ----------------------------------------------------------------------
1027    Name: Matthias Blume
1028    Date: 2001/07/10 17:30:00 EDT
1029    Tag: Release_110_34
1030    Description:
1031    
1032    New version number (110.34). New bootfiles.
1033    
1034    ----------------------------------------------------------------------
1035    Name: Matthias Blume
1036    Date: 2001/07/09 16:00:00 EDT
1037    Tag: blume-20010709-more-varargs
1038    Description:
1039    
1040    I changed the handling of varargs in ml-nlffigen again:
1041    The ellipsis ... will now simply be ignored (with an accompanying warning).
1042    
1043    The immediate effect is that you can actually call a varargs function
1044    from ML -- but you can't actually supply any arguments beyond the ones
1045    specified explicitly.  (For example, you can call printf with its format
1046    string, but you cannot pass additional arguments.)
1047    
1048    This behavior is only marginally more useful than the one before, but
1049    it has the advantage that a function or, more importantly, a function
1050    type never gets dropped on the floor, thus avoiding follow-up problems with
1051    other types that refer to the offending one.
1052    
1053    ----------------------------------------------------------------------
1054    Name: Matthias Blume
1055    Date: 2001/07/09 11:25:00 EDT
1056    Tag: blume-20010709-varargs
1057    Description:
1058    
1059    1. ckit-lib.cm now exports structure Error
1060    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1061       with a warning accompanied by a source location.  Moreover, it
1062       merely skips the offending function or type and proceeds with the
1063       rest of its work.u  As a result, one can safely feed C code containing
1064       "..." to ml-nlffigen.
1065    3. There are some internal improvements to CM, providing slightly
1066       more general string substitutions in the tools subsystem.
1067    
1068    ----------------------------------------------------------------------
1069    Name: Matthias Blume
1070    Date: 2001/06/27 15:10:00 EDT
1071    Tag: blume-20010627-concur
1072    Description:
1073    
1074    Fixed a small bug in CM's handling of parallel compilation.
1075    (You could observe the bug by Control-C-interrupting an ordinary
1076    CMB.make or CM.stabilize and then attaching some compile servers.
1077    The result was that all of a sudden the previously interrupted
1078    compilation would continue on its own.  This was because of
1079    an over-optimization: CM did not bother to clean out certain queues
1080    when no servers were attached "anyway", resulting in the contents
1081    of these queues to grab control when new servers did get attached.)
1082    
1083    There is also another minor update to the CM manual.
1084    
1085    ----------------------------------------------------------------------
1086    Name: Matthias Blume
1087    Date: 2001/06/26 16:15:00 EDT
1088    Tag: blume-20010626-cmdoc
1089    Description:
1090    
1091    Minor typo fixed in CM manual (syntax diagram for libraries).
1092    
1093    ----------------------------------------------------------------------
1094    Name: Matthias Blume
1095    Date: 2001/06/25 22:55:00 EDT
1096    Tag: blume-20010625-x86pc
1097    Description:
1098    
1099    Fixed a nasty bug in the X86 assembly code that caused signal
1100    handlers to fail (crash) randomly.
1101    
1102    ----------------------------------------------------------------------
1103    Name: Matthias Blume
1104    Date: 2001/06/25 12:05:00 EDT
1105    Tag: blume-20010625-nlffigen
1106    Description:
1107    
1108    This update fixes a number of minor bugs in ml-nlffigen as reported by
1109    Nick Carter <nbc@andrew.cmu.edu>.
1110    
1111      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
1112      2. Default names for generated files are now derived from the name of
1113         the C file *without its directory*.  In particular, this causes generated
1114         files to be placed locally even if the C file is in some system directory.
1115      3. Default names for generated signatures and structures are also derived
1116         from the C file name without its directory.  This avoids silly things
1117         like "structure GL/GL".
1118         (Other silly names are still possible because ml-nlffigen does not do
1119          a thorough check of whether generated names are legal ML identifiers.
1120          When in doubt, use command line arguments to force particular names.)
1121    
1122    ----------------------------------------------------------------------
1123    Name: Matthias Blume
1124    Date: 2001/06/21 12:25:00 EDT
1125    Tag: blume-20010621-eXene
1126    Description:
1127    
1128    eXene now compiles and (sort of) works again.
1129    
1130    The library name (for version > 110.33) is $/eXene.cm.
1131    
1132    I also added an new example in src/eXene/examples/nbody.  See the
1133    README file there for details.
1134    
1135    ----------------------------------------------------------------------
1136    Name: Matthias Blume
1137    Date: 2001/06/20 16:40:00 EDT
1138    Tag: blume-20010620-cml
1139    Description:
1140    
1141    CML now compiles and works again.
1142    
1143    Libraries (for version > 110.33):
1144    
1145      $cml/cml.cm            Main CML library.
1146      $cml/basis.cm          CML's version of $/basis.cm.
1147      $cml/cml-internal.cm   Internal helper library.
1148      $cml/core-cml.cm       Internal helper library.
1149      $cml-lib/trace-cml.cm  Tracing facility.
1150      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
1151    
1152    The installer (config/install.sh) has been taught how to properly
1153    install this stuff.
1154    
1155    ----------------------------------------------------------------------
1156    Name: Matthias Blume
1157    Date: 2001/06/19 17:55:00 EDT
1158    Tag: blume-20010619-instantiate
1159    Description:
1160    
1161    This un-breaks the fix for bug 1432.
1162    (The bug was originally fixed in 110.9 but I broke it again some
1163    time after that.)
1164    
1165    ----------------------------------------------------------------------
1166    Name: Matthias Blume
1167    Date: 2001/06/19 17:25:00 EDT
1168    Tag: blume-20010619-signals
1169    Description:
1170    
1171    This should (hopefully) fix the long-standing signal handling bug.
1172    (The runtime system was constructing a continuation record with an
1173    incorrect descriptor which would cause the GC to drop data on the floor...)
1174    
1175    ----------------------------------------------------------------------
1176    Name: Matthias Blume
1177    Date: 2001/06/15 15:05:00 EDT
1178    Tag: blume-20010615-moresparc
1179    Description:
1180    
1181    Here is a short late-hour update related to Sparc c-calls:
1182    
1183     -- made handling of double-word arguments a bit smarter
1184    
1185     -- instruction selection phase tries to collapse certain clumsily
1186        constructed ML-Trees; typical example:
1187    
1188            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
1189    
1190        This currently has no further impact on SML/NJ since mlriscGen does
1191        not seem to generate such patterns in the first place, and c-calls
1192        (which did generate them in the beginning) has meanwhile been fixed
1193        so as to avoid them as well.
1194    
1195    ----------------------------------------------------------------------
1196    Name: Matthias Blume
1197    Date: 2001/06/15 15:05:00 EDT
1198    Tag: blume-20010615-sparc
1199    Description:
1200    
1201    The purpose of this update is to provide an implementation of NLFFI
1202    on Sparc machines.
1203    
1204    Here are the changes in detail:
1205    
1206       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
1207       the Sparc implementation of the c-calls API.
1208       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
1209       for accessing the ML frame.  Thus, we have a real frame pointer and
1210       can freely modify %sp without need for an omit-frame-ptr phase.
1211       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
1212       this case.
1213       * ml-nlffigen has been taught to produce code for different architectures
1214       and calling conventions.
1215       * In a way similar to what was done in the x86 case, the Sparc
1216       backend uses its own specific extension to mltree.  (For example,
1217       it needs to be able to generate UNIMP instructions which are part
1218       of the calling convention.)
1219       * ml-nlffi-lib was reorganized to make it more modular (in particular,
1220       to make it easier to plug in new machine- and os-dependent parts).
1221    
1222    There are some other fairly unrelated bug fixes and cleanups as well:
1223    
1224       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
1225       that they properly share their libraries with existing SML/NJ libraries.
1226       * I fixed a minor cosmetic bug in CM, supressing certain spurious
1227       follow-up error messages.
1228       * Updates to CM/CMB documentation.
1229    
1230    TODO items:
1231    
1232       * MLRISC should use a different register as its asmTemp on the Sparc.
1233         (The current %o2 is a really bad choice because it is part of the
1234         calling conventions, so things might interfere in unexpected ways.)
1235    
1236    ----------------------------------------------------------------------
1237    Name: Matthias Blume
1238    Date: 2001/06/07
1239    Tag: blume-20010607-calls
1240    Description:
1241    
1242    A number of internal changes related to C calls and calling conventions:
1243    
1244    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
1245       number of bytes popped implicitly (by the callee).  In most cases
1246       this field is 0 but on x86/win32 it is some non-zero value.  This
1247       is information provided for the benefit of the "omit-frameptr" pass.
1248    2. The CALL instruction on the x86 carries a similar "pops" field.
1249       The instruction selection phase copies its value from the ML-Tree
1250       CALL statement.
1251    3. On all other architectures, the instruction selection phase checks
1252       whether "pops=0" and complains if not.
1253    4. The c-calls implementation for x86 now accepts two calling conventions:
1254       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
1255       up after the call and pops is set to 0.  For "stdcall", the caller
1256       does nothing, leaving the cleanup to the callee; pops is set to
1257       the number of bytes that were pushed onto the stack.
1258    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
1259       distinguish between "ccall" and "stdcall".
1260    6. The UNIMP instruction has been added to the supported Sparc instruction
1261       set. (This is needed for implementing the official C calling convention
1262       on this architecture.)
1263    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
1264       work with the latest CM.
1265    
1266    ----------------------------------------------------------------------
1267    Name: Matthias Blume
1268    Date: 2001/06/05 15:10:00 EDT
1269    Tag: blume-20010605-cm-index
1270    Description:
1271    
1272    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
1273    
1274    1. CM can now generate "index files".  These are human-readable files
1275       that list on a per-.cm-file basis each toplevel symbol defined or
1276       imported.  The location of the index file for
1277       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
1278       To enable index-file generation, set CM.Control.generate_index to true
1279       or export an environment-symbol: export CM_GENERATE_INDEX=true.
1280    
1281       The CM manual has been updated accordingly.
1282    
1283    2. I made some slight modifications to the c-calls API in MLRISC.
1284    
1285         a) There is now a callback to support saving/restoring of
1286            dedicated but caller-save registers around the actual call
1287            instruction.
1288         b) One can optionally specify a comment-annotation for the
1289            call instruction.
1290    
1291    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
1292       (For example, the comment annotation shows the C prototype of
1293        the function being called.)
1294    
1295    ----------------------------------------------------------------------
1296    Name: Matthias Blume
1297    Date: 2001/06/01 13:30:00 EDT
1298    Tag: blume-20010601-nlffi-cleanup
1299    Description:
1300    
1301    This is mostly a cleanup of MLFFI stuff:
1302    
1303       - some signature files have been put into a more exposed place
1304       - the ugly 'f type parameter is gone (simplifies types tremendously!)
1305       - ml-nlffigen changed accordingly
1306       - tutorial updated
1307    
1308    Other changes:
1309    
1310       - author's affiliation in CM manual(s) updated
1311       - some more recognized keywords added to Allen's sml.sty
1312    
1313    ----------------------------------------------------------------------
1314    Name: Matthias Blume
1315    Date: 2001/05/25 15:30:00 EDT
1316    Tag: blume-20010525-iptr
1317    Description:
1318    
1319      - put the official 110.33-README (as it appears on the ftp server) under
1320        CVS
1321      - fixed a small bug related to incomplete pointer types in
1322        ml-nlffigen
1323      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
1324        (it does not need the 'f type parameter)
1325    
1326    ----------------------------------------------------------------------
1327    Name: Matthias Blume
1328    Date: 2001/05/23 14:30:00 EDT
1329    Tag: Release_110_33
1330    Description:
1331    
1332    New version number (110.33).  New bootfiles.
1333    
1334    ----------------------------------------------------------------------
1335    Name: Matthias Blume
1336    Date: 2001/05/22 18:06:00 EDT
1337    Tag: blume-20010522-targets
1338    Description:
1339    
1340    Made install.sh use file config/targets.customized if it exists, falling
1341    back to config/targets if it doesn't.  This way one can have a customized
1342    version of the targets file without touching the "real thing", thus
1343    eliminating the constant fear of accidentally checking something bogus
1344    back into the CVS repository...  (File config/targets.customized must
1345    not be added to the repository!)
