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 670, Sun Jun 18 13:10:57 2000 UTC revision 1066, Thu Feb 14 16:50:02 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: 2002/02/14 11:50:00 EST
18    Tag: blume-20020214-quote
19    Description:
20    
21    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
22    because of the new and improved tag generation scheme.  (Thanks to Allen
23    Leung for pointing it out.)
24    
25    ----------------------------------------------------------------------
26    Name: Lal George
27    Date: 2002/02/14 09:55:27 EST 2002
28    Tag: george-20020214-isabelle-bug
29    Description:
30    
31    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
32    of Isabelle on the x86.
33    
34    From Allen:
35    -----------
36     I've found the problem:
37    
38         in ra-core.sml, I use the counter "blocked" to keep track of the
39         true number of elements in the freeze queue.  When the counter goes
40         to zero, I skip examining the queue.  But I've messed up the
41         bookkeeping in combine():
42    
43             else ();
44             case !ucol of
45               PSEUDO => (if !cntv > 0 then
46                     (if !cntu > 0 then blocked := !blocked - 1 else ();
47                                        ^^^^^^^^^^^^^^^^^^^^^^^
48                      moveu := mergeMoveList(!movev, !moveu)
49                     )
50                  else ();
51    
52         combine() is called to coalesce two nodes u and v.
53         I think I was thinking that if the move counts of u and v are both
54         greater than zero then after they are coalesced then one node is
55         removed from the freeze queue.  Apparently I was thinking that
56         both u and v are of low degree, but that's clearly not necessarily true.
57    
58    
59    02/12/2002:
60        Here's the patch.  HOL now compiles.
61    
62        I don't know how this impact on performance (compile
63        time or runtime).  This bug caused the RA (especially on the x86)
64        to go thru the potential spill phase when there are still nodes on the
65        freeze queue.
66    
67    
68    
69    
70    ----------------------------------------------------------------------
71    Name: Matthias Blume
72    Date: 2002/02/13 22:40:00 EST
73    Tag: blume-20020213-fptr-rtti
74    Description:
75    
76    Fixed a bug in ml-nlffigen that was introduced with one of the previous
77    updates.
78    
79    ----------------------------------------------------------------------
80    Name: Matthias Blume
81    Date: 2002/02/13 16:41:00 EST
82    Tag: blume-20020213-cmlpq
83    Description:
84    
85    Added new priority queue export symbols (which have just been added to
86    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
87    not compile and the installer would choke.)
88    
89    ----------------------------------------------------------------------
90    Name: Matthias Blume
91    Date: 2002/02/13 16:15:00 EST
92    Tag: blume-20020213-various
93    Description:
94    
95    1. More tweaks to ml-nlffigen:
96    
97       - better internal datastructures (resulting in slight speedup)
98       - "-match" option requires exact match
99       - "localized" gensym counters (untagged structs/unions nested within
100         other structs/unions or within typedefs get a fresh counter; their
101         tag will be prefixed by a concatenation of their parents' tags)
102       - bug fixes (related to calculation of transitive closure of types
103         to be included in the output)
104    
105    2. Minor Basis updates:
106    
107       - added implementations for List.collate and Option.app
108    
109    ----------------------------------------------------------------------
110    Name: Matthias Blume
111    Date: 2002/02/11 15:55:00 EST
112    Tag: blume-20020211-gensym
113    Description:
114    
115    Added a "-gensym" option to command line of ml-nlffigen.  This can be
116    used to specify a "stem" -- a string that is inserted in all "gensym'd"
117    names (ML structure names that correspond to unnamed C structs, unions,
118    and enums), so that separate runs of ml-nlffigen do not clash.
119    
120    ----------------------------------------------------------------------
121    Name: Matthias Blume
122    Date: 2002/02/11 12:05:00 EST
123    Tag: blume-20020211-gensml
124    Description:
125    
126    A quick fix for a problem with GenSML (in the pgraph-util library):
127    Make generation of toplevel "local" optional.  (Strictly speaking,
128    signature definitions within "local" are not legal SML.)
129    
130    Other than that: updates to INSTALL and cm/TODO.
131    
132    ----------------------------------------------------------------------
133    Name: Matthias Blume
134    Date: 2002/02/08 15:00:00 EST
135    Tag: blume-20020208-uniquepid
136    Description:
137    
138    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
139    
140    1. The installer (config/install.sh) has gotten smarter:
141    
142         - Configuration options are a bit easier to specify now
143           (in config/targets).
144         - Bug in recognizing .tar.bz2 files fixed.
145         - Installer automatically resolves dependencies between
146           configuration options (e.g., if you ask for eXene, you will
147           also get cml -- regardless whether you asked for it or not).
148         - Installer can run in "quieter mode" by setting the environment
149           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
150           "completely silent", though.
151         - Build HashCons library as part of smlnj-lib.
152    
153    2. A new scheme for assigning persistent identifiers to compilation
154       units (and, by extension, to types etc.) has been put into place.
155       This fixes a long-standing bug where types and even dynamic values
156       can get internally confused, thereby compromising type safety
157       (abstraction) and dynamic correctness.  See
158    
159         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
160    
161       for an example of how things could go wrong until now.
162    
163       The downside of the new scheme is that pids are not quite as
164       persistent as they used to be: CM will generate a fresh pid
165       for every compilation unit that it thinks it sees for the first
166       time.  That means that if you compile starting from a clean, fresh
167       source tree at two different times, you end up with different
168       binaries.
169    
170       Cutoff recompilation, however, has not been compromised because
171       CM keeps pid information in special caches between runs.
172    
173    ----------------------------------------------------------------------
174    Name: Lal George
175    Date: 2002/02/07 15:34:13 EST 2002
176    Tag: <none>
177    Description:
178    
179    Compilers that generate assembly code may produce  global labels
180    whose value is resolved at link time. The various peephole optimization
181    modules did not take this in account.
182    
183    TODO. The Labels.addrOf function should really return an option
184    type so that clients are forced to deal with this issue, rather
185    than an exception being raised.
186    
187    ----------------------------------------------------------------------
188    Name: Lal George
189    Date: 2002/02/06 13:55:02 EST
190    Tag: george-20020206-ra-breakup
191    Description:
192    
193    1. A bug fix from Allen.
194    
195        A typo causes extra fstp %st(0)'s to be generated at compensation
196        edges, which might cause stack underflow traps at runtime.  This
197        occurs in fft where there are extraneous fstps right before the 'into'
198        trap instruction (in this case they are harmless since none of the
199        integers overflow.)
200    
201    2. Pulled out various utility modules that were embedded in the modules
202       of the register allocator. I need these modules for other purposes, but
203       they are not complete enough to put into a library (just yet).
204    ----------------------------------------------------------------------
205    Name: Matthias Blume
206    Date: 2002/01/31 16:05:00 EST
207    Tag: blume-20020131-sparc-ccalls
208    Description:
209    
210    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
211       of extra stack space by mistake.  Fixed.
212    
213    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
214    
215    ----------------------------------------------------------------------
216    Name: Allen Leung
217    Date: 2002/01/30
218    Tag: leunga-20020130-nowhere-bug-fix
219    Description:
220    
221       MLRISC bug fixes:
222       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
223       2. MachineInt.fromString was negating its value.
224    
225    ----------------------------------------------------------------------
226    Name: Matthias Blume
227    Date: 2002/01/29
228    Tag: blume-20020129-INSTALL
229    Description:
230    
231    - Added somewhat detailed installation instructions (file INSTALL).
232    - Fixed curl-detection bug in config/install.sh.
233    - It is now possible to select the URL getter using the URLGETTER
234      environment variable:
235    
236          not set / "unknown"      --> automatic detection (script tries wget,
237                                       curl, and lynx)
238          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
239                                       how to properly invoke them)
240          other                    --> use $URLGETTER directly, it must take
241                                       precisely two command-line arguments
242                                       (source URL and destination file name)
243    
244    ----------------------------------------------------------------------
245    Name: Matthias Blume
246    Date: 2002/01/28
247    Tag: blume-20020128-sparc-ccalls
248    Description:
249    
250    - Fixed problem with calculation of "used" registers in sparc-c-calls.
251    - Make use of the allocParam argument in sparc-c-calls.
252    
253    ----------------------------------------------------------------------
254    Name: Matthias Blume
255    Date: 2002/01/28
256    Tag: blume-20020128-allocParam
257    Description:
258    
259    John Reppy:  Changes c-calls API to accept client-callback for
260    allocating extra stack space.
261    me: Corresponding changes to mlriscGen (using a dummy argument that
262        does not change the current behavior).
263    
264    ----------------------------------------------------------------------
265    Name: Matthias Blume
266    Date: 2002/01/28 12:00:00
267    Tag: Release_110_38
268    Description:
269    
270    This time for real!!!
271    
272    ----------------------------------------------------------------------
273    Name: Matthias Blume
274    Date: 2002/01/28 10:56:00 EST
275    Tag: blume-20020128-retraction
276    Description:
277    
278    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
279       with blume-Release_110_38-retracted.)
280    
281    1. Fixed a problem with incorrect rounding modes in real64.sml.
282       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
283    
284    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
285       and enums fixed.  The naming of corresponding ML identifiers should
286       now be consistent again.
287    
288    ----------------------------------------------------------------------
289    Name: Allen Leung
290    Date: 2002/01/27
291    Tag: leunga-20020127-nowhere
292    Description:
293    
294       Added a target called nowhere in the configuration scripts.
295       Enabling this will build the MLRISC 'nowhere' tool (for translating
296       programs with where-clauses into legal SML code) during installation.
297    
298    ----------------------------------------------------------------------
299    Name: Matthias Blume
300    Date: 2002/01/25 21:27:00 EST
301    Tag: blume-Release_110_38-retracted
302    Description:
303    
304    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
305    
306    README will be added later.
307    
308    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
309    (Release_110_38) removed.  Reason: Last-minute bug fixes.
310    
311    ----------------------------------------------------------------------
312    Name: Matthias Blume
313    Date: 2002/01/25
314    Tag: blume-20020125-ffi
315    Description:
316    
317    A large number of tweaks and improvements to ml-nlffi-lib and
318    ml-nlffigen:
319    
320       - ML represenation types have been streamlined
321       - getter and setter functions work with concrete values, not abstract
322         ones where possible
323       - ml-nlffigen command line more flexible (see README file there)
324       - some bugs have been fixed (hopefully)
325    
326    ----------------------------------------------------------------------
327    Name: Lal George
328    Date: 2002/01/24
329    Tag: george-20020124-risc-ra-interface
330    Description:
331    
332       There is a dramatic simplification in the interface to the
333       register allocator for RISC architectures as a result of making
334       parallel copy instructions explicit.
335    
336    ----------------------------------------------------------------------
337    Name: Matthias Blume
338    Date: 2002/01/22
339    Tag: blume-20020122-x86-ccalls
340    Description:
341    
342    Bug fix for c-calls on x86 (having to do with how char- and
343    short-arguments are being handled).
344    
345    ----------------------------------------------------------------------
346    Name: Matthias Blume
347    Date: 2002/01/21
348    Tag: blume-20020121-ff
349    Description:
350    
351    Another day of fiddling with the FFI...
352    
353    1. Bug fix/workaround:  CKIT does not complain about negative array
354       dimensions, so ml-nlffigen has to guard itself against this possibility.
355       (Otherwise a negative dimension would send it into an infinite loop.)
356    
357    2. Some of the abstract types (light objects, light pointers, most "base"
358       types) in structure C are now eqtypes.
359    
360    3. Added constructors and test functions for NULL function pointers.
361    
362    ----------------------------------------------------------------------
363    Name: Matthias Blume
364    Date: 2002/01/18
365    Tag: blume-20020118-ready-for-new-release
366    Description:
367    
368    Made config/srcarchiveurl point to a new place.  (Will provide boot
369    files shortly.)
370    
371    Maybe we christen this to be 110.38?
372    
373    ----------------------------------------------------------------------
374    Name: Matthias Blume
375    Date: 2002/01/18
376    Tag: blume-20020118-more-ffifiddle
377    Description:
378    
379    Today's FFI fiddling:
380    
381      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
382        C.Set.* that use concrete (MLRep.*) arguments and results instead
383        of abstract ones.
384    
385      - Provided word-style bit operations etc. for "int" representation
386        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
387        and Long.
388    
389    ----------------------------------------------------------------------
390    Name: Matthias Blume
391    Date: 2002/01/18
392    Tag: blume-20020118-use-x86-fp
393    Description:
394    
395    Now that x86-fast-fp seems to be working, I turned it back on again
396    by default.  (Seems to work fine now, even with the FFI.)
397    
398    Other than that, I added some documentation about the FFI to
399    src/ml-nlffigen/README and updated the FFI test examples in
400    src/ml-nlffi-lib/Tests/*.
401    
402    ----------------------------------------------------------------------
403    Name: Allen Leung
404    Date: 2002/01/17
405    Tag: leunga-20020117-x86-fast-fp-call
406    Description:
407    
408       1. Fixed a problem with handling return fp values when x86's fast fp
409          mode is turned on.
410    
411       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
412          of %f32.
413    
414       3. Added a constructor INT32lit to the ast of MLRISC tools.
415    
416    ----------------------------------------------------------------------
417    Name: Matthias Blume
418    Date: 2002/01/16
419    Tag: blume-20020116-ffifiddle
420    Description:
421    
422    More fiddling with the FFI interface:
423    
424     - Make constness 'c instead of rw wherever possible.  This eliminates
425       the need for certain explicit coercions.  (However, due to ML's
426       value polymorphism, there will still be many cases where explicit
427       coercions are necessary.  Phantom types are not the whole answer
428       to modeling a subtyping relationship in ML.)
429    
430     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
431    
432     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
433       clumsy workaround for varargs problem.)
434    
435    ----------------------------------------------------------------------
436    Name: Lal George
437    Date: 2002/01/15
438    Tag: <none>
439    Description:
440    
441    1. Since COPY instructions are no longer native to the architecture,
442       a generic functor can be used to implement the expandCopies function.
443    
444    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
445       TEXT segment.
446    
447    ----------------------------------------------------------------------
448    Name: Matthias Blume
449    Date: 2002/01/15
450    Tag: blume-20020115-ffiupdates
451    Description:
452    
453    1. Fix for bug resulting in single-precision float values being returned
454       incorrectly from FFI calls.
455    
456    2. Small modifications to C FFI API:
457    
458        - memory-allocation routines return straight objects (no options)
459          and raise an exception in out-of-memory situations
460        - unsafe extensions to cast between function pointers and pointers
461          from/to ints
462        - added structure C_Debug as an alternative to structure C where
463          pointer-dereferencing (|*| and |*!) always check for null-pointers
464        - added open_lib' to DynLinkage;  open_lib' works like open_lib
465          but also takes a (possibly empty) list of existing library handles
466          that the current library depends on
467    
468    ----------------------------------------------------------------------
469    Name: Matthias Blume
470    Date: 2002/01/10
471    Tag: blume-20020110-newffigen
472    Description:
473    
474    1. Updates to portable graph code.
475    
476    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
477       more scalable now so that even huge interfaces such as the one
478       for GTK compile in finite time and space. :-)
479       See src/ml-nlffigen/README for details on what's new.
480    
481    ----------------------------------------------------------------------
482    Name: Lal George
483    Date: 2001/01/09 14:31:35 EST 2002
484    Tag: george-20011206-rm-native-copy
485    Description:
486    
487            Removed the native COPY and FCOPY instructions
488            from all the architectures and replaced it with the
489            explicit COPY instruction from the previous commit.
490    
491            It is now possible to simplify many of the optimizations
492            modules that manipulate copies. This has not been
493            done in this change.
494    
495    ----------------------------------------------------------------------
496    Name: Lal George
497    Date: 2001/12/06 16:50:13 EST 2001
498    Tag: george-20011206-mlrisc-instruction
499    Description:
500    
501    Changed the representation of instructions from being fully abstract
502    to being partially concrete. That is to say:
503    
504      from
505            type instruction
506    
507      to
508            type instr                              (* machine instruction *)
509    
510            datatype instruction =
511                LIVE of {regs: C.cellset, spilled: C.cellset}
512              | KILL of {regs: C.cellset, spilled: C.cellset}
513              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
514              | ANNOTATION of {i: instruction, a: Annotations.annotation}
515              | INSTR of instr
516    
517    This makes the handling of certain special instructions that appear on
518    all architectures easier and uniform.
519    
520    LIVE and KILL say that a list of registers are live or killed at the
521    program point where they appear. No spill code is generated when an
522    element of the 'regs' field is spilled, but the register is moved to
523    the 'spilled' (which is present, more for debugging than anything else).
524    
525    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
526    We used to generate:
527    
528            DEFFREG f1
529            f1 := f2 + f3
530            trapb
531    
532    but now generate:
533    
534            f1 := f2 + f3
535            trapb
536            LIVE {regs=[f1,f2,f3], spilled=[]}
537    
538    Furthermore, the DEFFREG (hack) required that all floating point instruction
539    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
540    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
541    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
542    this architecture).
543    
544    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
545    all the architectures. This will result in further simplification of the
546    register allocator that must be aware of them for coalescing purposes, and
547    will also simplify certain aspects of the machine description that provides
548    callbacks related to parallel copies.
549    
550    ANNOTATION should be obvious, and now INSTR represents the honest to God
551    machine instruction set!
552    
553    The <arch>/instructions/<arch>Instr.sml files define certain utility
554    functions for making porting easier -- essentially converting upper case
555    to lower case. All machine instructions (of type instr) are in upper case,
556    and the lower case form generates an MLRISC instruction. For example on
557    the alpha we have:
558    
559      datatype instr =
560         LDA of {r:cell, b:cell, d:operand}
561       | ...
562    
563      val lda : {r:cell, b:cell, d:operand} -> instruction
564        ...
565    
566    where lda is just (INSTR o LDA), etc.
567    
568    ----------------------------------------------------------------------
569    Name: Matthias Blume
570    Date: 2001/11/22 21:40:00 EST
571    Tag: Release_110_37
572    Description:
573    
574    Release 110.37.  This time for real.
575    
576    ----------------------------------------------------------------------
577    Name: Matthias Blume
578    Date: 2001/11/21 16:35:00 EST
579    Tag: blume-20011121-foot-in-mouth
580    Description:
581    
582    Removed the "Release_110_37" tag because of a serious bug.
583    This will be re-tagged once the bug is fixed.
584    
585    ----------------------------------------------------------------------
586    Name: Matthias Blume
587    Date: 2001/11/21 16:14:00 EST
588    Tag: blume-20011121-forgottenfile
589    Description:
590    
591    Forgot to add a file.  (Just a .tex-file -- part of
592    the CM manual source.)
593    
594    ----------------------------------------------------------------------
595    Name: Matthias Blume
596    Date: 2001/11/21 16:10:00 EST
597    Tag: blume-20011121-invalid_110_37
598    Description:
599    
600    Note: I removed the original tag "Release_110_37" from this commit
601          because we found a serious bug in all non-x86 backends.
602          - Matthias
603    
604    1. Modifications to the SML/NJ code generator and to the runtime system
605       so that code object name strings are directly inserted into code
606       objects at code generation time.  The only business the runtime system
607       has with this is now to read the name strings on occasions.
608       (The encoding of the name string has also changed somewhat.)
609    
610    2. CM now implements a simple "set calculus" for specifying export lists.
611       In particular, it is now possible to refer to the export lists of
612       other libraries/groups/sources and form unions as well as differences.
613       See the latest CM manual for details.
614    
615    3. An separate notion of "proxy" libraries has again be eliminated from
616       CM's model.  (Proxy libraries are now simply a special case of using
617       the export list calculus.)
618    
619    4. Some of the existing libraries now take advantage of the new set
620       calculus.
621       (Notice that not all libraries have been converted because some
622       of the existing .cm-files are supposed to be backward compatible
623       with 110.0.x.)
624    
625    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
626       "General.exnMessage"!  The former relies on a certain hook to be
627       initialized, and that often does not happen in the stand-alone case.)
628    
629    ----------------------------------------------------------------------
630    Name: Lal George
631    Date: 2001/11/21  13:56:18 EST
632    Tag: george-2001121-pseudo-ops
633    Description:
634    
635      Implemented a complete redesign of MLRISC pseudo-ops. Now there
636      ought to never be any question of incompatabilities with
637      pseudo-op syntax expected by host assemblers.
638    
639      For now, only modules supporting GAS syntax are implemented
640      but more should follow, such as MASM, and vendor assembler
641      syntax, e.g. IBM as, Sun as, etc.
642    
643    ----------------------------------------------------------------------
644    Name: Matthias Blume
645    Date: 2001/11/14 11:52:00 EST
646    Tag: blume-20011114-srcname
647    Description:
648    
649    1. Routed the name of the current source file to mlriscgen where it
650       should be directly emitted into the code object.  (This last part
651       is yet to be done.)
652    
653    2. Some cleanup of the pgraph code to make it match the proposal that
654       I put out the other day.  (The proposal notwithstanding, things are
655       still in flux here.)
656    
657    ----------------------------------------------------------------------
658    Name: Lal George
659    Date: 2001/11/14 09:44:04 EST
660    Tag:
661    Description:
662    
663      Fix for a backpatching bug reported by Allen.
664    
665      Because the boundary between short and long span-dependent
666      instructions is +/- 128, there are an astounding number of
667      span-dependent instructions whose size is over estimated.
668    
669      Allen came up with the idea of letting the size of span
670      dependent instructions be non-monotonic, for a maxIter
671      number of times, after which the size must be monotonically
672      increasing.
673    
674      This table shows the number of span-dependent instructions
675      whose size was over-estimated as a function of maxIter, for the
676      file Parse/parse/ml.grm.sml:
677    
678         maxIter            # of instructions:
679            10                      687
680            20                      438
681            30                      198
682            40                        0
683    
684      In compiling the compiler, there is no significant difference in
685      compilation speed between maxIter=10 and maxIter=40. Actually,
686      my measurements showed that maxIter=40 was a tad faster than
687      maxIter=10! Also 96% of the  files in the compiler reach a fix
688      point within 13 iterations, so fixing maxIter at 40, while high,
689      is okay.