1346    
1347    ----------------------------------------------------------------------
1348    Name: Matthias Blume
1349    Date: 2001/05/22 16:30:00 EDT
1350    Tag: blume-20010522-minitut
1351    Description:
1352    
1353    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
1354       struct returns.
1355    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
1356       incomplete, preliminary documentation for NLFFI.
1357    
1358    ----------------------------------------------------------------------
1359    Name: Matthias Blume
1360    Date: 2001/05/14 11:30:00 EDT
1361    Tag: blume-20010514-script
1362    Description:
1363    
1364    Some bugs in install script fixed.
1365    
1366    In addition to that I also made a slight change to the NLFFI API:
1367    Functors generated by ml-nlffigen now take the dynamic library as a
1368    straight functor argument, not as a suspended one.  (The original
1369    functor code used to force the suspension right away anyway, so there
1370    was nothing gained by this complication of the interface.)
1371    
1372    ----------------------------------------------------------------------
1373    Name: Matthias Blume
1374    Date: 2001/05/11 14:35:00 EDT
1375    Tag: blume-20010511-ml-nlffi
1376    Description:
1377    
1378    I finally took the plunge and added my new FFI code to the main
1379    repository.  For x86-linux it is now ready for prime-time.
1380    
1381    There are two new subdirectories of "src":
1382    
1383      - ml-nlffi-lib:
1384           The utility library for programs using the FFI interface.
1385           Here is the implementation of $/c.cm and its associated low-level
1386           partners $/c-int.cm and $/memory.cm.
1387      - ml-nlffigen:
1388           A stand-alone program for generating ML glue code from C source
1389           code.
1390    
1391    Building ml-nlffigen requires $/ckit-lib.cm.
1392    
1393    The config/install.sh script has been updates to do the Right Thing
1394    (hopefully).
1395    
1396    Notice that the source tree for the C-Kit will not be put under "src"
1397    but directly under the installation root directory.  (This is the
1398    structure that currently exists on the CVS server when you check out
1399    module "sml".)  Fortunately, config/install.sh knows about this oddity.
1400    
1401    Bugs: No documentation yet.
1402    
1403    ----------------------------------------------------------------------
1404    Name: Matthias Blume
1405    Date: 2001/05/09 16:35:00 EDT
1406    Tag: blume-20010509-cpscontract
1407    Description:
1408    
1409    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
1410    wrapper/unwrapper elimination did not decrement usage counts and some
1411    dead variables got overlooked by the dead-up logic.)
1412    
1413    ----------------------------------------------------------------------
1414    Name: Lal George
1415    Date: 2001/05/08  17:26:09 EDT
1416    Tag: george-20010508-omit-frameptr
1417    Description:
1418    
1419    Changes to implement the omit-frame-pointer optimization to support
1420    raw C calls. For now, there is only support on the Intel x86, but
1421    other architectures will follow as more experience is gained with this.
1422    
1423    
1424    ----------------------------------------------------------------------
1425    Name: Matthias Blume
1426    Date: 2001/05/07 14:40:00 EDT
1427    Tag: blume-20010507-proxies
1428    Description:
1429    
1430    I made into "proxy libraries" all libraries that qualify for such a
1431    change.  (A qualifying library is a library that has another library or
1432    groups as its sole member and repeats that member's export list
1433    verbatim.  A proxy library avoids this repetition by omitting its export
1434    list, effectively inheriting the list that its (only) member exports.
1435    See the CM manual for more explanation.)
1436    The main effect is that explicit export lists for these libraries
1437    do not have to be kepts in sync, making maintenance a bit easier.
1438    
1439    I also added copyright notices to many .cm-files.
1440    
1441    Last but not least, I made a new set of bootfiles.
1442    
1443    ----------------------------------------------------------------------
1444    Name: Matthias Blume
1445    Date: 2001/05/04 17:00:00 EDT
1446    Tag: blume-20010504-cm-lsplit
1447    Description:
1448    
1449    0. John merged pending changes to $/smlnj-lib.cm
1450    
1451    1. Allen's previous change accidentally backed out of one of Lal's
1452       earlier changes.  I undid this mistake (re-introducing Lal's change).
1453    
1454    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
1455       within the compiler where applicable.  There is some code simplification
1456       because of that.
1457    
1458    3. The "split" phase (in FLINT) is now part of the default list of phases.
1459       Compiler.Control.LambdaSplitting.* can be used to globally control the
1460       lambda-splitting (cross-module-inlining) engine.  In addition to that,
1461       it can now also be controlled on a per-source basis: CM has been taught
1462       a new tool parameter applicable to ML source files.
1463    
1464       - To turn lambda-splitting off completely:
1465            local open Compiler.Control.LambdaSplitting in
1466                val _ = set Off
1467            end
1468       - To make "no lambda-splitting" the global default (but allow per-source
1469         overriding); this is the initial setting:
1470            local open Compiler.Control.LambdaSplitting in
1471                val _ = set (Default NONE)
1472            end
1473       - To make "lambda-splitting with aggressiveness a" the global default
1474         (and allow per-source overriding):
1475            local open Compiler.Control.LambdaSplitting in
1476                val _ = set (Default (SOME a))
1477            end
1478    
1479       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
1480         write (in the respective .cm-file):
1481            a.sml (lambdasplitting:off)
1482       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
1483            a.sml (lambdasplitting:on)
1484       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
1485         <a> is a decimal non-negative integer):
1486            a.sml (lambdasplitting:<a>)
1487       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
1488            a.sml (lambdasplitting:infinity)
1489       - To use the global default for a.sml:
1490            a.sml (lambdasplitting:default)
1491         or simply
1492            a.sml
1493    
1494    ----------------------------------------------------------------------
1495    Name: Allen Leung
1496    Date: 2001/05/04 01:57:00 EDT
1497    Tag: leunga-20010504-sync
1498    Description:
1499    
1500      MLRISC features.
1501    
1502      1. Fix to CMPXCHG instructions.
1503      2. Changed RA interface to allow annotations in callbacks.
1504      3. Added a new method to the stream interface to allow annotations updates.
1505    
1506    ----------------------------------------------------------------------
1507    Name: Matthias Blume
1508    Date: 2001/05/01 11:45:00 EDT
1509    Tag: blume-20010501-pcedittmp
1510    Description:
1511    
1512    Changed install.sh to use the current working directory instead of
1513    /usr/tmp for a temporary file (pcedittmp).  The previous choice
1514    of /usr/tmp caused trouble with MacOS X because of file premission
1515    problems.
1516    
1517    ----------------------------------------------------------------------
1518    Name: Matthias Blume
1519    Date: 2001/04/20 11:10:00 EDT
1520    Tag: blume-20010420-inMLflag
1521    Description:
1522    
1523     - added vp_limitPtrMask to vproc-state.h
1524       (for use by the raw-C-calls mechanism to implement proper interrupt
1525        handling)
1526     - made the ML compiler aware of various data-structure offsets so it
1527       can generate code for accessing the vp_inML flag and vp_limitPtrMask
1528     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
1529       raw C-calls
1530    
1531    ----------------------------------------------------------------------
1532    Name: Lal George
1533    Date: 2001/04/20 09:15:28 EDT
1534    Tag: george-20010420-macosX
1535    Description:
1536    
1537     - Changes to port to Mac OS X; Darwin.
1538    
1539     - In the process I found that sqrt was broken on the PPC, because the
1540       fsqrt instruction is not implemented.
1541    
1542    ----------------------------------------------------------------------
1543    Name: Matthias Blume
1544    Date: 2001/04/18 12:45:00 EDT
1545    Tag: blume-20010418-ccalls
1546    Description:
1547    
1548     - fixed two off-by-4 errors in the x86-specific c-calls implementation
1549       (this bug prevented structure arguments containing pointers from being
1550        passed correctly)
1551     - changed the raw-C-call code in mlriscGen.sml in such a way that
1552       structure arguments are represented as a pointer to the beginning
1553       of the structure (instead of having a series of synthesized arguments,
1554       one for each structure member)
1555    
1556     - made makeml script's verbosity level configurable via environment
1557       variable (MAKEML_VERBOSITY)
1558    
1559     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
1560       in rawmem-x86.sml; we are now using the real thing
1561    
1562    ----------------------------------------------------------------------
1563    Name: Matthias Blume
1564    Date: 2001/03/22 16:25:00 EST
1565    Tag: blume-20010322-bootfiles
1566    Description:
1567    
1568    Created a new set of bootfiles (for your automatic installation convenience).
1569    
1570    ----------------------------------------------------------------------
1571    Name: Matthias Blume
1572    Date: 2001/03/22 15:10:00 EST
1573    Tag: blume-20010322-rawmem-parcm
1574    Description:
1575    
1576    1. All "raw memory access" primitives for the new FFI are implemented now
1577       (at least on the x86).
1578    2. Some further cleanup of CM's parallel make mechanism.
1579    
1580    ----------------------------------------------------------------------
1581    Name: Matthias Blume
1582    Date: 2001/03/19 17:53:00 EST
1583    Tag: blume-20010319-parallel
1584    Description:
1585    
1586    Parallel make (using compile servers) now works again.
1587    
1588    To this end, CM.stabilize and CMB.make have been modified to work in
1589    two passes when compile servers are attached:
1590       1. Compile everything, do not perform stabilization; this pass
1591          uses compile servers
1592       2. Stabilize everything; this pass does not use compile servers
1593    If there are no compile servers, the two passes are combined into one
1594    (as before).  Splitting the passes increases the inherent parallelism
1595    in the dependency graph because the entire graph including all
1596    libraries is available at the same time.  This, in turn, improves
1597    server utilization.  The downside is that the master process will
1598    have to do some extra work after compilation is done (because for
1599    technical reasons it must re-read all the binfiles during stabilization).
1600    
1601    ----------------------------------------------------------------------
1602    Name: Matthias Blume
1603    Date: 2001/03/16 12:22:00 EST
1604    Tag: blume-20010316-bootfiles
1605    Description:
1606    
1607    Created a new set of bootfiles (for your automatic installation convenience).
1608    
1609    ----------------------------------------------------------------------
1610    Name: Matthias Blume
1611    Date: 2001/03/16 11:00:00 EST
1612    Tag: blume-20010316-MLTREE-fixup
1613    Description:
1614    
1615    This is a minor fixup for an (untagged) earlier commit by Allen.
1616    (A file was missing).
1617    
1618    ----------------------------------------------------------------------
1619    Name: Allen Leung
1620    Date: Mon Mar  5 18:54:57 EST 2001
1621    Tag: leunga-20010305-cut-support
1622    
1623    1. New support for alternative control-flow in MLTREE.
1624       Currently we support
1625    
1626          FLOW_TO(CALL ...., [k1,...,kn])
1627    
1628       This is needed for 'cuts to' in C-- and try/handle-like constructs
1629       in Moby
1630    
1631       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
1632    
1633    2. Register Allocator
1634    
1635       Changes in interface [from Fermin, John]
1636    
1637    3. Alpha 8-bit SLL support [Fermin]
1638    
1639    4. All architectures
1640    
1641       A new module (ClusterExpandCopies) for expanding parallel copies.
1642    
1643    ----------------------------------------------------------------------
1644    Name: Allen Leung
1645    Date: 2001/02/27 23:07:00 EST
1646    Tag: leunga-20010227-minor-stuff
1647    
1648    1. Alpha bug fix for CMOVNE
1649    2. Handle mltree COND(..,FCMP ...,...)