690    
691    ----------------------------------------------------------------------
692    Name: Matthias Blume
693    Date: 2001/10/31 15:25:00 EST
694    Tag: blume-20011031-pgraph
695    Description:
696    
697    CKIT:
698    * Changed the "Function" constructor of type Ast.ctype to carry optional
699      argument identifiers.
700    * Changed the return type of TypeUtil.getFunction accordingly.
701    * Type equality ignores the argument names.
702    * TypeUtil.composite tries to preserve argument names but gives up quickly
703      if there is a mismatch.
704    
705    installation script:
706    * attempts to use "curl" if available (unless "wget" is available as well)
707    
708    CM:
709    * has an experimental implementation of "portable graphs" which I will
710      soon propose as an implementation-independent library format
711    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
712    
713    NLFFI-LIB:
714    * some cleanup (all cosmetic)
715    
716    NLFFIGEN:
717    * temporarily disabled the mechanism that suppresses ML output for
718      C definitions whose identifiers start with an underscore character
719    * generate val bindings for enum constants
720    * user can request that only one style (light or heavy) is being used;
721      default is to use both (command-line arguments: -heavy and -light)
722    * fixed bug in handling of function types involving incomplete pointers
723    * generate ML entry points that take record arguments (i.e., using
724      named arguments) for C functions that have a prototype with named
725      arguments
726      (see changes to CKIT)
727    
728    ----------------------------------------------------------------------
729    Name: Allen Leung
730    Date: 2001/10/27 20:34:00 EDT
731    Tag: leunga-20011027-x86-fast-fp-call
732    Description:
733    
734       Fixed the bug described in blume-20010920-slowfp.
735    
736       The fix involves
737          1. generating FCOPYs in FSTP in ia32-svid
738          2. marking a CALL with the appropriate annotation
739    
740    ----------------------------------------------------------------------
741    Name: Matthias Blume
742    Date: 2001/10/16 11:32:00 EDT
743    Tag: blume-20011016-netbsd
744    Description:
745    
746    Underscore patch from Chris Richards (fixing problem with compiling
747    runtime system under recent NetBSD).
748    
749    ----------------------------------------------------------------------
750    Name: Allen Leung
751    Date: 2001/10/12 17:18:32 EDT 2001
752    Tag: leung-20011012-x86-printflowgraph
753    Description:
754    
755    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
756    
757    ----------------------------------------------------------------------
758    Name: Lal George
759    Date: 2001/10/11 23:51:34 EDT
760    Tag: george-20011011-too-many-instrs
761    Description:
762    
763    The representation of a program point never expected to see more
764    than 65536 instructions in a basic block!
765    
766    ----------------------------------------------------------------------
767    Name: Lal George
768    Date: 2001/10/09 09:41:37 EDT
769    Tag: george-20011008-mlrisc-labels
770    Description:
771    
772    Changed the machine description files to support printing of
773    local and global labels in assembly code, based on host assembler
774    conventions.
775    
776    ----------------------------------------------------------------------
777    Name: Matthias Blume
778    Date: 2001/09/25 15:25:00 EDT
779    Tag: blume-20010925-exninfo
780    Description:
781    
782    I provided a non-hook implementation of exnName (at the toplevel) and
783    made the "dummy" implementation of exnMessage (at the toplevel) more
784    useful: if nothing gets "hooked in", then at least you are going to
785    see the exception name and a message indicating why you don't see more.
786    
787    [For the time being, programs that need exnMessage and want to use
788    ml-build should either use General.exnMessage (strongly recommended) or
789    refer to structure General at some other point so that CM sees a
790    static dependency.]
791    
792    [Similar remarks go for "print" and "use":  If you want to use their
793    functionality in stand-alone programs generated by ml-build, then use
794    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
795    
796    ----------------------------------------------------------------------
797    Name: Matthias Blume
798    Date: 2001/09/20 17:28:00 EDT
799    Tag: blume-20010920-slowfp
800    Description:
801    
802    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
803    
804    ----------------------------------------------------------------------
805    Name: Matthias Blume
806    Date: 2001/09/20 17:20:00 EDT
807    Tag: blume-20010920-canonicalpaths
808    Description:
809    
810    0. Updated the BOOT file (something that I forgot to do earlier).
811    
812    1. Small internal change to CM so that it avoids "/../" in filenames
813       as much as possible (but only where it is safe).
814    
815    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
816       that arguments that contain delimiters are passed through correctly.
817       This change also means that all "special" arguments of the form
818       @SMLxxx... must come first.
819    
820    3. Changed install script to put relative anchor names for tool commands
821       into pathconfig.
822    
823    ----------------------------------------------------------------------
824    Name: Matthias Blume
825    Date: 2001/09/18 15:35:00 EDT
826    Tag: blume-20010918-readme11036
827    Description:
828    
829    Added README files.
830    
831    ----------------------------------------------------------------------
832    Name: Matthias Blume
833    Date: 2001/09/18 11:45:00 EDT
834    Tag: Release_110_36 (retag)
835    Description:
836    
837    Fixed mistake in config/preloads. Retagged as 110.36.
838    
839    ----------------------------------------------------------------------
840    Name: Matthias Blume
841    Date: 2001/09/18 09:40:00 EDT
842    Tag: Release_110_36_orig (tag changed)
843    Description:
844    
845    New version (110.36).  New bootfiles.
846    
847    ----------------------------------------------------------------------
848    Name: Matthias Blume
849    Date: 2001/09/14 16:15:00 EDT
850    Tag: blume-20010914-x86fastfp
851    Description:
852    
853    John committed some changes that Allen made, in particular a (hopefully)
854    correctly working version of the x86-fp module.
855    
856    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
857    flag to "true".  Everything seems to compile to a fixpoint ok, and
858    "mandelbrot" speeds up by about 15%.
859    
860    ----------------------------------------------------------------------
861    Name: Matthias Blume
862    Date: 2001/09/13 11:20:00 EDT
863    Tag: blume-20010913-minimal
864    Description:
865    
866    1. Stefan Monnier's patch to fix a miscompilation problem that
867       was brought to light by John Reppy's work on Moby.
868    
869    2. Implemented a minimal "structure Compiler" that contains just
870       "version" and "architecture".  The minimal version will be
871       available when the full version is not.  This is for backward-
872       compatibility with code that wants to test Compiler.version.
873    
874    ----------------------------------------------------------------------
875    Name: Matthias Blume
876    Date: 2001/08/28 14:03:00 EDT
877    Tag: blume-20010828-ml-lex
878    Description:
879    
880    Fix for bug 1581, received from Neophytos Michael.
881    
882    ----------------------------------------------------------------------
883    Name: Matthias Blume
884    Date: 2001/08/27 11:20:00 EDT
885    Tag: blume-20010827-readme11035
886    Description:
887    
888    Fleshed out the README file for 110.35.
889    
890    ----------------------------------------------------------------------
891    Name: Matthias Blume
892    Date: 2001/08/24 17:10:00 EDT
893    Tag: Release_110_35
894    Description:
895    
896    New version number (110.35).  New bootfiles.
897    
898    ----------------------------------------------------------------------
899    Name: Lal George
900    Date: 2001/08/24 13:47:18 EDT 2001
901    Tag: george-20010824-MLRISC-graphs
902    Description:
903    
904     removed clusters from MLRISC completely and replaced with graphs.
905    
906    ----------------------------------------------------------------------
907    Name: Matthias Blume
908    Date: 2001/08/23 17:50:00 EDT
909    Tag: blume-20010823-toplevel
910    Description:
911    
912    - some reorganization of the code that implements various kinds of
913      environments in the compiler (static, dynamic, symbolic, combined)
914    - re-implemented the EnvRef module so that evalStream works properly
915      (if the stream contains references to "use", "CM.make", etc.)
916    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
917    
918    ----------------------------------------------------------------------
919    Name: Matthias Blume
920    Date: 2001/08/20 15:50 EDT
921    Tag: blume20010820-slipup
922    Description:
923    
924    I forgot to commit a few files.  Here they are...
925    
926    ----------------------------------------------------------------------
927    Name: Matthias Blume
928    Date: 2001/08/20 15:35:00 EDT
929    Tag: blume-20010820-debugprof
930    Description:
931    
932    !!!! NEW BOOTFILES !!!!
933    
934    This is another round of reorganizing the compiler sources.  This
935    time the main goal was to factor out all the "instrumentation"
936    passes (for profiling and backtracing) into their own library.
937    The difficulty was to do it in such a way that it does not depend
938    on elaborate.cm but only on elabdata.cm.
939    
940    Therefore there have been further changes to both elaborate.cm and
941    elabdata.cm -- more "generic" things have been moved from the former
942    to the latter.  As a result, I was forced to split the assignment
943    of numbers indicating "primtyc"s into two portions: SML-generic and
944    SML/NJ-specific.  Since it would have been awkward to maintain,
945    I bit the bullet and actually _changed_ the mapping between these
946    numbers and primtycs.  The bottom line of this is that you need
947    a new set of bin- and bootfiles.
948    
949    I have built new bootfiles for all architectures, so doing a fresh
950    checkout and config/install.sh should be all you need.
951    
952    The newly created library's name is
953    
954        $smlnj/viscomp/debugprof.cm
955    
956    and its sources live under
957    
958        src/compiler/DebugProf
959    
960    ----------------------------------------------------------------------
961    Name: Matthias Blume
962    Date: 2001/08/15 17:15:00 EDT
963    Tag: blume-20010815-compreorg
964    Description:
965    
966    This is a first cut at reorganizing the CM libraries that make up the
967    core of the compiler.  The idea is to separate out pieces that could
968    be used independently by tools, e.g., the parser, the typechecker, etc.
969    
970    The current status is a step in this direction, but it is not quite
971    satisfactory yet.  Expect more changes in the future.
972    
973    Here is the current (new) organization...
974    
975        What used to be $smlnj/viscomp/core.cm is now divided into
976        six CM libraries:
977    
978             $smlnj/viscomp/basics.cm
979                           /parser.cm
980                           /elabdata.cm
981                           /elaborate.cm
982                           /execute.cm
983                           /core.cm
984    
985        The CM files for these libraries live under src/system/smlnj/viscomp.
986        All these libraries are proxy libraries that contain precisely
987        one CM library component.  Here are the locations of the components
988        (all within the src/compiler tree):
989    
990             Basics/basics.cm
991             Parse/parser.cm
992             ElabData/elabdata.cm
993             Elaborator/elaborate.cm
994             Execution/execute.cm
995             core.cm
996    
997         [This organization is the same that has been used already
998         for a while for the architecture-specific parts of the visible
999         compiler and for the old version of core.cm.]
1000    
1001         As you will notice, many source files have been moved from their
1002         respective original locations to a new home in one of the above
1003         subtrees.
1004    
1005         The division of labor between the new libraries is the following:
1006    
1007             basics.cm:
1008                - Simple, basic definitions that pertain to many (or all) of
1009                  the other libraries.
1010             parser.cm:
1011                - The SML parser, producing output of type Ast.dec.