1650    3. Bug fix in simplifier
1651    
1652    ----------------------------------------------------------------------
1653    Name: Matthias Blume
1654    Date: 2001/01/30 17:50:00 EST
1655    Tag: blume-20010130-sync
1656    Description:
1657    
1658    This is just a minor update to sync my devel branch with the main brach.
1659    The only visible change is the addition of some README files.
1660    
1661    ----------------------------------------------------------------------
1662    Name: Matthias Blume
1663    Date: 2001/01/12 23:30:00 JST
1664    Tag: blume-20010112-bootfiles
1665    Description:
1666    
1667    Made a new set of bootfiles that goes with the current state of the
1668    repository.
1669    
1670    ----------------------------------------------------------------------
1671    Name: Matthias Blume
1672    Date: 2001/01/12 21:20:00 JST
1673    Tag: blume-20010112-sync
1674    Description:
1675    
1676    I am just flushing out some minor changes that had accumulated in
1677    my private branch in order to sync with the main tree.  (This is
1678    mainly because I had CVS trouble when trying to merge _into_ my
1679    private branch.)
1680    
1681    Most people should be completely unaffected by this.
1682    
1683    ----------------------------------------------------------------------
1684    Name: Allen Leung
1685    Date: Thu Jan 11 21:03:00 EST 2001
1686    Tag: leunga-20010111-labexp=mltree
1687    Description:
1688    
1689    1.  Removed the type LabelExp and replace it by MLTree.
1690    2.  Rewritten mltree-simplify with the pattern matcher tool.
1691    3.  There were some bugs in alpha code generator which would break
1692        64-bit code generation.
1693    4.  Redo the tools to generate code with the
1694    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
1695        are now generated by perl scripts.
1696    
1697    ----------------------------------------------------------------------
1698    Name: Matthias Blume
1699    Date: 2001/01/10 21:55:00 JST
1700    Tag: blume-20010110-rcc
1701    Description:
1702    
1703    The RCC stuff now seems to work (but only on the x86).
1704    This required hacking of the c-calls interface (and -implementation) in
1705    MLRISC.
1706    
1707    Normal compiler users should be unaffected.
1708    
1709    ----------------------------------------------------------------------
1710    Name: Matthias Blume
1711    Date: 2001/01/09 01:20:00 JST
1712    Tag: blume-20010109-rcc
1713    Description:
1714    
1715    This is a fairly big patch, flushing out a large number of pending
1716    changes that I made to my development copy over the last couple of days.
1717    
1718    Of practical relevance at this moment is a workaround for a pickling
1719    bug that Allen ran into the other day.  The cause of the bug itself is
1720    still unknown and it might be hard to fix it properly, but the
1721    workaround has some merits of its own (namely somewhat reducing pickling
1722    overhead for certain libraries).  Therefore, I think this solution should
1723    be satisfactory at this time.
1724    
1725    The rest of the changes (i.e., the vast majority) has to do with my
1726    ongoing efforts of providing direct support for C function calls from
1727    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
1728    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
1729    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
1730    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
1731    handle RCC.
1732    
1733    The last part (the changes to mlriscGen.sml) are still known to be
1734    wrong on the x86 and not implemented on all other architectures.  But
1735    the infrastructure is in place. I had to change a few functor
1736    signatures in the backend to be able to route the CCalls interface
1737    from MLRISC there, and I had to specialize the mltree type (on the
1738    x86) to include the necessary extensions. (The extensions themselves
1739    were already there and redy to go in MLRISC/x86).
1740    
1741    Everything should be very happy as soon as someone helps me with
1742    mlriscGen.sml...
1743    
1744    In any case, nothing of this should matter to anyone as long as the
1745    new primop is not being used (which is going to be the case unless you
1746    find it where I hid it :). The rest of the compiler is completely
1747    unaffected.
1748    
1749    ----------------------------------------------------------------------
1750    Name: Matthias Blume
1751    Date: 2001/01/05 00:30:00 JST
1752    Tag: blume-20010105-primops
1753    Description:
1754    
1755    Added some experimental support for work that I am doing right now.
1756    These changes mostly concern added primops, but there is also a new
1757    experimental C library in the runtime system (but currently not enabled
1758    anywhere except on Linux/X86).
1759    
1760    In the course of adding primops (and playing with them), I discovered that
1761    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
1762    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
1763    To recover, I made type information in INL_PRIM mandatory and changed
1764    prim.sml as well as built-in.sml accordingly.  The InLine structure now
1765    has complete, correct type information (i.e., no bottom types).
1766    
1767    Since all these changes mean that we need new binfiles, I also bumped the
1768    version number to 110.32.1.
1769    
1770    ----------------------------------------------------------------------
1771    Name: Matthias Blume
1772    Date: 2000/12/30 22:10:00 JST
1773    Tag: blume-20001230-various
1774    Description:
1775    
1776    Added proxy libraries for MLRISC and let MLRISC libraries refer
1777    to each other using path anchors.  (See CM manual for explanation.)
1778    
1779    Updated CM documentation.
1780    
1781    Fixed some bugs in CM.
1782    
1783    Implemented "proxy" libraries (= syntactic sugar for CM).
1784    
1785    Added "-quiet" option to makeml and changed runtime system accordingly.
1786    
1787    Added cleanup handler for exportML to reset timers and compiler stats.
1788    
1789    ----------------------------------------------------------------------
1790    Name: Lal George
1791    Date: 2000/12/22 22:22:58 EST 2000
1792    Tag: Release_110_32
1793    Description:
1794    
1795            Infinite precision used throughout MLRISC.
1796            see MLRISC/mltree/machine-int.sig
1797    
1798    ----------------------------------------------------------------------
1799    Name: Matthias Blume
1800    Date: 2000/12/22 23:16:00 JST
1801    Tag: blume-20001222-warn
1802    Description:
1803    
1804    Corrected wording and formatting of some CM warning message which I
1805    broke in my previous patch.
1806    
1807    ----------------------------------------------------------------------
1808    Name: Matthias Blume
1809    Date: 2000/12/22 21:20:00 JST
1810    Tag: blume-20001222-anchorenv
1811    Description:
1812    
1813    Fixed CM's handling of anchor environments in connection with CMB.make.
1814    
1815    ----------------------------------------------------------------------
1816    Name: Matthias Blume
1817    Date: 2000/12/22 13:15:00 JST
1818    Tag: blume-20001222-cleanup
1819    Description:
1820    
1821    Removed src/cm/ffi which does not (and did not) belong here.
1822    
1823    ----------------------------------------------------------------------
1824    Name: Matthias Blume
1825    Date: 2000/12/21 23:55:00 JST
1826    Tag: blume-20001221-exn
1827    Description:
1828    
1829    Probably most important: CM no longer silently swallows all exceptions
1830    in the compiler.
1831    Plus: some other minor CM changes.  For example, CM now reports some
1832    sizes for generated binfiles (code, data, envpickle, lambdapickle).
1833    
1834    ----------------------------------------------------------------------
1835    Name: Matthias Blume
1836    Date: 2000/12/15 00:01:05 JST
1837    Tag: blume-20001215-dirtool
1838    Description:
1839    
1840    - "dir" tool added.
1841    - improvements and cleanup to Tools structure
1842    - documentation updates
1843    
1844    ----------------------------------------------------------------------
1845    Name: Allen Leung
1846    Date: Thu Dec 14 03:45:24 EST 2000
1847    Description:
1848    Tag:  leunga-20001214-int-inf
1849    Description:
1850    
1851       In IntInf, added these standard functions, which are missing from our
1852    implementation:
1853    
1854        andb : int * int -> int
1855        xorb : int * int -> int
1856        orb  : int * int -> int
1857        notb : int -> int
1858         <<   : int * word -> int
1859        ~>>  : int * word -> int
1860    
1861       Not tested, I hope they are correct.
1862    
1863    ----------------------------------------------------------------------
1864    Name: Allen Leung
1865    Date: Fri Dec  8 19:23:26 EST 2000
1866    Description:
1867    Tag:  leunga-20001208-nowhere
1868    Description:
1869    
1870      Slight improvements to the 'nowhere' tool to handle OR-patterns,
1871    to generate better error messages etc.  Plus a brief manual.
1872    
1873    ----------------------------------------------------------------------
1874    Name: Lal George
1875    Date: 2000/12/08 09:54:02 EST 2000
1876    Tag: Release_110_31
1877    Description:
1878    
1879    - Version 110.31
1880    ----------------------------------------------------------------------
1881    Name: Allen Leung
1882    Date: Thu Dec  7 22:01:04 EST 2000
1883    Tag:  leunga-20001207-cell-monster-hack
1884    Description:
1885    
1886    Major MLRISC internal changes.  Affect all clients.
1887    Summary:
1888    
1889    1.  Type CELLS.cell = int is now replaced by a datatype.
1890        As a result, the old regmap is now gone.  Almost all interfaces
1891        in MLRISC change as a consequence.
1892    
1893    2.  A new brand version of machine description tool (v3.0) that generates
1894        modules expecting the new interface.  The old version is removed.
1895    
1896    3.  The RA interface has been further abstracted into two new functors.
1897        RISC_RA and X86RA.  These functors have much simpler interfaces.
1898        [See also directory MLRISC/demo.]
1899    
1900    4.  Some other new source->source code generation tools are available:
1901    
1902        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
1903        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
1904           I use this tool to generate the peephole optimizers---with the new
1905           cell type changes, peephole rules are becoming difficult to write
1906           without conditional pattern matching.
1907    
1908    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
1909        cover the entire MLRISC source tree so many things broke.
1910    
1911    6.  CM files have been moved to the subdirectory MLRISC/cm.
1912        They are moved because there are a lot of them and they clutter up the
1913        root dir.
1914    
1915    7.  More detailed documentation to come...
1916    
1917        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
1918        a makeml -rebuild first.  This is because of other other
1919        changes that Matthias has made (see below).
1920    
1921    
1922    ----------------------------------------------------------------------
1923    Name: Matthias Blume
1924    Date: 2000/11/30 23:12:00 JST
1925    Tag: blume-20001130-filereorg
1926    Description:
1927    
1928    Some manual updates and some file reorganizations in CM.
1929    
1930    ----------------------------------------------------------------------
1931    Name: Matthias Blume
1932    Date: 2000/11/24 17:45:00 JST
1933    Tag: blume-20001124-link
1934    Description:
1935    
1936    Drastically improved link traversal code for the case that the dynamic
1937    value was already loaded at bootstrap time.  As a result, CM and CMB
1938    now both load blazingly fast -- even on a very slow machine.  Also,
1939    memory consumption has been further reduced by this.
1940    
1941    Warning: The format of the PIDMAP file has changed.  THerefore, to
1942    bootstrap you have to do this:
1943    
1944    1. Run CMB.make
1945    2. Make a symbolic link for the boot directory:
1946         ln -s sml.boot.ARCH-OS xxx
1947    3. "Rebuild" the boot directory:
1948         ./makeml -boot xxx -rebuild sml ; rm xxx
1949    4. Boot normally:
1950          ./makeml
1951    
1952    ----------------------------------------------------------------------
1953    Name: Matthias Blume
1954    Date: 2000/11/21 21:20:00 JST
1955    Tag: blume-20001121-tools
1956    Description:
1957    
1958    Continued hacking on autoloading problem -- with success this time.
1959    Also changed tool-plugin mechanism.  See new CM manual.
1960    
1961    ----------------------------------------------------------------------
1962    Name: Matthias Blume
1963    Date: 2000/11/19 14:30:00 JST
1964    Tag:  blume-20001119-autoload
1965    Description:
1966    
1967    Some hacking to make autoloading faster.  Success for CMB, no success
1968    so far for CM.  There is a reduced structure CM' that autoloads faster.
1969    (This is a temporary, non-documented hack to be eliminated again when
1970    the general problem is solved.)
1971    
1972    ----------------------------------------------------------------------
1973    Name: Matthias Blume
1974    Date: 2000/11/17 14:10:00 JST
1975    Tag: blume-20001117-pickle-lib
1976    Description:
1977    
1978    1. Eliminated comp-lib.cm
1979    2. Made pickle-lib.cm
1980    3. Eliminated all uses of intset.sml (from comp-lib.cm)
1981    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
1982       equivalent constructs from smlnj-lib.cm (INtHashTable).
1983    5. Point 4. also goes for those uses of intmap.* in MLRISC.
1984       Duplicated intmap modules thrown out.
1985    6. Hunted down all duplicated SCC code and replaced it with
1986       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
1987    7. Rewrote Feedback module.
1988    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
1989       should be thrown out and equivalent modules from smlnj-lib.cm
1990       should be used (IntRedBlackSet, IntListSet, ...).
1991    
1992    Confirmed that compiler compiles to fixpoint.
1993    
1994    ----------------------------------------------------------------------
1995    Name: Allen Leung
1996    Date: 2000/11/10 18:00:00
1997    Tag: leunga-20001110-new-x86-fp
1998    
1999    A new x86 floating point code generator has been added.
2000    By default this is turned off.  To turn this on, do:
2001    
2002        CM.autoload "$smlnj/compiler.cm";
2003        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
2004    
2005    Changes:
2006    
2007    1.  Changed FTAN to FPTAN so that the assembly output is correct.
2008    2.  Changed the extension callback for FTANGENT to generate:
2009    
2010              fptan
2011              fstp  %st(0)
2012        instead of
2013              fptan
2014              fstpl ftempmem
2015    
2016    3.  Numerous assembly fixes for x86.
2017    
2018    5.  Cleaned up the machine code output module x86/x86MC.sml and added
2019        support for a whole bunch of instructions and addressing modes:
2020    
2021          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
2022          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
2023          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
2024          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
2025          fxch %st(n)
2026          fld %st(n)
2027          fst %st(n)
2028          fst mem
2029          fstp %st(n)
2030          fucom %st(n)
2031          fucomp %st(n)
2032    
2033        All these are now generated when the fast fp mode is turned on.
2034    
2035    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
2036    
2037    ----------------------------------------------------------------------
2038    Name: Matthias Blume
2039    Date: 2000/11/09 11:20:00 JST
2040    Tag: blume-20001109-scc
2041    Description:
2042    
2043    Eliminated some code duplication:
2044    
2045    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
2046       (Otherwise the functor is useless.)
2047    2. Used GraphSCCFn where SCCUtilFun was used previously.
2048    3. Got rid of SCCUtilFun (in comp-lib.cm).
2049    
2050    ----------------------------------------------------------------------
2051    Name: Lal George
2052    Date: 2000/11/06 09:02:21 EST 2000
2053    Tag: Release_110_30
2054    Description:
2055    
2056    - Version 110.30
2057    ----------------------------------------------------------------------
2058    Name: Matthias Blume
2059    Date: 2000/11/04 14:45:00
2060    Tag: blume-20001104-mlbuild
2061    Description:
2062    
2063    - Made ml-build faster on startup.
2064    - Documentation fixes.
2065    
2066    ----------------------------------------------------------------------
2067    Name: Matthias Blume
2068    Date: 2000/11/02 17:00:00 JST
2069    Tag: blume-20001102-condcomp
2070    Description:
2071    
2072    - Small tweaks to pickler -- new BOOTFILES!
2073    - Version bumped to 110.29.2.
2074    - Added conditional compilation facility to init.cmi (see comment there).
2075    ----------------------------------------------------------------------
2076    Name: Allen Leung
2077    Date: 2000/10/23 19:31:00
2078    Tag: leunga-20001023-demo-ra
2079    
2080    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
2081    2. Test programs for the graph library updated
2082    3. Some new MLRISC demo programs added
2083    
2084    ----------------------------------------------------------------------
2085    Name: Matthias Blume
2086    Date: 2000/08/31 22:15:00 JST
2087    Tag: blume-20001017-errmsg
2088    Description:
2089    
2090    More error message grief: Where there used to be no messages, there
2091    now were some that had bogus error regions.  Fixed.
2092    
2093    ----------------------------------------------------------------------
2094    Name: Matthias Blume
2095    Date: 2000/08/31 17:30:00 JST
2096    Tag: blume-20001017-v110p29p1
2097    Description:
2098    
2099    I made a version 110.29.1 with new bootfiles.
2100    
2101    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
2102              CM documentation changes and a small bugfix in CM's error reporting.
2103    
2104    ----------------------------------------------------------------------
2105    Name: Lal George
2106    Date: 2000/09/27 14:42:35 EDT
2107    Tag: george-20000927-nodestatus
2108    Description:
2109    
2110    Changed the type of the nodestatus, so that:
2111    
2112            SPILLED(~1)             is now SPILLED
2113            SPILLED(m) where m>=0   is now MEMREG(m)
2114            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
2115    
2116    ----------------------------------------------------------------------
2117    Name: Matthias Blume
2118    Date: 2000/09/07 14:45:00 JST
2119    Tag: blume-20000907-cmerrmsg
2120    Description:
2121    
2122    Small tweak to CM to avoid getting ML syntax error messages twice.
2123    
2124    ----------------------------------------------------------------------
2125    Name: Matthias Blume
2126    Date: 2000/08/31 18:00:00 JST
2127    Tag: blume-20000831-cvsbootfiles
2128    Description:
2129    
2130    New URL for boot files (because the 110.29 files on the BL server do
2131    now work correctly with my updated install scripts for yacc and lex).
2132    
2133    ----------------------------------------------------------------------
2134    Name: Matthias Blume
2135    Date: 2000/08/08 12:33:00 JST
2136    Tag: blume-20000808-manual
2137    Description:
2138    
2139    Tiny update to CM manual.
2140    
2141    ----------------------------------------------------------------------
2142    Name: Allen Leung
2143    Date: 2000/08/7 19:31:00
2144    Tag: leunga-20000807-a-whole-bunch-of-stuff
2145    
2146      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
2147    and Mobdy.
2148    
2149    1.  x86
2150    
2151       a.  Fixes to peephole module by John and Dan.
2152       b.  Assembly fix to SETcc by Allen.
2153       c.  Fix to c-call by John.
2154       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
2155       e.  Instruction selection optimization to SETcc as suggested by John.
2156    
2157           For example,
2158    
2159            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
2160    
2161           should generate:
2162    
2163            MOVL a, x
2164            SUBL b, x
2165            SHRL 31, x
2166    
2167    2.  IR stuff
2168    
2169         A bunch of new DJ-graph related algorithms added.  These
2170         speed up SSA construction.
2171    
2172    3.  SSA + Scheduling
2173    
2174         Added code for SSA and scheduling to the repository
2175    
2176    ----------------------------------------------------------------------
2177    Name: Lal George
2178    Date: 2000/07/27 11:53:14 EDT
2179    
2180    Tag: lal-20000727-linux-ppc
2181    Description:
2182    
2183     Made changes to support Linux PPC.
2184     p.s. I have confirmation that the 110.29 boot files work fine.
2185    
2186    ----------------------------------------------------------------------
2187    Name: Matthias Blume
2188    Date: 2000/07/27 17:40:00 JST
2189    Tag: blume-20000727-scripts
2190    Description:
2191    
2192    !!!! WARNING !!!!
2193    You must recompile the runtime system!
2194    !!!! WARNING !!!!
2195    
2196    This is basically another round of script-enhancements:
2197    
2198    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
2199       and undefine CM preprocessor symbols.
2200    
2201    2. ml-build avoids generating a new heap image if it finds that the
2202       existing one is still ok.  (The condition is that no ML file had to
2203       be recompiled and all ML files are found to be older that the heap
2204       file.)
2205    
2206       To make this work smoothly, I also hacked the runtime system as
2207       well as SMLofNJ.SysInfo to get access to the heap image suffix
2208       (.sparc-solaris, ...) that is currently being used.
2209    
2210       Moreover, the signature of CM.mk_standalone has changed.  See the
2211       CM manual.
2212    
2213    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
2214       CM manual for details.)
2215    
2216    4. More CM manual updates:
2217        - all of the above has been documented.
2218        - there is now a section describing the (CM-related) command line
2219          arguments that are accepted by the "sml" command
2220    
2221    ----------------------------------------------------------------------
2222    Name: Matthias Blume
2223    Date: 2000/07/25 16:20:00 JST
2224    Tag: blume-20000725-makedepend
2225    Description:
2226    
2227    Added a script called ml-makedepend.  This can be used in makefiles
2228    for Unix' make in a way very similar to the "makedepend" command for
2229    C.
2230    
2231    The script internally uses function CM.sources.
2232    
2233    Synopsis:
2234    
2235        ml-makedepend [-f makefile] cmfile targetname
2236    
2237    The default for the makefile is "makefile" (or "Makefile" should
2238    "makefile" not exist).
2239    
2240    ml-makedepend adds a cmfile/targetname-specific section to this
2241    makefile (after removing the previous version of this section).  The
2242    section contains a single dependency specification with targetname on
2243    the LHS (targetname is an arbitrary name), and a list of files derived
2244    from the cmfile on the RHS.  Some of the files on the RHS are
2245    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
2246    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
2247    path names.  The makefile writer is responsible for making sure that
2248    these variables have correct at the time "make" is invoked.
2249    
2250    ----------------------------------------------------------------------
2251    Name: Matthias Blume
2252    Date: 2000/07/22 23:30:00 JST
2253    Tag: blume-20000722-urlupdate
2254    Description:
2255    
2256    Changed BOOT and config/srcarchiveurl to point to BL server:
2257    
2258        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
2259    
2260    ----------------------------------------------------------------------
2261    Name: Matthias Blume
2262    Date: 2000/07/18 18:00:00 JST
2263    Tag: blume-20000718-Version_110_29
2264    Description:
2265    
2266    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
2267    
2268    2. Updated config/version to 110.29
2269    
2270    3. Updated config/srcarchiveurl
2271    
2272    3. New boot files!
2273       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
2274    
2275    ----------------------------------------------------------------------
2276    Name: Matthias Blume
2277    Date: 2000/07/11 13:58:00 JST
2278    Tag: blume-20000711-doctypo
2279    Description:
2280    
2281    Fixed a few typos in CM manual.
2282    
2283    ----------------------------------------------------------------------
2284    Name: Allen Leung
2285    Date: 2000/06/15 00:38:00
2286    Tag: leunga-20000704-sparc-x86
2287    
2288    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
2289    2. fix to x86 RET bug [found by Dan Grossman]
2290    3. sparc assembly bug fix for ticc instructions [found by Fermin]
2291    
2292       Affects c-- and moby only
2293    
2294    ----------------------------------------------------------------------
2295    Name: Matthias Blume
2296    Date: 2000/07/04 15:26:00
2297    Tag: blume-20000704-trigger
2298    Description:
2299    
2300    1. Improvements to CM manual.
2301    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
2302       of getting a back-trace.  The function, when called, raises an
2303       internal exception which explicitly carries the full back-trace history,
2304       so it is unaffected by any intervening handle-raise pairs ("trivial"
2305       or not).  The interactive loop will print that history once it arrives
2306       at top level.
2307       Short of having all exceptions implicitly carry the full history, the
2308       recommended way of using this facility is:
2309         - compile your program with instrumentation "on"
2310         - run it, when it raises an exception, look at the history
2311         - if the history is "cut off" because of some handler, go and modify
2312           your program so that it explicitly calls BTrace.trigger
2313         - recompile (still instrumented), and rerun; look at the full history
2314    
2315    ----------------------------------------------------------------------
2316    Name: Matthias Blume
2317    Date: 2000/07/03 15:36:00 JST
2318    Tag: blume-20000702-manual
2319    Description:
2320    
2321    Small corrections and updates to CM manual.
2322    
2323    ----------------------------------------------------------------------
2324    Name: Matthias Blume
2325    Date: 2000/06/29 16:04:00 JST
2326    Tag: blume-20000629-yacctool
2327    Description:
2328    
2329    Changes:
2330    
2331    1. Class "mlyacc" now takes separate arguments to pass options to
2332       generated .sml- and .sig-files independently.
2333    2. Corresponding CM manual updates.