1012                - The type family for Ast is also defined and exported here.
1013             elabdata.cm:
1014                - The datatypes that describe input and output of the elaborator.
1015                  This includes types, absyn, and static environments.
1016             elaborator.cm:
1017                - The SML/NJ type checker and elaborator.
1018                  This maps an Ast.dec (with a given static environment) to
1019                  an Absyn.dec (with a new static environment).
1020                - This libraries implements certain modules that used to be
1021                  structures as functors (to remove dependencies on FLINT).
1022             execute.cm:
1023                - Everything having to do with executing binary code objects.
1024                - Dynamic environments.
1025             core.cm:
1026                - SML/NJ-specific instantiations of the elaborator and MLRISC.
1027                - Top-level modules.
1028                - FLINT (this should eventually become its own library)
1029    
1030    Notes:
1031    
1032    I am not 100% happy with the way I separated the elaborator (and its
1033    data structures) from FLINT.  Two instances of the same problem:
1034    
1035        1. Data structures contain certain fields that carry FLINT-specific
1036           information.  I hacked around this using exn and the property list
1037           module from smlnj-lib.  But the fact that there are middle-end
1038           specific fields around at all is a bit annoying.
1039    
1040        2. The elaborator calculates certain FLINT-related information.  I tried
1041           to make this as abstract as I could using functorization, but, again,
1042           the fact that the elaborator has to perform calculations on behalf
1043           of the middle-end at all is not nice.
1044    
1045        3. Having to used exn and property lists is unfortunate because it
1046           weakens type checking.  The other alternative (parameterizing
1047           nearly *everything*) is not appealing, though.
1048    
1049    I removed the "rebinding =" warning hack because due to the new organization
1050    it was awkward to maintain it.  As a result, the compiler now issues some of
1051    these warnings when compiling init.cmi during bootstrap compilation. On
1052    the plus side, you also get a warning when you do, for example:
1053       val op = = Int32.+
1054    which was not the case up to now.
1055    
1056    I placed "assign" and "deref" into the _Core structure so that the
1057    code that deals with the "lazy" keyword can find them there.  This
1058    removes the need for having access to the primitive environment
1059    during elaboration.
1060    
1061    ----------------------------------------------------------------------
1062    Name: Matthias Blume
1063    Date: 2001/08/13
1064    Tag: blume-20010813-closures
1065    Description:
1066    
1067    This fix was sent to us by Zhong Shao.  It is supposed to improve the
1068    performance of certain loops by avoiding needless closure allocation.
1069    
1070    ----------------------------------------------------------------------
1071    Name: Lal George
1072    Date: 2001/07/31 10:03:23 EDT 2001
1073    Tag: george-20010731-x86-fmalloc
1074    Description: Fixed bug in x86 calls
1075    
1076        There was a bug where call instructions would mysteriously
1077        vanish. The call instruction had to be one that returned
1078        a floating point value.
1079    
1080    ----------------------------------------------------------------------
1081    Name: Lal George
1082    Date: 2001/07/19 16:36:29 EDT 2001
1083    Tag: george-20010719-simple-cells
1084    Description:
1085    
1086    I have dramatically simplified the interface for CELLS in MLRISC.
1087    
1088    In summary, the cells interface is broken up into three parts:
1089    
1090      1. CellsBasis : CELLS_BASIS
1091    
1092            CellsBasis is a top level structure and common for all
1093            architectures.  it contains the definitions of basic datatypes
1094            and utility  functions over these types.
1095    
1096      2. functor Cells() : CELLS
1097    
1098            Cells generates an interface for CELLS that incorporates the
1099            specific resources on the target architecture, such as the
1100            presence of special register classes, their number and size,
1101            and various useful substructures.
1102    
1103      3. <ARCH>CELLS
1104    
1105            e.g. SparcCells: SPARCCELLS
1106    
1107            <ARCH>CELLS usually contains additional bindings for special
1108            registers  on the architecture, such as:
1109    
1110                    val r0 : cell           (* register zero *)
1111                    val y : cell            (* Y register *)
1112                    val psr : cell          (* processor status register *)
1113                    ...
1114    
1115            The structure returned by applying the Cells functor is opened
1116            in this interface.
1117    
1118    The main implication of all this is that the datatypes for cells is
1119    split between CellsBasis and CELLS -- a fairly simple change for user
1120    code.
1121    
1122    In the old scheme the CELLS interface had a definitional binding of
1123    the form:
1124    
1125            signature CELLS = sig
1126    
1127               structure CellsBasis = CellsBasis
1128    
1129               ...
1130    
1131            end
1132    
1133    With all the sharing constraints that goes on in MLRISC, this old
1134    design  quickly leads to errors such as:
1135    
1136            "structure definition spec inside of sharing ... "
1137    
1138    
1139    and appears to require an unacceptable amount of sharing and where
1140    constraint hackery.
1141    
1142    I think this error message (the interaction of definitional specs and
1143    sharing) requires more explanation on our web page.
1144    
1145    ----------------------------------------------------------------------
1146    Name: Matthias Blume
1147    Date: 2001/07/19 15:00:00 EDT
1148    Tag: blume-20010719-libreorg
1149    Description:
1150    
1151    This update puts together a fairly extensive but straightforward change
1152    to the way the libraries that implement the interactive system are
1153    organized:
1154    
1155       The biggest change is the elimination of structure Compiler.  As a
1156       replacement for this structure, there is now a CM library
1157       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
1158       that exports all the substructures of the original structure Compiler
1159       directly.  So instead of saying Compiler.Foo.bar one now simply
1160       says Foo.bar.  (The CM libraries actually export a collection of
1161       structures that is richer than the collection of substructures of
1162       structure Compiler.)
1163    
1164       To make the transition smooth, there is a separate library called
1165       $smlnj/compiler/compiler.cm which puts together and exports the
1166       original structure Compiler (or at least something very close to it).
1167    
1168       There are five members of the original structure Compiler
1169       that are not exported directly but which instead became members
1170       of a new structure Backend (described by signature BACKEND).  These are:
1171       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
1172       Interact (: INTERACT), structure Machine (: MACHINE), and val
1173       architecture (: string).
1174    
1175       Structure Compiler.Version has become structure CompilerVersion.
1176    
1177       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
1178       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
1179       or x86, respectively.
1180       Each of these exports the same frontend structures that
1181       $smlnj/compiler.cm exports.  But they do not have a structure Backend
1182       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
1183       Hppa, PPC, Sparc, or X86, respectively.
1184    
1185       Library $smlnj/compiler/all.cm exports the union of the exports of
1186       $smlnj/compiler/<arch>.cm
1187    
1188       There are no structures <Arch>Compiler anymore, use
1189       $smlnj/compiler/<arch>.cm instead.
1190    
1191       Library host-compiler-0.cm is gone.  Instead, the internal library
1192       that instantiates CM is now called cm0.cm.  Selection of the host
1193       compiler (backend) is no longer done here but. (Responsibility for it
1194       now lies with $smlnj/compiler/current.cm.  This seems to be more
1195       logical.)
1196    
1197       Many individual files have been moved or renamed.  Some files have
1198       been split into multiple files, and some "dead" files have been deleted.
1199    
1200    Aside from these changes to library organization, there are also changes
1201    to the way the code itself is organized:
1202    
1203       Structure Binfile has been re-implemented in such a way that it no
1204       longer needs any knowledge of the compiler.  It exclusively deals
1205       with the details of binfile layout.  It no longer invokes the
1206       compiler (for the purpose of creating new prospective binfile
1207       content), and it no longer has any knowledge of how to interpret
1208       pickles.
1209    
1210       Structure Compile (: COMPILE) has been stripped down to the bare
1211       essentials of compilation.  It no longer deals with linking/execution.
1212       The interface has been cleaned up considerably.
1213    
1214       Utility routines for dealing with linking and execution have been
1215       moved into their own substructures.
1216    
1217       (The ultimate goal of these changes is to provide a light-weight
1218       binfile loader/linker (at least for, e.g., stable libraries) that
1219       does not require CM or the compiler to be present.)
1220    
1221    CM documentation has been updated to reflect the changes to library
1222    organization.
1223    
1224    ----------------------------------------------------------------------
1225    Name: Matthias Blume
1226    Date: 2001/07/10 17:30:00 EDT
1227    Tag: Release_110_34
1228    Description:
1229    
1230    Minor tweak to 110.34 (re-tagged):
1231    
1232      - README.html file added to CVS repository
1233      - runtime compiles properly under FreeBSD 3.X and 4.X
1234    
1235    ----------------------------------------------------------------------
1236    Name: Matthias Blume
1237    Date: 2001/07/10 17:30:00 EDT
1238    Tag: Release_110_34
1239    Description:
1240    
1241    New version number (110.34). New bootfiles.
1242    
1243    ----------------------------------------------------------------------
1244    Name: Matthias Blume
1245    Date: 2001/07/09 16:00:00 EDT
1246    Tag: blume-20010709-more-varargs
1247    Description:
1248    
1249    I changed the handling of varargs in ml-nlffigen again:
1250    The ellipsis ... will now simply be ignored (with an accompanying warning).
1251    
1252    The immediate effect is that you can actually call a varargs function
1253    from ML -- but you can't actually supply any arguments beyond the ones
1254    specified explicitly.  (For example, you can call printf with its format
1255    string, but you cannot pass additional arguments.)
1256    
1257    This behavior is only marginally more useful than the one before, but
1258    it has the advantage that a function or, more importantly, a function
1259    type never gets dropped on the floor, thus avoiding follow-up problems with
1260    other types that refer to the offending one.
1261    
1262    ----------------------------------------------------------------------
1263    Name: Matthias Blume
1264    Date: 2001/07/09 11:25:00 EDT
1265    Tag: blume-20010709-varargs
1266    Description:
1267    
1268    1. ckit-lib.cm now exports structure Error
1269    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1270       with a warning accompanied by a source location.  Moreover, it
1271       merely skips the offending function or type and proceeds with the
1272       rest of its work.u  As a result, one can safely feed C code containing
1273       "..." to ml-nlffigen.
1274    3. There are some internal improvements to CM, providing slightly
1275       more general string substitutions in the tools subsystem.
1276    
1277    ----------------------------------------------------------------------
1278    Name: Matthias Blume
1279    Date: 2001/06/27 15:10:00 EDT
1280    Tag: blume-20010627-concur
1281    Description:
1282    
1283    Fixed a small bug in CM's handling of parallel compilation.
1284    (You could observe the bug by Control-C-interrupting an ordinary
1285    CMB.make or CM.stabilize and then attaching some compile servers.
1286    The result was that all of a sudden the previously interrupted
1287    compilation would continue on its own.  This was because of
1288    an over-optimization: CM did not bother to clean out certain queues
1289    when no servers were attached "anyway", resulting in the contents
1290    of these queues to grab control when new servers did get attached.)