2334    3. BTrace module now also reports call sites.  (However, for loop clusters
2335       it only shows from where the cluster was entered.)  There are associated
2336       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
2337    
2338    ----------------------------------------------------------------------
2339    Name: Matthias Blume
2340    Date: 2000/06/27 16:51:00 JST
2341    Tag: blume-20000627-noweb
2342    Description:
2343    
2344    Changes:
2345    
2346     1. Implemented "subdir" and "witness" options for noweb tool.
2347        This caused some slight internal changes in CM's tool implementation.
2348     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
2349        some remaining issues from earlier path anchor changes.
2350     3. Updated CM manual accordingly.
2351    
2352     4. Changed implementation of back-tracing so that I now consider it
2353        ready for prime-time.
2354    
2355        In particular, you don't have to explicitly trigger the back-trace
2356        anymore.  Instead, if you are running BTrace-instrumented code and
2357        there is an uncaught exception (regardless of whether or not it was
2358        raised in instrumented code), the top-level evalloop will print
2359        the back-trace.
2360    
2361        Features:
2362    
2363          - Instrumented and uninstrumented code work together seemlessly.
2364            (Of course, uninstrumented code is never mentioned in actual
2365             back-traces.)
2366    
2367          - Asymptotic time- and space-complexity of instrumented code is
2368            equal to that of uninstrumented code.  (This means that
2369            tail-recursion is preserved by the instrumentation phase.)
2370    
2371          - Modules whose code has been instrumented in different sessions
2372            work together without problem.
2373    
2374          - There is no penalty whatsoever on uninstrumented code.
2375    
2376          - There is no penalty on "raise" expressions, even in
2377            instrumented code.
2378    
2379        A potential bug (or perhaps it is a feature, too):
2380    
2381          A back-trace reaches no further than the outermost instrumented
2382          non-trivial "raise".  Here, a "trivial" raise is one that is the
2383          sole RHS of a "handle" rule.  Thus, back-traces reach trough
2384    
2385               <exp> handle e => raise e
2386    
2387          and even
2388    
2389               <exp> handle Foo => raise Bar
2390    
2391          and, of course, through
2392    
2393               <exp> handle Foo => ...
2394    
2395         if the exception was not Foo.
2396    
2397         Back-traces always reach right through any un-instrumented code
2398         including any of its "handle" expressions, trivial or not.
2399    
2400       To try this out, do the following:
2401    
2402         - Erase all existing binfiles for your program.
2403           (You may keep binfiles for those modules where you think you
2404            definitely don't need back-tracing.)
2405         - Turn on back-trace instrumentation:
2406              SMLofNJ.Internals.BTrace.mode (SOME true);
2407         - Recompile your program.  (I.e., run "CM.make" or "use".)
2408         - You may now turn instrumentation off again (if you want):
2409              SMLofNJ.Internals.BTrace.mode (SOME false);
2410         - Run your program as usual.  If it raises an exception that
2411           reaches the interactive toplevel, then a back-trace will
2412           automatically be printed.  After that, the toplevel loop
2413           will print the exception history as usual.
2414    
2415    ----------------------------------------------------------------------
2416    Name: Matthias Blume
2417    Date: 2000/06/26 09:56:46 JST
2418    Tag: blume-20000626-setup
2419    Description:
2420    
2421    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
2422          ML code before and after compiling a file (e.g., to set compiler
2423          flags)
2424    
2425    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
2426              - associated changes to btrace.sml (BTrace instrumentation pass)
2427              - cleaner implementation of btimp.sml (BTrace tracing and report
2428                module)
2429    
2430    CM manual: * new path encoding documented
2431               * description of setup-parameter to "sml" added
2432    
2433    The biggest user-visible change to back-tracing is that it is no
2434    longer necessary to compile all traced modules within the same
2435    session.  (This was a real limitation.)
2436    
2437    ----------------------------------------------------------------------
2438    Name: Matthias Blume
2439    Date: 2000/06/24 12:40:00 JST
2440    Tag: blume-20000624-startup
2441    Description:
2442    
2443    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
2444    bit too often -- to put it mildly. :)
2445    
2446    ----------------------------------------------------------------------
2447    Name: Matthias Blume
2448    Date: 2000/06/23 18:20:00 JST
2449    Tag: blume-20000623-btrace
2450    Description:
2451    
2452    This updates adds a backtrace facility to aid programmers in debugging
2453    their programs.  This involves the following changes:
2454    
2455    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
2456       keeping track of the current call stack.  When programs are compiled
2457       in a special mode, the compiler will insert calls to these hooks
2458       into the user program.
2459       "Hook" means that it is possible for different implementations of
2460       back-tracing to register themselves (at different times).
2461    
2462    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
2463       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
2464       It can be turned on using this call:
2465         SMLofNJ.Internals.BTrace.mode (SOME true);
2466       Turning it off again:
2467         SMLofNJ.Internals.BTrace.mode (SOME false);
2468       Querying the current status:
2469         SMLofNJ.Internals.BTrace.mode NONE;
2470       Annotated programs are about twice as big as normal ones, and they
2471       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
2472       where all hooks do nothing).  The slowdown with a plugin that is
2473       actually useful (such as the one supplied by default) is even greater,
2474       but in the case of the default plugin it is still only an constant
2475       factor (amortized).
2476    
2477    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
2478       with a sub-structure BTrace for controlling back-tracing.  In particular,
2479       the above-mentioned function "mode" controls whether the annotation
2480       phase is invoked by the compiler.  Another important function is
2481       "trigger": when called it aborts the current execution and causes
2482       the top-level loop to print a full back-trace.
2483    
2484    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
2485       for back-tracing.  It keeps track of the dynamic call stack and in
2486       addition to that it keeps a partial history at each "level" of that
2487       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
2488       d and b (at separate times, dynamically), then the report will show:
2489    
2490       GOTO   d
2491             /c
2492       GOTO  \b
2493       CALL   a
2494    
2495       This shows that there was an initial non-tail call of a, then a
2496       tail-call to b or c, looping behavior in a cluster of functions that
2497       consist of b and c, and then a goto from that cluster (i.e., either from
2498       b or from c) to d.
2499    
2500       Note that (depending on the user program) the amount of information
2501       that the back-trace module has to keep track of at each level is bounded
2502       by a constant.  Thus, the whole implementation has the same asymptotical
2503       complexity as the original program (both in space and in time).
2504    
2505    5. compiler/TopLevel/interact/evalloop.sml has been modified to
2506       handle the special exception SMLofNJ.Internals.BTrace.BTrace
2507       which is raised by the "trigger" function mentioned above.
2508    
2509    Notes on usage:
2510    
2511    - Annotated code works well together with unannotated code:
2512    Unannotated calls simply do not show up at all in the backtrace.
2513    
2514    - It is not a good idea to let modules that were annotated during
2515    different sessions run at the same time.  This is because the compiler
2516    chooses small integers to identify individual functions, and there
2517    will be clashes if different modules were compiled in separate sessions.
2518    (Nothing will crash, and you will even be told about the clashes, but
2519    back-trace information will in general not be useful.)
2520    
2521    - Back-tracing can be confused by callcc and capture.
2522    
2523    - The only way of getting a back-trace right now is to explicitly
2524    invoke the "trigger" function from your user program.  Eventually, we
2525    should make every exception carry back-trace information (if
2526    available).  But since this creates more overhead at "raise"-time
2527    (similar to the current exnHistory overhead), I have not yet
2528    implemented this.  (The implementation will be rather easy.)  With
2529    exceptions carrying back-trace information, this facility will be even
2530    more useful because users don't need to modify their programs...
2531    
2532    - While it is possible to compile the compiler with back-trace
2533    annotations turned on (I did it to get some confidence in
2534    correctness), you must make absolutely sure that core.sml and
2535    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
2536    be compiled with annotation because there is no core access yet, but
2537    if you compile btimp.sml with annotation, then the system will go into
2538    an infinite recursion and crash.)
2539    Since CM currently does not know about BTrace, the only way to turn
2540    annotations on and off for different modules of the compiler is to
2541    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
2542    this is awkward and clumsy.
2543    
2544    Sample sessions:
2545    
2546    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
2547    - SMLofNJ.Internals.BTrace.mode (SOME true);
2548    [autoloading]
2549    [autoloading done]
2550    val it = false : bool
2551    - structure X = struct
2552    -     fun main n = let
2553    -         fun a (x, 0) = d x
2554    -           | a (x, n) = b (x, n - 1)
2555    -         and b (x, n) = c (x, n)
2556    -         and c (x, n) = a (x, n)
2557    -         and d x = e (x, 3)
2558    -         and e (x, 0) = f x
2559    -           | e (x, n) = e (x, n - 1)
2560    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
2561    -           | f n = n * g (n - 1)
2562    -         and g n = a (n, 3)
2563    -     in
2564    -         f n
2565    -     end
2566    - end;
2567    structure X : sig val main : int -> int end
2568    - X.main 3;
2569    *** BACK-TRACE ***
2570    GOTO   stdIn:4.2-13.20: X.main[2].f
2571    GOTO-( stdIn:4.2-13.20: X.main[2].e
2572    GOTO   stdIn:4.2-13.20: X.main[2].d
2573         / stdIn:4.2-13.20: X.main[2].a
2574         | stdIn:4.2-13.20: X.main[2].b
2575    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2576    CALL   stdIn:4.2-13.20: X.main[2].g
2577    GOTO   stdIn:4.2-13.20: X.main[2].f
2578    GOTO-( stdIn:4.2-13.20: X.main[2].e
2579    GOTO   stdIn:4.2-13.20: X.main[2].d
2580         / stdIn:4.2-13.20: X.main[2].a
2581         | stdIn:4.2-13.20: X.main[2].b
2582    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2583    CALL   stdIn:4.2-13.20: X.main[2].g
2584    GOTO   stdIn:4.2-13.20: X.main[2].f
2585    GOTO-( stdIn:4.2-13.20: X.main[2].e
2586    GOTO   stdIn:4.2-13.20: X.main[2].d
2587         / stdIn:4.2-13.20: X.main[2].a
2588         | stdIn:4.2-13.20: X.main[2].b
2589    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2590    CALL   stdIn:4.2-13.20: X.main[2].g
2591    GOTO   stdIn:4.2-13.20: X.main[2].f
2592    CALL   stdIn:2.15-17.4: X.main[2]
2593    -
2594    
2595    (Note that because of a FLINt bug the above code currently does not
2596    compile without BTrace turned on.)
2597    
2598    Here is another example, using my modified Tiger compiler:
2599    
2600    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
2601    - SMLofNJ.Internals.BTrace.mode (SOME true);
2602    [autoloading]
2603    [autoloading done]
2604    val it = false : bool
2605    - CM.make "sources.cm";
2606    [autoloading]
2607    ...
2608    [autoloading done]
2609    [scanning sources.cm]
2610    [parsing (sources.cm):parse.sml]
2611    [creating directory CM/SKEL ...]
2612    [parsing (sources.cm):tiger.lex.sml]
2613    ...
2614    [wrote CM/sparc-unix/semant.sml]
2615    [compiling (sources.cm):main.sml]
2616    [wrote CM/sparc-unix/main.sml]
2617    [New bindings added.]
2618    val it = true : bool
2619    - Main.compile ("../testcases/merge.tig", "foo.out");
2620    *** BACK-TRACE ***
2621    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
2622    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2623    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2624    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2625    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2626    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2627    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
2628         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
2629    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
2630    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2631    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
2632    CALL   main.sml:1.18-118.4: Main.compile[2]
2633    -
2634    
2635    ----------------------------------------------------------------------
2636    Name: Matthias Blumen
2637    Date: 2000/06/21 18:00:00 JST
2638    Tag: blume-20000621-manual
2639    Description:
2640    
2641    CM manual update: Path environments documented.
2642    
2643    ----------------------------------------------------------------------
2644    Name: Matthias Blume
2645    Date: 2000/06/19 13:40:00
2646    Tag: blume-20000619-manual
2647    Description:
2648    
2649    CM manual and system/README update.  This only covers the fact that
2650    there are no more implicit anchors.  (Path environments and the "bind"
2651    option to "cm" have yet to be documented.)
2652    
2653    ----------------------------------------------------------------------
2654    Name: Matthias Blume
2655    Date: 2000/06/19 11:05:00 JST
2656    Tag: blume-20000619-chdir-bugfix
2657    Description:
2658    
2659    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
2660    
2661    ----------------------------------------------------------------------
2662    Name: Matthias Blume
2663    Date: 2000/06/18 22:00:10 JST
2664    Tag: blume-20000618-implicit-anchors-really-gone
2665    Description:
2666    
2667    I updates the previous HISTORY entry where I forgot to mention that
2668    implicit anchors are no longer with us.
2669    
2670    The current update also gets rid of the (now useless) controller
2671    CM.Control.implicit_anchors.
2672    
2673    ----------------------------------------------------------------------
2674    Name: Matthias Blume
2675    Date: 2000/06/16 17:30:00 JST
2676    Tag: blume-20000616-anchorenv
2677    Description:
2678    
2679    This patch implements the long anticipated (just kidding :) "anchor
2680    environment" mechanism.  In the course of doing this, I also
2681    re-implemented CM's internal "SrcPath" module from scratch.  The new
2682    one should be more robust in certain boundary cases.  In any case, it
2683    is a lot cleaner than its predecessor (IMHO).
2684    
2685    This time, although there is yet another boot file format change, I
2686    kept the unpickler backward-compatible.  As a result, no new bootfiles
2687    are necessary and bootstrapping is straightforward.  (You cannot read
2688    new bootfiles into an old system, but the other way around is no
2689    problem.)
2690    
2691    Visible changes:
2692    
2693    ** 0. Implicit path anchors (without the leading $-symbol) are no
2694    longer recognized at all. This means that such path names are not
2695    illegal either.  For example, the name basis.cm simply refers to a
2696    local file called "basis.cm" (i.e, the name is an ordinary path
2697    relative to .cm-files directory).  Or, to put it differently, only
2698    names that start with $ are anchored paths.
2699    
2700    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
2701    vanished.
2702    
2703    John (Reppy) had critizised this as soon as I originally proposed and
2704    implemented it, but at that time I did not really deeply believe
2705    him. :) Now I came full-circle because I need the $<singlearc> syntax
2706    in another place where it cannot be seen as an abbreviation for
2707    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
2708    seems to mean (i.e., it "expands" into the corresponding anchor
2709    value).
2710    
2711    However, when paths are used as members in CM description files, it
2712    continues to be true that there must be at least another arc after the
2713    anchor.  This is now enforced separately during semantic analysis
2714    (i.e., from a lexical/syntactical point of view, the notation is ok.)
2715    
2716    ** 2. The "cm" class now accepts an option "bind".  The option's value
2717    is a sub-option list of precisely two items -- one labeled "anchor"
2718    and the other one labeled "value".  As you might expect, "anchor" is
2719    used to specify an anchor name to be bound, and "value" specifies what
2720    the anchor is being bound to.
2721    
2722    The value must be a directory name and can be given in either standard
2723    syntax (including the possibility that it is itself an anchored path)
2724    or native syntax.
2725    
2726    Examples:
2727    
2728       foo.cm (bind:(anchor:bar value:$mystuff/bar))
2729       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
2730    
2731    and so on.
2732    
2733    The meaning of this is that the .cm-file will be processed with an
2734    augmented anchor environment where the given anchor(s) is/are bound to
2735    the given values(s).
2736    
2737    The rationale for having this feature is this: Suppose you are trying
2738    to use two different (already stable) libraries a.cm and b.cm (that
2739    you perhaps didn't write yourself).  Further, suppose each of these
2740    two libraries internally uses its own auxiliary library $aux/lib.cm.
2741    Normally you would now have a problem because the anchor "lib" can not
2742    be bound to more than one value globally.  Therefore, the project that
2743    uses both a.cm and b.cm must locally redirect the anchor to some other
2744    place:
2745    
2746       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
2747       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
2748    
2749    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
2750    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
2751    
2752    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
2753    you when you do so at the time of CM.stabilize).  Therefore, you can
2754    also use an anchored path as the value:
2755    
2756      a.cm (bind:(anchor:lib value:$a-lib))
2757      b.cm (bind:(anchor:lib value:$b-lib))
2758    
2759    Now you can globally configure (using the usual CM.Anchor.anchor or
2760    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
2761    itself is always locally bound, setting it globally is no longer
2762    meaningful or necessary (but it does not hurt either).  In fact, "lib"
2763    can still be used as a global anchor for separate purposes.  As a
2764    matter of fact, one can locally define "lib" in terms of a global
2765    "lib":
2766    
2767      a.cm (bind:(anchor:lib value:$lib/a))
2768      b.cm (bind:(anchor:lib value:$lib/b))
2769    
2770    ** 3: The encoding of path names has changed.  This affects the way
2771    path names are shown in CM's progress report and also the internal
2772    protocol encoding used for parallel make.
2773    
2774    The encoding now uses one or more ':'-separated segments.  Each
2775    segments corresponds to a file that has been specified relative to the
2776    file given by its preceding segment.  The first segment is either
2777    relative to the CWD, absolute, or anchored.  Each segment itself is
2778    basically a Unix pathname; all segments but the first are relative.
2779    
2780    Example:
2781    
2782       $foo/bar/baz.cm:a/b/c.sml
2783    
2784    This path denotes the file bar/a/b/c.sml relative to the directory
2785    denoted by anchor "foo".  Notice that the encoding also includes
2786    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
2787    paths are resolved relative to the .cm-files directory, so baz.cm must
2788    be ignored to get the "real" pathname.
2789    
2790    To make this fact more obvious, CM puts the names of such "virtual
2791    arcs" into parentheses when they appear in progress reports. (No
2792    parentheses will appear in the internal protocol encoding.)  Thus,
2793    what you really see is:
2794    
2795      $foo/bar/(baz.cm):a/b/c.sml
2796    
2797    I find this notation to be much more informative than before.
2798    
2799    Another new feature of the encoding is that special characters
2800    including parentheses, colons, (back)slashes, and white space are
2801    written as \ddd (where ddd is the decimal encoding of the character).
2802    
2803    *** The CM manual still needs to be updated.
2804    
2805    ----------------------------------------------------------------------
2806    Name: Allen Leung
2807    Date: 2000/06/15 00:38:00
2808    Tag: leunga-20000615-x86-peephole
2809    
2810    x86 Peephole fix by Fermin.  Affects c-- and moby only.
2811    
2812    ----------------------------------------------------------------------
2813    Name: Matthias Blume
2814    Date: 2000/06/12 11:40:00
2815    Tag: blume-20000612-parmakefix
2816    Description:
2817    
2818    More cleanup after changing the file naming scheme: This time I
2819    repaired the parallel make mechanism for CMB.make which I broke earlier.
2820    
2821    ----------------------------------------------------------------------
2822    Name: Allen Leung
2823    Date: 2000/06/09 01:25:00
2824    Tag: leunga-20000609-various
2825    
2826    None of these things should affect normal SML/NJ operations
2827    
2828    1. Peephole improvements provided by Fermin (c--)
2829    2. New annotation DEFUSE for adding extra dependence (moby)
2830    3. New X86 LOCK instructions (moby)
2831    4. New machine description language for reservation tables (scheduling)
2832    5. Fixes to various optimization/analysis modules (branch chaining, dominator
2833       trees etc.)
2834    6. I've changed the CM files so that they can work with versions
2835       110.0.6, 110.25 and 110.28
2836    
2837    ----------------------------------------------------------------------
2838    Name: Matthias Blume
2839    Date: 2000/06/09 12:40:00
2840    Tag: blume-20000609-log
2841    Description:
2842    
2843    - Removed all(?) remaining RCS Log entries from sources.
2844    
2845    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
2846      anchored paths).
2847    
2848    ----------------------------------------------------------------------
2849    Name: Matthias Blume
2850    Date: 2000/06/07 17:00:00 JST
2851    Tag: blume-20000607-no-implicit-anchors
2852    Description:
2853    
2854    1. This update changes the default setting for
2855    CM.Control.implicit_anchors from true to false.  This means that
2856    implicit anchors are no longer permitted by default.  I also tried to
2857    make sure that nothing else still relies on implicit anchors.
2858    (This is the next step on the schedule towards a CM that does not even
2859    have the notion of implicit anchors anymore.)
2860    
2861    2. More CM manual updates.
2862    
2863    3. I managed to track down and fix the pickling bug I mentioned last
2864    time.  Because of the previously existing workaround, this entails no
2865    immediate practical changes.
2866    
2867    ----------------------------------------------------------------------
2868    Name: Matthias Blume
2869    Date: 2000/06/06 11:15:00 JST
2870    Tag: blume-20000606-lazierpickle
2871    Description:
2872    
2873    !!!! NEW BOOT FILES !!!!
2874    
2875    * The main purpose of this update is to make library pickles lazier in
2876    order to reduce the initial space penalty for autoloading a library.
2877    As a result, it is now possible to have $smlnj/compiler.cm
2878    pre-registered.  This should take care of the many complaints or
2879    inquiries about missing structure Compiler.  This required changes to
2880    CM's internal data structures and small tweaks to some algorithms.
2881    
2882    As a neat additional effect, it is no longer necessary (for the sake
2883    of lean heap image files) to distinguish between a "minimal" CM and a
2884    "full" CM.  Now, there is only one CM (i.e., the "full" version:
2885    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
2886    interactive top level. ($smlnj/cm/minimal.cm is gone.)
2887    
2888    To make the life of compiler-hackers easier, "makeml" now also
2889    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
2890    words, after you bootstrap a new sml for the first time, you will not
2891    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
2892    around you will still have to do it, though.)
2893    
2894    * A second change consists of major updates to the CM manual.  There
2895    are now several appendices with summary information and also a full
2896    specification of the CM description file syntax.
2897    
2898    * In directory src/system I added the script "allcross".  This script
2899    invokes sml and cross-compiles the compiler for all supported
2900    architectures.  (Useful when providing a new set of boot files.)
2901    
2902    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
2903    added a small tweak to pickle-util.sml to work around this problem,
2904    but it is not a proper fix yet.  I will investigate further.  (The
2905    effect of the bug was an inflation of library pickle size.)
2906    
2907    * Version number increased to 110.28.1 (to avoid compatibility problems).
2908    
2909    ----------------------------------------------------------------------
2910    Name: Allen Leung
2911    Date: 2000/05/25 17:28 EDT
2912    Tag: leunga-20000525-ra
2913    Description:
2914    
2915      Fixed a bug in freezing phase of the register allocator.
2916    
2917    ----------------------------------------------------------------------
2918    Name: Allen Leung
2919    Date: 2000/05/15 22:53 EDT
2920    Tag: leunga-20000515-alpha-x86-ra
2921    Description:
2922    
2923      1. Alpha
2924    
2925          Slight cleanup.  Removed the instruction SGNXL
2926    
2927      2. X86
2928    
2929          Added the following instructions to the instruction set:
2930    
2931            ROLx, RORx,
2932            BTx, BTSx, BTLx, BTRx,
2933            XCHGx, and variants with the LOCK prefix
2934    
2935      3. Register Allocation
2936    
2937          The module ra-rewrite-with-renaming has been improved.
2938    
2939      These have no effect on SML/NJ.
2940    
2941    ----------------------------------------------------------------------
2942    Name: Matthias Blume
2943    Date: 2000/05/15 16:20:00 JST
2944    Tag: blume-20000515-lightrebuild
2945    Description:
2946    
2947    1. I added an alternative to "-rebuild" to "makeml".  The difference is
2948       that prior to calling CMB.make' the CM-variable "LIGHT" will be
2949       defined.  In effect, the command will not build any cross-compiler
2950       backends and therefore finish more quickly.
2951    
2952       The "fixpt" script also takes a "-light" switch to be able to use
2953       this new facility while compiling for a fixpoint.
2954    
2955    2. I replaced all mentions of anchored paths in group owner specifications
2956       with simple relative paths (usually starting with "..").