1291    
1292    There is also another minor update to the CM manual.
1293    
1294    ----------------------------------------------------------------------
1295    Name: Matthias Blume
1296    Date: 2001/06/26 16:15:00 EDT
1297    Tag: blume-20010626-cmdoc
1298    Description:
1299    
1300    Minor typo fixed in CM manual (syntax diagram for libraries).
1301    
1302    ----------------------------------------------------------------------
1303    Name: Matthias Blume
1304    Date: 2001/06/25 22:55:00 EDT
1305    Tag: blume-20010625-x86pc
1306    Description:
1307    
1308    Fixed a nasty bug in the X86 assembly code that caused signal
1309    handlers to fail (crash) randomly.
1310    
1311    ----------------------------------------------------------------------
1312    Name: Matthias Blume
1313    Date: 2001/06/25 12:05:00 EDT
1314    Tag: blume-20010625-nlffigen
1315    Description:
1316    
1317    This update fixes a number of minor bugs in ml-nlffigen as reported by
1318    Nick Carter <nbc@andrew.cmu.edu>.
1319    
1320      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
1321      2. Default names for generated files are now derived from the name of
1322         the C file *without its directory*.  In particular, this causes generated
1323         files to be placed locally even if the C file is in some system directory.
1324      3. Default names for generated signatures and structures are also derived
1325         from the C file name without its directory.  This avoids silly things
1326         like "structure GL/GL".
1327         (Other silly names are still possible because ml-nlffigen does not do
1328          a thorough check of whether generated names are legal ML identifiers.
1329          When in doubt, use command line arguments to force particular names.)
1330    
1331    ----------------------------------------------------------------------
1332    Name: Matthias Blume
1333    Date: 2001/06/21 12:25:00 EDT
1334    Tag: blume-20010621-eXene
1335    Description:
1336    
1337    eXene now compiles and (sort of) works again.
1338    
1339    The library name (for version > 110.33) is $/eXene.cm.
1340    
1341    I also added an new example in src/eXene/examples/nbody.  See the
1342    README file there for details.
1343    
1344    ----------------------------------------------------------------------
1345    Name: Matthias Blume
1346    Date: 2001/06/20 16:40:00 EDT
1347    Tag: blume-20010620-cml
1348    Description:
1349    
1350    CML now compiles and works again.
1351    
1352    Libraries (for version > 110.33):
1353    
1354      $cml/cml.cm            Main CML library.
1355      $cml/basis.cm          CML's version of $/basis.cm.
1356      $cml/cml-internal.cm   Internal helper library.
1357      $cml/core-cml.cm       Internal helper library.
1358      $cml-lib/trace-cml.cm  Tracing facility.
1359      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
1360    
1361    The installer (config/install.sh) has been taught how to properly
1362    install this stuff.
1363    
1364    ----------------------------------------------------------------------
1365    Name: Matthias Blume
1366    Date: 2001/06/19 17:55:00 EDT
1367    Tag: blume-20010619-instantiate
1368    Description:
1369    
1370    This un-breaks the fix for bug 1432.
1371    (The bug was originally fixed in 110.9 but I broke it again some
1372    time after that.)
1373    
1374    ----------------------------------------------------------------------
1375    Name: Matthias Blume
1376    Date: 2001/06/19 17:25:00 EDT
1377    Tag: blume-20010619-signals
1378    Description:
1379    
1380    This should (hopefully) fix the long-standing signal handling bug.
1381    (The runtime system was constructing a continuation record with an
1382    incorrect descriptor which would cause the GC to drop data on the floor...)
1383    
1384    ----------------------------------------------------------------------
1385    Name: Matthias Blume
1386    Date: 2001/06/15 15:05:00 EDT
1387    Tag: blume-20010615-moresparc
1388    Description:
1389    
1390    Here is a short late-hour update related to Sparc c-calls:
1391    
1392     -- made handling of double-word arguments a bit smarter
1393    
1394     -- instruction selection phase tries to collapse certain clumsily
1395        constructed ML-Trees; typical example:
1396    
1397            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
1398    
1399        This currently has no further impact on SML/NJ since mlriscGen does
1400        not seem to generate such patterns in the first place, and c-calls
1401        (which did generate them in the beginning) has meanwhile been fixed
1402        so as to avoid them as well.
1403    
1404    ----------------------------------------------------------------------
1405    Name: Matthias Blume
1406    Date: 2001/06/15 15:05:00 EDT
1407    Tag: blume-20010615-sparc
1408    Description:
1409    
1410    The purpose of this update is to provide an implementation of NLFFI
1411    on Sparc machines.
1412    
1413    Here are the changes in detail:
1414    
1415       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
1416       the Sparc implementation of the c-calls API.
1417       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
1418       for accessing the ML frame.  Thus, we have a real frame pointer and
1419       can freely modify %sp without need for an omit-frame-ptr phase.
1420       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
1421       this case.
1422       * ml-nlffigen has been taught to produce code for different architectures
1423       and calling conventions.
1424       * In a way similar to what was done in the x86 case, the Sparc
1425       backend uses its own specific extension to mltree.  (For example,
1426       it needs to be able to generate UNIMP instructions which are part
1427       of the calling convention.)
1428       * ml-nlffi-lib was reorganized to make it more modular (in particular,
1429       to make it easier to plug in new machine- and os-dependent parts).
1430    
1431    There are some other fairly unrelated bug fixes and cleanups as well:
1432    
1433       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
1434       that they properly share their libraries with existing SML/NJ libraries.
1435       * I fixed a minor cosmetic bug in CM, supressing certain spurious
1436       follow-up error messages.
1437       * Updates to CM/CMB documentation.
1438    
1439    TODO items:
1440    
1441       * MLRISC should use a different register as its asmTemp on the Sparc.
1442         (The current %o2 is a really bad choice because it is part of the
1443         calling conventions, so things might interfere in unexpected ways.)
1444    
1445    ----------------------------------------------------------------------
1446    Name: Matthias Blume
1447    Date: 2001/06/07
1448    Tag: blume-20010607-calls
1449    Description:
1450    
1451    A number of internal changes related to C calls and calling conventions:
1452    
1453    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
1454       number of bytes popped implicitly (by the callee).  In most cases
1455       this field is 0 but on x86/win32 it is some non-zero value.  This
1456       is information provided for the benefit of the "omit-frameptr" pass.
1457    2. The CALL instruction on the x86 carries a similar "pops" field.
1458       The instruction selection phase copies its value from the ML-Tree
1459       CALL statement.
1460    3. On all other architectures, the instruction selection phase checks
1461       whether "pops=0" and complains if not.
1462    4. The c-calls implementation for x86 now accepts two calling conventions:
1463       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
1464       up after the call and pops is set to 0.  For "stdcall", the caller
1465       does nothing, leaving the cleanup to the callee; pops is set to
1466       the number of bytes that were pushed onto the stack.
1467    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
1468       distinguish between "ccall" and "stdcall".
1469    6. The UNIMP instruction has been added to the supported Sparc instruction
1470       set. (This is needed for implementing the official C calling convention
1471       on this architecture.)
1472    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
1473       work with the latest CM.
1474    
1475    ----------------------------------------------------------------------
1476    Name: Matthias Blume
1477    Date: 2001/06/05 15:10:00 EDT
1478    Tag: blume-20010605-cm-index
1479    Description:
1480    
1481    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
1482    
1483    1. CM can now generate "index files".  These are human-readable files
1484       that list on a per-.cm-file basis each toplevel symbol defined or
1485       imported.  The location of the index file for
1486       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
1487       To enable index-file generation, set CM.Control.generate_index to true
1488       or export an environment-symbol: export CM_GENERATE_INDEX=true.
1489    
1490       The CM manual has been updated accordingly.
1491    
1492    2. I made some slight modifications to the c-calls API in MLRISC.
1493    
1494         a) There is now a callback to support saving/restoring of
1495            dedicated but caller-save registers around the actual call
1496            instruction.
1497         b) One can optionally specify a comment-annotation for the
1498            call instruction.
1499    
1500    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
1501       (For example, the comment annotation shows the C prototype of
1502        the function being called.)
1503    
1504    ----------------------------------------------------------------------
1505    Name: Matthias Blume
1506    Date: 2001/06/01 13:30:00 EDT
1507    Tag: blume-20010601-nlffi-cleanup
1508    Description:
1509    
1510    This is mostly a cleanup of MLFFI stuff:
1511    
1512       - some signature files have been put into a more exposed place
1513       - the ugly 'f type parameter is gone (simplifies types tremendously!)
1514       - ml-nlffigen changed accordingly
1515       - tutorial updated
1516    
1517    Other changes:
1518    
1519       - author's affiliation in CM manual(s) updated
1520       - some more recognized keywords added to Allen's sml.sty
1521    
1522    ----------------------------------------------------------------------
1523    Name: Matthias Blume
1524    Date: 2001/05/25 15:30:00 EDT
1525    Tag: blume-20010525-iptr
1526    Description:
1527    
1528      - put the official 110.33-README (as it appears on the ftp server) under
1529        CVS
1530      - fixed a small bug related to incomplete pointer types in
1531        ml-nlffigen
1532      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
1533        (it does not need the 'f type parameter)
1534    
1535    ----------------------------------------------------------------------
1536    Name: Matthias Blume
1537    Date: 2001/05/23 14:30:00 EDT
1538    Tag: Release_110_33
1539    Description:
1540    
1541    New version number (110.33).  New bootfiles.
1542    
1543    ----------------------------------------------------------------------
1544    Name: Matthias Blume
1545    Date: 2001/05/22 18:06:00 EDT
1546    Tag: blume-20010522-targets
1547    Description:
1548    
1549    Made install.sh use file config/targets.customized if it exists, falling
1550    back to config/targets if it doesn't.  This way one can have a customized
1551    version of the targets file without touching the "real thing", thus
1552    eliminating the constant fear of accidentally checking something bogus
1553    back into the CVS repository...  (File config/targets.customized must
1554    not be added to the repository!)
1555    
1556    ----------------------------------------------------------------------
1557    Name: Matthias Blume
1558    Date: 2001/05/22 16:30:00 EDT
1559    Tag: blume-20010522-minitut
1560    Description:
1561    
1562    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
1563       struct returns.
1564    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
1565       incomplete, preliminary documentation for NLFFI.
1566    
1567    ----------------------------------------------------------------------
1568    Name: Matthias Blume
1569    Date: 2001/05/14 11:30:00 EDT
1570    Tag: blume-20010514-script
1571    Description:
1572    
1573    Some bugs in install script fixed.
1574    
1575    In addition to that I also made a slight change to the NLFFI API:
1576    Functors generated by ml-nlffigen now take the dynamic library as a
1577    straight functor argument, not as a suspended one.  (The original
1578    functor code used to force the suspension right away anyway, so there
1579    was nothing gained by this complication of the interface.)
1580    
1581    ----------------------------------------------------------------------
1582    Name: Matthias Blume
1583    Date: 2001/05/11 14:35:00 EDT
1584    Tag: blume-20010511-ml-nlffi
1585    Description:
1586    
1587    I finally took the plunge and added my new FFI code to the main
1588    repository.  For x86-linux it is now ready for prime-time.
1589    
1590    There are two new subdirectories of "src":
1591    
1592      - ml-nlffi-lib:
1593           The utility library for programs using the FFI interface.
1594           Here is the implementation of $/c.cm and its associated low-level
1595           partners $/c-int.cm and $/memory.cm.
1596      - ml-nlffigen:
1597           A stand-alone program for generating ML glue code from C source
1598           code.
1599    
1600    Building ml-nlffigen requires $/ckit-lib.cm.
1601    
1602    The config/install.sh script has been updates to do the Right Thing
1603    (hopefully).
1604    
1605    Notice that the source tree for the C-Kit will not be put under "src"
1606    but directly under the installation root directory.  (This is the
1607    structure that currently exists on the CVS server when you check out
1608    module "sml".)  Fortunately, config/install.sh knows about this oddity.
1609    
1610    Bugs: No documentation yet.
1611    
1612    ----------------------------------------------------------------------
1613    Name: Matthias Blume
1614    Date: 2001/05/09 16:35:00 EDT
1615    Tag: blume-20010509-cpscontract
1616    Description:
1617    
1618    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
1619    wrapper/unwrapper elimination did not decrement usage counts and some
1620    dead variables got overlooked by the dead-up logic.)
1621    
1622    ----------------------------------------------------------------------
1623    Name: Lal George
1624    Date: 2001/05/08  17:26:09 EDT
1625    Tag: george-20010508-omit-frameptr
1626    Description:
1627    
1628    Changes to implement the omit-frame-pointer optimization to support
1629    raw C calls. For now, there is only support on the Intel x86, but
1630    other architectures will follow as more experience is gained with this.
1631    
1632    
1633    ----------------------------------------------------------------------
1634    Name: Matthias Blume
1635    Date: 2001/05/07 14:40:00 EDT
1636    Tag: blume-20010507-proxies
1637    Description:
1638    
1639    I made into "proxy libraries" all libraries that qualify for such a
1640    change.  (A qualifying library is a library that has another library or
1641    groups as its sole member and repeats that member's export list
1642    verbatim.  A proxy library avoids this repetition by omitting its export
1643    list, effectively inheriting the list that its (only) member exports.
1644    See the CM manual for more explanation.)
1645    The main effect is that explicit export lists for these libraries
1646    do not have to be kepts in sync, making maintenance a bit easier.
1647    
1648    I also added copyright notices to many .cm-files.
1649    
1650    Last but not least, I made a new set of bootfiles.
1651    
1652    ----------------------------------------------------------------------
1653    Name: Matthias Blume
1654    Date: 2001/05/04 17:00:00 EDT
1655    Tag: blume-20010504-cm-lsplit
1656    Description:
1657    
1658    0. John merged pending changes to $/smlnj-lib.cm
1659    
1660    1. Allen's previous change accidentally backed out of one of Lal's
1661       earlier changes.  I undid this mistake (re-introducing Lal's change).
1662    
1663    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
1664       within the compiler where applicable.  There is some code simplification
1665       because of that.
1666    
1667    3. The "split" phase (in FLINT) is now part of the default list of phases.
1668       Compiler.Control.LambdaSplitting.* can be used to globally control the
1669       lambda-splitting (cross-module-inlining) engine.  In addition to that,
1670       it can now also be controlled on a per-source basis: CM has been taught
1671       a new tool parameter applicable to ML source files.
1672    
1673       - To turn lambda-splitting off completely:
1674            local open Compiler.Control.LambdaSplitting in
1675                val _ = set Off
1676            end
1677       - To make "no lambda-splitting" the global default (but allow per-source
1678         overriding); this is the initial setting:
1679            local open Compiler.Control.LambdaSplitting in
1680                val _ = set (Default NONE)
1681            end
1682       - To make "lambda-splitting with aggressiveness a" the global default
1683         (and allow per-source overriding):
1684            local open Compiler.Control.LambdaSplitting in
1685                val _ = set (Default (SOME a))
1686            end
1687    
1688       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
1689         write (in the respective .cm-file):
1690            a.sml (lambdasplitting:off)
1691       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
1692            a.sml (lambdasplitting:on)
1693       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
1694         <a> is a decimal non-negative integer):
1695            a.sml (lambdasplitting:<a>)
1696       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
1697            a.sml (lambdasplitting:infinity)
1698       - To use the global default for a.sml:
1699            a.sml (lambdasplitting:default)
1700         or simply
1701            a.sml
1702    
1703    ----------------------------------------------------------------------
1704    Name: Allen Leung
1705    Date: 2001/05/04 01:57:00 EDT
1706    Tag: leunga-20010504-sync
1707    Description:
1708    
1709      MLRISC features.
1710    
1711      1. Fix to CMPXCHG instructions.
1712      2. Changed RA interface to allow annotations in callbacks.
1713      3. Added a new method to the stream interface to allow annotations updates.
1714    
1715    ----------------------------------------------------------------------
1716    Name: Matthias Blume
1717    Date: 2001/05/01 11:45:00 EDT
1718    Tag: blume-20010501-pcedittmp
1719    Description:
1720    
1721    Changed install.sh to use the current working directory instead of
1722    /usr/tmp for a temporary file (pcedittmp).  The previous choice
1723    of /usr/tmp caused trouble with MacOS X because of file premission
1724    problems.
1725    
1726    ----------------------------------------------------------------------
1727    Name: Matthias Blume
1728    Date: 2001/04/20 11:10:00 EDT
1729    Tag: blume-20010420-inMLflag
1730    Description:
1731    
1732     - added vp_limitPtrMask to vproc-state.h
1733       (for use by the raw-C-calls mechanism to implement proper interrupt
1734        handling)
1735     - made the ML compiler aware of various data-structure offsets so it
1736       can generate code for accessing the vp_inML flag and vp_limitPtrMask
1737     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
1738       raw C-calls
1739    
1740    ----------------------------------------------------------------------
1741    Name: Lal George
1742    Date: 2001/04/20 09:15:28 EDT
1743    Tag: george-20010420-macosX
1744    Description:
1745    
1746     - Changes to port to Mac OS X; Darwin.
1747    
1748     - In the process I found that sqrt was broken on the PPC, because the
1749       fsqrt instruction is not implemented.
1750    
1751    ----------------------------------------------------------------------
1752    Name: Matthias Blume
1753    Date: 2001/04/18 12:45:00 EDT
1754    Tag: blume-20010418-ccalls
1755    Description:
1756    
1757     - fixed two off-by-4 errors in the x86-specific c-calls implementation
1758       (this bug prevented structure arguments containing pointers from being
1759        passed correctly)
1760     - changed the raw-C-call code in mlriscGen.sml in such a way that
1761       structure arguments are represented as a pointer to the beginning
1762       of the structure (instead of having a series of synthesized arguments,
1763       one for each structure member)
1764    
1765     - made makeml script's verbosity level configurable via environment
1766       variable (MAKEML_VERBOSITY)
1767    
1768     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
1769       in rawmem-x86.sml; we are now using the real thing
1770    
1771    ----------------------------------------------------------------------
1772    Name: Matthias Blume
1773    Date: 2001/03/22 16:25:00 EST
1774    Tag: blume-20010322-bootfiles
1775    Description:
1776    
1777    Created a new set of bootfiles (for your automatic installation convenience).
1778    
1779    ----------------------------------------------------------------------
1780    Name: Matthias Blume
1781    Date: 2001/03/22 15:10:00 EST
1782    Tag: blume-20010322-rawmem-parcm
1783    Description:
1784    
1785    1. All "raw memory access" primitives for the new FFI are implemented now
1786       (at least on the x86).
1787    2. Some further cleanup of CM's parallel make mechanism.
1788    
1789    ----------------------------------------------------------------------
1790    Name: Matthias Blume
1791    Date: 2001/03/19 17:53:00 EST
1792    Tag: blume-20010319-parallel
1793    Description:
1794    
1795    Parallel make (using compile servers) now works again.
1796    
1797    To this end, CM.stabilize and CMB.make have been modified to work in
1798    two passes when compile servers are attached:
1799       1. Compile everything, do not perform stabilization; this pass
1800          uses compile servers
1801       2. Stabilize everything; this pass does not use compile servers
1802    If there are no compile servers, the two passes are combined into one
1803    (as before).  Splitting the passes increases the inherent parallelism
1804    in the dependency graph because the entire graph including all
1805    libraries is available at the same time.  This, in turn, improves
1806    server utilization.  The downside is that the master process will
1807    have to do some extra work after compilation is done (because for
1808    technical reasons it must re-read all the binfiles during stabilization).
1809    
1810    ----------------------------------------------------------------------
1811    Name: Matthias Blume
1812    Date: 2001/03/16 12:22:00 EST
1813    Tag: blume-20010316-bootfiles
1814    Description:
1815    
1816    Created a new set of bootfiles (for your automatic installation convenience).
1817    
1818    ----------------------------------------------------------------------
1819    Name: Matthias Blume
1820    Date: 2001/03/16 11:00:00 EST
1821    Tag: blume-20010316-MLTREE-fixup
1822    Description:
1823    
1824    This is a minor fixup for an (untagged) earlier commit by Allen.
1825    (A file was missing).
1826    
1827    ----------------------------------------------------------------------
1828    Name: Allen Leung
1829    Date: Mon Mar  5 18:54:57 EST 2001
1830    Tag: leunga-20010305-cut-support
1831    
1832    1. New support for alternative control-flow in MLTREE.
1833       Currently we support
1834    
1835          FLOW_TO(CALL ...., [k1,...,kn])
1836    
1837       This is needed for 'cuts to' in C-- and try/handle-like constructs
1838       in Moby
1839    
1840       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
1841    
1842    2. Register Allocator
1843    
1844       Changes in interface [from Fermin, John]
1845    
1846    3. Alpha 8-bit SLL support [Fermin]
1847    
1848    4. All architectures
1849    
1850       A new module (ClusterExpandCopies) for expanding parallel copies.
1851    
1852    ----------------------------------------------------------------------
1853    Name: Allen Leung
1854    Date: 2001/02/27 23:07:00 EST
1855    Tag: leunga-20010227-minor-stuff
1856    
1857    1. Alpha bug fix for CMOVNE
1858    2. Handle mltree COND(..,FCMP ...,...)
1859    3. Bug fix in simplifier
1860    
1861    ----------------------------------------------------------------------
1862    Name: Matthias Blume
1863    Date: 2001/01/30 17:50:00 EST
1864    Tag: blume-20010130-sync
1865    Description:
1866    
1867    This is just a minor update to sync my devel branch with the main brach.
1868    The only visible change is the addition of some README files.
1869    
1870    ----------------------------------------------------------------------
1871    Name: Matthias Blume
1872    Date: 2001/01/12 23:30:00 JST
1873    Tag: blume-20010112-bootfiles
1874    Description:
1875    
1876    Made a new set of bootfiles that goes with the current state of the
1877    repository.
1878    
1879    ----------------------------------------------------------------------
1880    Name: Matthias Blume
1881    Date: 2001/01/12 21:20:00 JST
1882    Tag: blume-20010112-sync
1883    Description:
1884    
1885    I am just flushing out some minor changes that had accumulated in
1886    my private branch in order to sync with the main tree.  (This is
1887    mainly because I had CVS trouble when trying to merge _into_ my
1888    private branch.)