2957       The rationale is that a library's internal workings should not be
2958       compromised by the lack of some anchor.  (An anchor is necessary
2959       for someone who wants to refer to the library by an anchored path,
2960       but it should not be necessary to build the same library in the first
2961       place.)
2962    
2963    3. I changed the way CM's tool mechanism determines the shell command
2964       string used for things like ml-yacc etc. so that it does not break
2965       when CM.Control.implicit_anchors is turned off.
2966    
2967    ----------------------------------------------------------------------
2968    Name: Matthias Blume
2969    Date: 2000/05/12 18:20:00 JST
2970    Tag: blume-20000512-ml-build
2971    Description:
2972    
2973    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
2974    getting installed properly (by config/install.sh).
2975    
2976    ----------------------------------------------------------------------
2977    Name: Matthias Blume
2978    Date: 2000/05/12 17:30:00 JST
2979    Tag: blume-20000512-anchors
2980    Description:
2981    
2982    !!! NEW BOOT FILES !!!
2983    
2984    This change is in preparation of fading out support for "implicitly
2985    anchored path names".  I went through all sources and used the
2986    explicit (and relatively new) $-notation.  See system/README and the
2987    CM manual for more info on this.
2988    
2989    I also modified the anchoring scheme for some things such as "smlnj",
2990    "MLRISC", "cm", etc. to take advantage of the fact that explicit
2991    anchors are more expressive: anchor name and first arc do not have to
2992    coincide.  This entails the following user-visible change:
2993    
2994    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
2995    particular, when you fire up sml with a command-line argument, say,
2996    e.g.:
2997    
2998       sml '$smlnj/cmb.cm'
2999    
3000    At the ML toplevel prompt:
3001    
3002       CM.autoload "$smlnj/cmb.cm";
3003    
3004    There is also a new controller in CM.Control that can be used to turn
3005    off all remaining support for implicit anchors by saying:
3006    
3007        CM.autoload "$smlnj/
3008        #set CM.Control.implicit_anchors false;
3009    
3010    This causes CM to reject implicitly anchored paths.  This is (for the
3011    time being) less permissive than the "final" version where there will
3012    be no more such implicit anchors and relative paths will be just that:
3013    relative.
3014    
3015    The next step (version after next version?) will be to make the
3016    default for CM.Control.implicit_anchors false.  After the dust has
3017    settled, I can then produce the "final" version of this...
3018    
3019    Note: Since bootstrapping is a bit tricky, I provided new boot files.
3020    
3021    ----------------------------------------------------------------------
3022    Name: Matthias Blume
3023    Date: 2000/05/11 16:30:00 JST
3024    Tag: blume-20000511-sources
3025    Description:
3026    
3027    The main change is that I added function CM.sources as a generalized
3028    version of the earlier CM.makedepend.  This entails the following
3029    additional changes:
3030    
3031      - CM.makedepend has been dropped.
3032    
3033      - CM manual has been updated.
3034    
3035      - TOOLS signature and API have been changed.
3036    
3037    ----------------------------------------------------------------------
3038    Name: Allen Leung
3039    Date: 2000/05/10 21:17 EDT
3040    Tag: leunga-20000510-moby-c--ssa
3041    Description:
3042    
3043      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
3044    None of these affect SML/NJ.
3045    
3046    1. Register Allocation
3047    
3048        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
3049           This module tries to remove local (i.e. basic block level) redundancies
3050           during spilling.
3051    
3052        b. A new framework for performing region based register allocation.
3053           Not yet entirely functional.
3054    
3055    2. X86
3056    
3057       a. DefUse for POP was missing the stack pointer [found by Lal]
3058       b. Reload for CALL was incorrect in X86Spill [found by John]
3059       c. Various fixes in X86Spill so that it can be used correctly for
3060          the new spilling module.
3061    
3062    3. SSA/IR
3063    
3064       a. New module ir/dj-dataflow.sml implements elimination based
3065          data flow analysis.
3066    
3067    4. MLRiscGen
3068    
3069       a. Fix for gc type annotation
3070    
3071    5. MDGen
3072    
3073       Various fixes for machine description -> ml code translation.  For ssa
3074       only.
3075    
3076    ----------------------------------------------------------------------
3077    Name: Allen Leung
3078    Date: 2000/05/08 22:17 EDT
3079    Tag: leunga-20000508-labexp
3080    Description:
3081    
3082      Fermin has found a few assembly problems with constant expressions
3083      generated in LabelExp.  Mostly, the problems involve extra parentheses,
3084      which choke on dumb assemblers.  This is his fix.
3085    
3086    ----------------------------------------------------------------------
3087    Name: Dave MacQueen
3088    Date: 2000/04/09 14:00 EDT
3089    Tag: dbm-20000502-Version_110_28
3090    Description:
3091    
3092    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
3093    
3094    2. Updated config/version to 110.28
3095    
3096    3. Updated config/srcarchiveurl
3097    
3098    3. New boot files!
3099       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
3100    
3101    ----------------------------------------------------------------------
3102    Name: Matthias Blume
3103    Date: 2000/05/01 19:05:00 JST
3104    Tag: blume-20000501-noweb
3105    Description:
3106    
3107    A new noweb tool has been added.  The existing system is entirely
3108    unaffected by this, but some CM users have asked for renewed noweb
3109    support.  Everything is documented in the CM manual.
3110    
3111    New (plugin) libraries:
3112    
3113       noweb-tool.cm
3114       nw-ext.cm
3115    
3116    ----------------------------------------------------------------------
3117    Name: Dave MacQueen
3118    Date: 2000/04/30 12:40PM EDT
3119    Tag: dbm-20000430-bug_fixes
3120    Description:
3121    
3122    1. Fix for bug 1498
3123       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
3124       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
3125         added toRealArray function
3126       smlnj/src/compiler/MiscUtil/print/ppobj.sml
3127         added check for tag Obj.RealArray to array printing case in ppObj
3128    
3129    2. Fix for bug 1510
3130       smlnj/src/compiler/Semant/types/typesutil.sml
3131         fixed definition of dummyargs (used by equalTycon) so that
3132         dummy args are distinct types
3133    
3134    ----------------------------------------------------------------------
3135    Name: Matthias Blume
3136    Date: 2000/04/30 01:00:00 JST
3137    Tag: blume-20000430-versions
3138    Description:
3139    
3140    1. CM version numbering added.  This is an implementation of Lal's
3141       proposal for adding version numbers and version checking to .cm
3142       files.  Lal said that his proposal was just that -- a proposal.
3143       For the time being I went ahead and implemented it so that people
3144       can comment on it.  Everything is completely backward-compatible
3145       (except for the stable library format, i.e., new bootfiles!).
3146    
3147       As usual, see the CM manual for details.
3148    
3149    2. An alternative syntax for anchored paths has been implemented.
3150       Dave has recently voiced the same concerns that I had when I did
3151       this, so there should be some support.  My take is that eventually
3152       I will let support for the current syntax (where anchors are
3153       "implicit") fade out in favor of the new, explicit syntax.
3154       In order to be backward-compatible, both old and new syntax are
3155       currently supported.
3156    
3157       Again, see the CM manual for details.
3158    
3159    3. Parallel make is trying to be slightly smarter:  When the master
3160       process finds a "bottleneck", i.e., when there is only one
3161       compilation unit that can be compiled and everybody else is
3162       waiting on it, then it will simply compile it directly instead
3163       of clumsily telling one of the slaves to do it.
3164    
3165    4. Support for "unsharing" added.  This is necessary in order to be
3166       able to have two different versions of the same library running
3167       at the same time (e.g., for trying out a new MLRISC while still
3168       having the old MLRISC linked into the current compiler, etc.)
3169       See the CM manual.
3170    
3171    5. Simple "makedepend" functionality added for generating Makefile
3172       dependency information.  (This is rather crude at the moment.
3173       Expect some changes here in the future.)
3174    
3175    6. ".fun" added as a recognized suffix for ML files. Also documented
3176       explicitly in the manual that the fallback behavior (unknown suffix
3177       -> ML file) is not an official feature!
3178    
3179    7. Small changes to the pickler for stable libraries.
3180    
3181    8. Several internal changes to CM (for cleanup/improvement).
3182    
3183    
3184    !!!! NEW BINFILES !!!!
3185    
3186    ----------------------------------------------------------------------
3187    Name: Matthias Blume
3188    Date: 2000/04/28 17:30:00 JST
3189    Tag: blume-20000428-pathconfig
3190    Description:
3191    
3192    1. I changed config/install.sh to remove duplicate entries from the
3193       lib/pathconfig file at the end.  Moreover, the final version of
3194       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
3195       in src/system/installml.
3196    
3197    2. The config/install.sh script now consistently uses relative
3198       pathnames in lib/pathconfig whenever the anchor is in the lib
3199       directory.  (So far this was true for the libraries that come
3200       pre-compiled and bundled as part of the bootfiles but not for
3201       libraries that are compiled by the script itself.)
3202    
3203    ----------------------------------------------------------------------
3204    Name: Matthias Blume
3205    Date: 2000/04/26 13:10:00 JST
3206    Tag: blume-20000426-fun_suffix
3207    Description:
3208    
3209    Added ".fun" as a recognized file name suffix (for ML code).
3210    
3211    ----------------------------------------------------------------------
3212    Name: Allen Leung
3213    Date: 2000/04/25 17:00:00 EST
3214    Tag: leunga-20000425-alpha-ra
3215    Description:
3216    
3217    1. Alpha
3218    
3219        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
3220    in C--.
3221    
3222    2. RA
3223    
3224       Added a flag "ra-dump-size" to print out the size of the flowgraph
3225       and the interference graph.
3226    
3227    ----------------------------------------------------------------------
3228    Name: Dave MacQueen
3229    Date: 2000/04/25/
3230    Tag: dbm-20000425-mlyacc_doc_examples
3231    Description:
3232      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
3233      Updated all three examples in src/ml-yacc/examples to run
3234      under 110.* using CM.make.
3235    
3236    ----------------------------------------------------------------------
3237    Name: Allen Leung
3238    Date: 2000/04/20 23:04:00 EST
3239    Tag: leunga-20000420-ssa-c---stuff
3240    Description:
3241    
3242      This update synchronizes my repository with Yale's.  Most of these
3243    changes, however, do not affect SML/NJ at all (the RA is an exception).
3244    
3245    1. Register Allocator
3246    
3247       a. An improvement in the interference graph construction:
3248          Given a copy
3249    
3250                s <- t
3251    
3252          no interference edge between s and t is added for this definition of s.
3253    
3254       b. I've added two new spill heuristic modules that Fermin and I developed
3255          (in the new library RA.cm). These are unused in SML/NJ but maybe
3256          useful for others (Moby?)
3257    
3258    2. X86
3259    
3260       a. Various fixes in the backend provided by Fermin [C--] and Lal.
3261    
3262    3. Alpha
3263    
3264       a. Added the BSR instruction and code generation that goes with it [C--]
3265       b. Other fixes too numerous to recount provided by Fermin [C--]
3266    
3267    4. Regmaps
3268    
3269       a. The regmaps are not initialized with the identity physical bindings
3270          at creation time.  This is unneeded.
3271    
3272    5. MLRISC Optimizations
3273    
3274       a. The DJ-Graph module can now compute the iterated dominance frontiers
3275          intersects with liveness incrementally in linear time! Woohoo!
3276          This is now used in my new SSA construction algorithm.
3277    
3278       b. THe branch reorganization module is now smarter about linear chains of
3279          basic blocks.
3280    
3281    
3282    ----------------------------------------------------------------------
3283    Name: Matthias Blume
3284    Date: 2000/04/12 13:52:00 JST
3285    Tag: blume_main_v110p27_1
3286    Description:
3287    
3288    Changed install.sh script to handle archive files without version number
3289    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
3290    name of the boot file archive.
3291    
3292    ----------------------------------------------------------------------
3293    Name: Dave MacQueen
3294    Date: 2000/04/09 14:00 EDT
3295    Tag: dbm-20000410-Version_110_27
3296    Description:
3297    
3298    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
3299    
3300    2. Updated src/config/version to 110.27
3301    
3302    3. New boot files!
3303    
3304    ----------------------------------------------------------------------
3305    Name: Allen Leung
3306    Date: 2000/04/09 19:09:00 EST
3307    Tag: leunga-20000409-misc
3308    Description:
3309    
3310    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
3311    
3312    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
3313    
3314    ----------------------------------------------------------------------
3315    Name: Stefan
3316    Date: 2000/04/07 10:00:00 EDT
3317    Tag: monnier-20000406-branch-handling
3318    Description:
3319    
3320    Improved handling of branches (mostly those generated from
3321    polymorphic equality), removed switchoff and changed the
3322    default optimization settings (more cpsopt and less flintopt).
3323    
3324    ----------------------------------------------------------------------
3325    Name: Allen Leung
3326    Date: 2000/04/06 01:30:00 EST
3327    Tag: leunga-20000406-peephole-x86-SSA-2
3328    Description:
3329    
3330       Forgot a few files.
3331    
3332    ----------------------------------------------------------------------
3333    Name: Allen Leung
3334    Date: 2000/04/06 00:36:00 EST
3335    Tag: leunga-20000406-peephole-x86-SSA
3336    Description:
3337    
3338    1.  New Peephole code
3339    
3340    2.  Minor improvement to X86 instruction selection
3341    
3342    3.  Various fixes to SSA and machine description -> code translator
3343    
3344    ----------------------------------------------------------------------
3345    Name: Matthias Blume
3346    Date: 2000/04/05 12:30:00 JST
3347    Tag: blume_main_v110p26p2_3
3348    Description:
3349    
3350    This update just merges three minor cosmetic updates to CM's sources
3351    to get ready for the 110.27 code freeze on Friday.  No functionality
3352    has changed.
3353    
3354    ----------------------------------------------------------------------
3355    Name: Allen Leung
3356    Date: 2000/04/04 19:39:00 EST
3357    Tag: leunga-20000404-x86-asm
3358    Description:
3359    
3360    1.  Fixed a problem in X86 assembly.
3361    
3362        Things like
3363    
3364           jmp %eax
3365           jmp (%eax)
3366    
3367        should be output as
3368    
3369           jmp *%eax
3370           jmp *(%eax)
3371    
3372    2.  Assembly output
3373    
3374          Added a new flag
3375    
3376              "asm-indent-copies" (default to false)
3377    
3378          When this flag is on, parallel copies will be indented an extra level.
3379    
3380    ----------------------------------------------------------------------
3381    Name: Allen Leung
3382    Date: 2000/04/04 03:18:00 EST
3383    Tag: leunga-20000404-C--Moby
3384    Description:
3385    
3386        All of these fixes are related to C--, Moby, and my own optimization
3387        stuff; so they shouldn't affect SML/NJ.
3388    
3389    1.  X86
3390    
3391        Various fixes related floating point, and extensions.
3392    
3393    2.  Alpha
3394    
3395        Some extra patterns related to loads with signed/zero extension
3396        provided by Fermin.
3397    
3398    3.  Assembly
3399    
3400        When generating assembly, resolve the value of client defined constants,
3401        instead of generating symbolic values.  This is controlled by the
3402        new flag "asm-resolve-constants", which is default to true.
3403    
3404    4.  Machine Descriptions
3405    
3406        a. The precedence parser was slightly broken when parsing infixr symbols.
3407        b. The type generalizing code had the bound variables reversed, resulting
3408           in a problem during arity raising.
3409        c. Various fixes in machine descriptions.
3410    
3411    ----------------------------------------------------------------------
3412    Name: Matthias Blume
3413    Date: 2000/04/03 16:05:00 JST
3414    Tag: blume_main_v110p26p2_2
3415    Description:
3416    
3417    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
3418    now done via the ordinary static environment that is context to each
3419    compilation unit.
3420    
3421    To this end, I arranged that instead of "structure Core" as "structure
3422    _Core" is bound in the pervasive environment.  Core access is done via
3423    _Core (which can never be accidentally rebound because _Core is not a
3424    legal surface-syntax symbol).
3425    
3426    The current solution is much cleaner because the core environment is
3427    now simply part of the pervasive environment which is part of every
3428    compilation unit's context anyway.  In particular, this eliminates all
3429    special-case handling that was necessary until now in order to deal
3430    with dynamic and symbolic parts of the core environment.
3431    
3432    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
3433    compilation manager's bootstrap compiler (actually: in the "init group"
3434    handling).  See the comments in src/system/smlnj/init/init.cmi for
3435    more details.
3436    
3437    I also tried to track down all mentions of "Core" (as string argument
3438    to Symbol.strSymbol) in the compiler and replaced them with a
3439    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
3440    name appears in one place only.
3441    
3442    Binfile and bootfile format have not changed, but the switchover from
3443    the old "init.cmi" to the new one is a bit tricky, so I supplied new
3444    bootfiles anyway.
3445    
3446    ----------------------------------------------------------------------
3447    Name: Allen Leung
3448    Date: 2000/04/02 21:17:00 EST
3449    Tag: leunga-20000402-mltree
3450    Description:
3451    
3452       1. Renamed the constructor CALL in MLTREE by popular demand.
3453       2. Added a bunch of files from my repository.  These are currently
3454          used by other non-SMLNJ backends.
3455    
3456    ----------------------------------------------------------------------
3457    Name: Allen Leung
3458    Date: 2000/03/31 21:15:00 EST
3459    Tag: leunga-20000331-aliasing
3460    Description:
3461    
3462    This update contains a rewritten (and hopefully more correct) module
3463    for extracting aliasing information from CPS.
3464    
3465       To turn on this feature:
3466    
3467            Compiler.Control.CG.memDisambiguate := true
3468    
3469       To pretty print the region information with assembly
3470    
3471           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
3472    
3473       To control how many levels of aliasing information are printed, use:
3474    
3475           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
3476    
3477       The default of n is 3.
3478    
3479    ----------------------------------------------------------------------
3480    Name: David MacQueen
3481    Date: 2000/03/31 11:15:00 EST
3482    Tag: dbm-20000331-runtime_fix
3483    Description:
3484    
3485    This update contains:
3486    
3487    1. runtime/c-lib/c-libraries.c
3488       includes added in revision 1.2 caused compilation errors on hppa-hpux
3489    
3490    2. fix for bug 1556
3491       system/Basis/Implementation/NJ/internal-signals.sml
3492    
3493    ----------------------------------------------------------------------
3494    Name: Matthias Blume
3495    Date: 2000/03/31 18:00:00 JST
3496    Tag: blume_main_v110p26p2_1
3497    Description:
3498    
3499    This update contains:
3500    
3501    1. A small change to CM's handling of stable libraries:
3502       CM now maintains one "global" modmap that is used for all stable
3503       libraries.  The use of such a global modmap maximizes sharing and
3504       minimizes the need for re-traversing parts of environments during
3505       modmap construction.  (However, this has minor impact since modmap
3506       construction seems to account for just one percent or less of total
3507       compile time.)
3508    
3509    2. I added a "genmap" phase to the statistics.  This is where I got the
3510       "one percent" number (see above).
3511    
3512    3. CM's new tool parameter mechanism just became _even_ better. :)
3513       - The parser understands named parameters and recursive options.
3514       - The "make" and "shell" tools use these new features.
3515         (This makes it a lot easier to cascade these tools.)
3516       - There is a small syntax change: named parameters use a
3517    
3518           <name> : ( <option> ... )            or
3519           <name> : <string>
3520    
3521         syntax.  Previously, named parameters were implemented in an
3522         ad-hoc fashion by each tool individually (by parsing strings)
3523         and had the form
3524    
3525           <name>=<string>
3526    
3527       See the CM manual for a full description of these issues.
3528    
3529    ----------------------------------------------------------------------
3530    Name: Matthias Blume
3531    Date: 2000/03/30 18:00:00 JST
3532    Tag: blume_main_v110p26p2_0
3533    Description:
3534    
3535    !!!!! WARNING !!!!!!
3536    !!  New binfiles  !!
3537    !!!!!!!!!!!!!!!!!!!!
3538    
3539    This update contains:
3540    
3541    1. Moderate changes to CM:
3542    
3543       - Changes to CM's tools mechanism.  In particular, it is now possible
3544       to have tools that accept additional "command line" parameters
3545       (specified in the .cm file at each instance where the tool's class is
3546       used).
3547    
3548       This was done to accommodate the new "make" and "shell" tools which
3549       facilitate fairly seamless hookup to portions of code managed using
3550       Makefiles or Shell scripts.
3551    
3552       There are no classes "shared" or "private" anymore.  Instead, the
3553       sharing annotation is now a parameter to the "sml" class.
3554    
3555       There is a bit of generic machinery for implementing one's own
3556       tools that accept command-line parameters.  However, I am not yet fully
3557       satisfied with that part, so expect changes here in the future.
3558    
3559       All existing tools are described in the CM manual.
3560    
3561       - Slightly better error handling.  (CM now suppresses many followup
3562       error messages that tended to be more annoying than helpful.)
3563    
3564    2. Major changes to the compiler's static environment data structures.
3565    
3566       - no CMStaticEnv anymore.
3567            - no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
3568              but it is called Environment), no conversions between different
3569              kinds of static environments
3570    
3571       - There is still a notion of a "modmap", but such modmaps are generated
3572         on demand at the time when they are needed.  This sounds slow, but I
3573         sped up the code that generates modmaps enough for this not to lead to
3574         a slowdown of the compiler (at least I didn't detect any).
3575    
3576       - To facilitate rapid modmap generation, static environments now
3577         contain an (optional) "modtree" structure.  Modtree annotations are
3578         constructed by the unpickler during unpickling.  (This means that
3579         the elaborator does not have to worry about modtrees at all.)
3580         Modtrees have the advantage that they are compositional in the same
3581         way as the environment data structure itself is compositional.
3582         As a result, modtrees never hang on to parts of an environment that
3583         has already been rendered "stale" by filtering or rebinding.
3584    
3585       - I went through many, many trials and errors before arriving at the
3586         current solution.  (The initial idea of "linkpaths" did not work.)
3587         But the result of all this is that I have touched a lot of files that
3588         depend on the "modules" and "types" data structures (most of the
3589         elaborator). There were a lot of changes during my "linkpath" trials
3590         that could have been reverted to their original state but weren't.
3591         Please, don't be too harsh on me for messing with this code a bit more
3592         than what was strictly necessary...  (I _did_ resist the temptation
3593         of doing any "global reformatting" to avoid an untimely death at
3594         Dave's hands. :)
3595    
3596       - One positive aspect of the previous point:  At least I made sure that
3597         all files that I touched now compile without warnings (other than
3598         "polyEqual").
3599    
3600       - compiler now tends to run "leaner" (i.e., ties up less memory in
3601         redundant modmaps)
3602    
3603    ----------------------------------------------------------------------
3604    Name: Allen Leung
3605    Date: 2000/03/29 18:00:00
3606    Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
3607    Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
3608    Description:
3609    
3610       This update contains *MAJOR* changes to the way code is generated from CPS
3611    in the module mlriscGen, and in various backend modules.
3612    
3613    CHANGES
3614    =======
3615    
3616    1. MLRiscGen: forward propagation fix.
3617    
3618       There was a bug in forward propagation introduced at about the same time
3619       as the MLRISC x86 backend, which prohibits coalescing to be
3620       performed effectively in loops.
3621    
3622       Effect: speed up of loops in RISC architectures.
3623               By itself, this actually slowed down certain benchmarks on the x86.
3624    
3625    2. MLRiscGen:  forward propagating addresses from consing.
3626    
3627       I've changed the way consing code is generated.  Basically I separated
3628       out the initialization part:
3629    
3630            store tag,   offset(allocptr)
3631            store elem1, offset+4(allocptr)
3632            store elem2, offset+8(allocptr)
3633            ...
3634            store elemn, offset+4n(allocptr)
3635