1889    
1890    Most people should be completely unaffected by this.
1891    
1892    ----------------------------------------------------------------------
1893    Name: Allen Leung
1894    Date: Thu Jan 11 21:03:00 EST 2001
1895    Tag: leunga-20010111-labexp=mltree
1896    Description:
1897    
1898    1.  Removed the type LabelExp and replace it by MLTree.
1899    2.  Rewritten mltree-simplify with the pattern matcher tool.
1900    3.  There were some bugs in alpha code generator which would break
1901        64-bit code generation.
1902    4.  Redo the tools to generate code with the
1903    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
1904        are now generated by perl scripts.
1905    
1906    ----------------------------------------------------------------------
1907    Name: Matthias Blume
1908    Date: 2001/01/10 21:55:00 JST
1909    Tag: blume-20010110-rcc
1910    Description:
1911    
1912    The RCC stuff now seems to work (but only on the x86).
1913    This required hacking of the c-calls interface (and -implementation) in
1914    MLRISC.
1915    
1916    Normal compiler users should be unaffected.
1917    
1918    ----------------------------------------------------------------------
1919    Name: Matthias Blume
1920    Date: 2001/01/09 01:20:00 JST
1921    Tag: blume-20010109-rcc
1922    Description:
1923    
1924    This is a fairly big patch, flushing out a large number of pending
1925    changes that I made to my development copy over the last couple of days.
1926    
1927    Of practical relevance at this moment is a workaround for a pickling
1928    bug that Allen ran into the other day.  The cause of the bug itself is
1929    still unknown and it might be hard to fix it properly, but the
1930    workaround has some merits of its own (namely somewhat reducing pickling
1931    overhead for certain libraries).  Therefore, I think this solution should
1932    be satisfactory at this time.
1933    
1934    The rest of the changes (i.e., the vast majority) has to do with my
1935    ongoing efforts of providing direct support for C function calls from
1936    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
1937    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
1938    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
1939    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
1940    handle RCC.
1941    
1942    The last part (the changes to mlriscGen.sml) are still known to be
1943    wrong on the x86 and not implemented on all other architectures.  But
1944    the infrastructure is in place. I had to change a few functor
1945    signatures in the backend to be able to route the CCalls interface
1946    from MLRISC there, and I had to specialize the mltree type (on the
1947    x86) to include the necessary extensions. (The extensions themselves
1948    were already there and redy to go in MLRISC/x86).
1949    
1950    Everything should be very happy as soon as someone helps me with
1951    mlriscGen.sml...
1952    
1953    In any case, nothing of this should matter to anyone as long as the
1954    new primop is not being used (which is going to be the case unless you
1955    find it where I hid it :). The rest of the compiler is completely
1956    unaffected.
1957    
1958    ----------------------------------------------------------------------
1959    Name: Matthias Blume
1960    Date: 2001/01/05 00:30:00 JST
1961    Tag: blume-20010105-primops
1962    Description:
1963    
1964    Added some experimental support for work that I am doing right now.
1965    These changes mostly concern added primops, but there is also a new
1966    experimental C library in the runtime system (but currently not enabled
1967    anywhere except on Linux/X86).
1968    
1969    In the course of adding primops (and playing with them), I discovered that
1970    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
1971    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
1972    To recover, I made type information in INL_PRIM mandatory and changed
1973    prim.sml as well as built-in.sml accordingly.  The InLine structure now
1974    has complete, correct type information (i.e., no bottom types).
1975    
1976    Since all these changes mean that we need new binfiles, I also bumped the
1977    version number to 110.32.1.
1978    
1979    ----------------------------------------------------------------------
1980    Name: Matthias Blume
1981    Date: 2000/12/30 22:10:00 JST
1982    Tag: blume-20001230-various
1983    Description:
1984    
1985    Added proxy libraries for MLRISC and let MLRISC libraries refer
1986    to each other using path anchors.  (See CM manual for explanation.)
1987    
1988    Updated CM documentation.
1989    
1990    Fixed some bugs in CM.
1991    
1992    Implemented "proxy" libraries (= syntactic sugar for CM).
1993    
1994    Added "-quiet" option to makeml and changed runtime system accordingly.
1995    
1996    Added cleanup handler for exportML to reset timers and compiler stats.
1997    
1998    ----------------------------------------------------------------------
1999    Name: Lal George
2000    Date: 2000/12/22 22:22:58 EST 2000
2001    Tag: Release_110_32
2002    Description:
2003    
2004            Infinite precision used throughout MLRISC.
2005            see MLRISC/mltree/machine-int.sig
2006    
2007    ----------------------------------------------------------------------
2008    Name: Matthias Blume
2009    Date: 2000/12/22 23:16:00 JST
2010    Tag: blume-20001222-warn
2011    Description:
2012    
2013    Corrected wording and formatting of some CM warning message which I
2014    broke in my previous patch.
2015    
2016    ----------------------------------------------------------------------
2017    Name: Matthias Blume
2018    Date: 2000/12/22 21:20:00 JST
2019    Tag: blume-20001222-anchorenv
2020    Description:
2021    
2022    Fixed CM's handling of anchor environments in connection with CMB.make.
2023    
2024    ----------------------------------------------------------------------
2025    Name: Matthias Blume
2026    Date: 2000/12/22 13:15:00 JST
2027    Tag: blume-20001222-cleanup
2028    Description:
2029    
2030    Removed src/cm/ffi which does not (and did not) belong here.
2031    
2032    ----------------------------------------------------------------------
2033    Name: Matthias Blume
2034    Date: 2000/12/21 23:55:00 JST
2035    Tag: blume-20001221-exn
2036    Description:
2037    
2038    Probably most important: CM no longer silently swallows all exceptions
2039    in the compiler.
2040    Plus: some other minor CM changes.  For example, CM now reports some
2041    sizes for generated binfiles (code, data, envpickle, lambdapickle).
2042    
2043    ----------------------------------------------------------------------
2044    Name: Matthias Blume
2045    Date: 2000/12/15 00:01:05 JST
2046    Tag: blume-20001215-dirtool
2047    Description:
2048    
2049    - "dir" tool added.
2050    - improvements and cleanup to Tools structure
2051    - documentation updates
2052    
2053    ----------------------------------------------------------------------
2054    Name: Allen Leung
2055    Date: Thu Dec 14 03:45:24 EST 2000
2056    Description:
2057    Tag:  leunga-20001214-int-inf
2058    Description:
2059    
2060       In IntInf, added these standard functions, which are missing from our
2061    implementation:
2062    
2063        andb : int * int -> int
2064        xorb : int * int -> int
2065        orb  : int * int -> int
2066        notb : int -> int
2067         <<   : int * word -> int
2068        ~>>  : int * word -> int
2069    
2070       Not tested, I hope they are correct.
2071    
2072    ----------------------------------------------------------------------
2073    Name: Allen Leung
2074    Date: Fri Dec  8 19:23:26 EST 2000
2075    Description:
2076    Tag:  leunga-20001208-nowhere
2077    Description:
2078    
2079      Slight improvements to the 'nowhere' tool to handle OR-patterns,
2080    to generate better error messages etc.  Plus a brief manual.
2081    
2082    ----------------------------------------------------------------------
2083    Name: Lal George
2084    Date: 2000/12/08 09:54:02 EST 2000
2085    Tag: Release_110_31
2086    Description:
2087    
2088    - Version 110.31
2089    ----------------------------------------------------------------------
2090    Name: Allen Leung
2091    Date: Thu Dec  7 22:01:04 EST 2000
2092    Tag:  leunga-20001207-cell-monster-hack
2093    Description:
2094    
2095    Major MLRISC internal changes.  Affect all clients.
2096    Summary:
2097    
2098    1.  Type CELLS.cell = int is now replaced by a datatype.
2099        As a result, the old regmap is now gone.  Almost all interfaces
2100        in MLRISC change as a consequence.
2101    
2102    2.  A new brand version of machine description tool (v3.0) that generates
2103        modules expecting the new interface.  The old version is removed.
2104    
2105    3.  The RA interface has been further abstracted into two new functors.
2106        RISC_RA and X86RA.  These functors have much simpler interfaces.
2107        [See also directory MLRISC/demo.]
2108    
2109    4.  Some other new source->source code generation tools are available:
2110    
2111        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
2112        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
2113           I use this tool to generate the peephole optimizers---with the new
2114           cell type changes, peephole rules are becoming difficult to write
2115           without conditional pattern matching.
2116    
2117    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
2118        cover the entire MLRISC source tree so many things broke.
2119    
2120    6.  CM files have been moved to the subdirectory MLRISC/cm.
2121        They are moved because there are a lot of them and they clutter up the
2122        root dir.
2123    
2124    7.  More detailed documentation to come...
2125    
2126        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
2127        a makeml -rebuild first.  This is because of other other
2128        changes that Matthias has made (see below).
2129    
2130    
2131    ----------------------------------------------------------------------
2132    Name: Matthias Blume
2133    Date: 2000/11/30 23:12:00 JST
2134    Tag: blume-20001130-filereorg
2135    Description:
2136    
2137    Some manual updates and some file reorganizations in CM.
2138    
2139    ----------------------------------------------------------------------
2140    Name: Matthias Blume
2141    Date: 2000/11/24 17:45:00 JST
2142    Tag: blume-20001124-link
2143    Description:
2144    
2145    Drastically improved link traversal code for the case that the dynamic
2146    value was already loaded at bootstrap time.  As a result, CM and CMB
2147    now both load blazingly fast -- even on a very slow machine.  Also,
2148    memory consumption has been further reduced by this.
2149    
2150    Warning: The format of the PIDMAP file has changed.  THerefore, to
2151    bootstrap you have to do this:
2152    
2153    1. Run CMB.make
2154    2. Make a symbolic link for the boot directory:
2155         ln -s sml.boot.ARCH-OS xxx
2156    3. "Rebuild" the boot directory:
2157         ./makeml -boot xxx -rebuild sml ; rm xxx
2158    4. Boot normally:
2159          ./makeml
2160    
2161    ----------------------------------------------------------------------
2162    Name: Matthias Blume
2163    Date: 2000/11/21 21:20:00 JST
2164    Tag: blume-20001121-tools
2165    Description:
2166    
2167    Continued hacking on autoloading problem -- with success this time.
2168    Also changed tool-plugin mechanism.  See new CM manual.
2169    
2170    ----------------------------------------------------------------------
2171    Name: Matthias Blume
2172    Date: 2000/11/19 14:30:00 JST
2173    Tag:  blume-20001119-autoload
2174    Description:
2175    
2176    Some hacking to make autoloading faster.  Success for CMB, no success
2177    so far for CM.  There is a reduced structure CM' that autoloads faster.
2178    (This is a temporary, non-documented hack to be eliminated again when
2179    the general problem is solved.)
2180    
2181    ----------------------------------------------------------------------
2182    Name: Matthias Blume
2183    Date: 2000/11/17 14:10:00 JST
2184    Tag: blume-20001117-pickle-lib
2185    Description:
2186    
2187    1. Eliminated comp-lib.cm
2188    2. Made pickle-lib.cm
2189    3. Eliminated all uses of intset.sml (from comp-lib.cm)
2190    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
2191       equivalent constructs from smlnj-lib.cm (INtHashTable).
2192    5. Point 4. also goes for those uses of intmap.* in MLRISC.
2193       Duplicated intmap modules thrown out.
2194    6. Hunted down all duplicated SCC code and replaced it with
2195       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
2196    7. Rewrote Feedback module.
2197    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
2198       should be thrown out and equivalent modules from smlnj-lib.cm
2199       should be used (IntRedBlackSet, IntListSet, ...).
2200    
2201    Confirmed that compiler compiles to fixpoint.
2202    
2203    ----------------------------------------------------------------------
2204    Name: Allen Leung
2205    Date: 2000/11/10 18:00:00
2206    Tag: leunga-20001110-new-x86-fp
2207    
2208    A new x86 floating point code generator has been added.
2209    By default this is turned off.  To turn this on, do:
2210    
2211        CM.autoload "$smlnj/compiler.cm";
2212        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
2213    
2214    Changes:
2215    
2216    1.  Changed FTAN to FPTAN so that the assembly output is correct.
2217    2.  Changed the extension callback for FTANGENT to generate:
2218    
2219              fptan
2220              fstp  %st(0)
2221        instead of
2222              fptan
2223              fstpl ftempmem
2224    
2225    3.  Numerous assembly fixes for x86.
2226    
2227    5.  Cleaned up the machine code output module x86/x86MC.sml and added
2228        support for a whole bunch of instructions and addressing modes:
2229    
2230          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
2231          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
2232          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
2233          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
2234          fxch %st(n)
2235          fld %st(n)
2236          fst %st(n)
2237          fst mem
2238          fstp %st(n)
2239          fucom %st(n)
2240          fucomp %st(n)
2241    
2242        All these are now generated when the fast fp mode is turned on.
2243    
2244    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
2245    
2246    ----------------------------------------------------------------------
2247    Name: Matthias Blume
2248    Date: 2000/11/09 11:20:00 JST
2249    Tag: blume-20001109-scc
2250    Description:
2251    
2252    Eliminated some code duplication:
2253    
2254    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
2255       (Otherwise the functor is useless.)
2256    2. Used GraphSCCFn where SCCUtilFun was used previously.
2257    3. Got rid of SCCUtilFun (in comp-lib.cm).
2258    
2259    ----------------------------------------------------------------------
2260    Name: Lal George
2261    Date: 2000/11/06 09:02:21 EST 2000
2262    Tag: Release_110_30
2263    Description:
2264    
2265    - Version 110.30
2266    ----------------------------------------------------------------------
2267    Name: Matthias Blume
2268    Date: 2000/11/04 14:45:00
2269    Tag: blume-20001104-mlbuild
2270    Description:
2271    
2272    - Made ml-build faster on startup.
2273    - Documentation fixes.
2274    
2275    ----------------------------------------------------------------------
2276    Name: Matthias Blume
2277    Date: 2000/11/02 17:00:00 JST
2278    Tag: blume-20001102-condcomp
2279    Description:
2280    
2281    - Small tweaks to pickler -- new BOOTFILES!
2282    - Version bumped to 110.29.2.
2283    - Added conditional compilation facility to init.cmi (see comment there).
2284    ----------------------------------------------------------------------
2285    Name: Allen Leung
2286    Date: 2000/10/23 19:31:00
2287    Tag: leunga-20001023-demo-ra
2288    
2289    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
2290    2. Test programs for the graph library updated
2291    3. Some new MLRISC demo programs added
2292    
2293    ----------------------------------------------------------------------
2294    Name: Matthias Blume
2295    Date: 2000/08/31 22:15:00 JST
2296    Tag: blume-20001017-errmsg
2297    Description:
2298    
2299    More error message grief: Where there used to be no messages, there
2300    now were some that had bogus error regions.  Fixed.
2301    
2302    ----------------------------------------------------------------------
2303    Name: Matthias Blume
2304    Date: 2000/08/31 17:30:00 JST
2305    Tag: blume-20001017-v110p29p1
2306    Description:
2307    
2308    I made a version 110.29.1 with new bootfiles.
2309    
2310    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
2311              CM documentation changes and a small bugfix in CM's error reporting.
2312    
2313    ----------------------------------------------------------------------
2314    Name: Lal George
2315    Date: 2000/09/27 14:42:35 EDT
2316    Tag: george-20000927-nodestatus
2317    Description:
2318    
2319    Changed the type of the nodestatus, so that:
2320    
2321            SPILLED(~1)             is now SPILLED
2322            SPILLED(m) where m>=0   is now MEMREG(m)
2323            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
2324    
2325    ----------------------------------------------------------------------
2326    Name: Matthias Blume
2327    Date: 2000/09/07 14:45:00 JST
2328    Tag: blume-20000907-cmerrmsg
2329    Description:
2330    
2331    Small tweak to CM to avoid getting ML syntax error messages twice.
2332    
2333    ----------------------------------------------------------------------
2334    Name: Matthias Blume
2335    Date: 2000/08/31 18:00:00 JST
2336    Tag: blume-20000831-cvsbootfiles
2337    Description:
2338    
2339    New URL for boot files (because the 110.29 files on the BL server do
2340    now work correctly with my updated install scripts for yacc and lex).
2341    
2342    ----------------------------------------------------------------------
2343    Name: Matthias Blume
2344    Date: 2000/08/08 12:33:00 JST
2345    Tag: blume-20000808-manual
2346    Description:
2347    
2348    Tiny update to CM manual.
2349    
2350    ----------------------------------------------------------------------
2351    Name: Allen Leung
2352    Date: 2000/08/7 19:31:00
2353    Tag: leunga-20000807-a-whole-bunch-of-stuff
2354    
2355      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
2356    and Mobdy.
2357    
2358    1.  x86
2359    
2360       a.  Fixes to peephole module by John and Dan.
2361       b.  Assembly fix to SETcc by Allen.
2362       c.  Fix to c-call by John.
2363       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
2364       e.  Instruction selection optimization to SETcc as suggested by John.
2365    
2366           For example,
2367    
2368            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
2369    
2370           should generate:
2371    
2372            MOVL a, x
2373            SUBL b, x
2374            SHRL 31, x
2375    
2376    2.  IR stuff
2377    
2378         A bunch of new DJ-graph related algorithms added.  These
2379         speed up SSA construction.
2380    
2381    3.  SSA + Scheduling
2382    
2383         Added code for SSA and scheduling to the repository
2384    
2385    ----------------------------------------------------------------------
2386    Name: Lal George
2387    Date: 2000/07/27 11:53:14 EDT
2388    
2389    Tag: lal-20000727-linux-ppc
2390    Description:
2391    
2392     Made changes to support Linux PPC.
2393     p.s. I have confirmation that the 110.29 boot files work fine.
2394    
2395    ----------------------------------------------------------------------
2396    Name: Matthias Blume
2397    Date: 2000/07/27 17:40:00 JST
2398    Tag: blume-20000727-scripts
2399    Description:
2400    
2401    !!!! WARNING !!!!
2402    You must recompile the runtime system!
2403    !!!! WARNING !!!!
2404    
2405    This is basically another round of script-enhancements:
2406    
2407    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
2408       and undefine CM preprocessor symbols.
2409    
2410    2. ml-build avoids generating a new heap image if it finds that the
2411       existing one is still ok.  (The condition is that no ML file had to
2412       be recompiled and all ML files are found to be older that the heap
2413       file.)
2414    
2415       To make this work smoothly, I also hacked the runtime system as
2416       well as SMLofNJ.SysInfo to get access to the heap image suffix
2417       (.sparc-solaris, ...) that is currently being used.
2418    
2419       Moreover, the signature of CM.mk_standalone has changed.  See the
2420       CM manual.
2421    
2422    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
2423       CM manual for details.)
2424    
2425    4. More CM manual updates:
2426        - all of the above has been documented.
2427        - there is now a section describing the (CM-related) command line
2428          arguments that are accepted by the "sml" command
2429    
2430    ----------------------------------------------------------------------
2431    Name: Matthias Blume
2432    Date: 2000/07/25 16:20:00 JST
2433    Tag: blume-20000725-makedepend
2434    Description:
2435    
2436    Added a script called ml-makedepend.  This can be used in makefiles
2437    for Unix' make in a way very similar to the "makedepend" command for
2438    C.
2439    
2440    The script internally uses function CM.sources.
2441    
2442    Synopsis:
2443    
2444        ml-makedepend [-f makefile] cmfile targetname
2445    
2446    The default for the makefile is "makefile" (or "Makefile" should
2447    "makefile" not exist).
2448    
2449    ml-makedepend adds a cmfile/targetname-specific section to this
2450    makefile (after removing the previous version of this section).  The
2451    section contains a single dependency specification with targetname on
2452    the LHS (targetname is an arbitrary name), and a list of files derived
2453    from the cmfile on the RHS.  Some of the files on the RHS are
2454    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
2455    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
2456    path names.  The makefile writer is responsible for making sure that
2457    these variables have correct at the time "make" is invoked.
2458    
2459    ----------------------------------------------------------------------
2460    Name: Matthias Blume
2461    Date: 2000/07/22 23:30:00 JST
2462    Tag: blume-20000722-urlupdate
2463    Description:
2464    
2465    Changed BOOT and config/srcarchiveurl to point to BL server:
2466    
2467        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
2468    
2469    ----------------------------------------------------------------------
2470    Name: Matthias Blume
2471    Date: 2000/07/18 18:00:00 JST
2472    Tag: blume-20000718-Version_110_29
2473    Description:
2474    
2475    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
2476    
2477    2. Updated config/version to 110.29
2478    
2479    3. Updated config/srcarchiveurl
2480    
2481    3. New boot files!
2482       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
2483    
2484    ----------------------------------------------------------------------
2485    Name: Matthias Blume
2486    Date: 2000/07/11 13:58:00 JST
2487    Tag: blume-20000711-doctypo
2488    Description:
2489    
2490    Fixed a few typos in CM manual.
2491    
2492    ----------------------------------------------------------------------
2493    Name: Allen Leung
2494    Date: 2000/06/15 00:38:00
2495    Tag: leunga-20000704-sparc-x86
2496    
2497    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
2498    2. fix to x86 RET bug [found by Dan Grossman]
2499    3. sparc assembly bug fix for ticc instructions [found by Fermin]
2500    
2501       Affects c-- and moby only
2502    
2503    ----------------------------------------------------------------------
2504    Name: Matthias Blume
2505    Date: 2000/07/04 15:26:00
2506    Tag: blume-20000704-trigger
2507    Description:
2508    
2509    1. Improvements to CM manual.
2510    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
2511       of getting a back-trace.  The function, when called, raises an
2512       internal exception which explicitly carries the full back-trace history,
2513       so it is unaffected by any intervening handle-raise pairs ("trivial"
2514       or not).  The interactive loop will print that history once it arrives
2515       at top level.
2516       Short of having all exceptions implicitly carry the full history, the
2517       recommended way of using this facility is:
2518         - compile your program with instrumentation "on"
2519         - run it, when it raises an exception, look at the history
2520         - if the history is "cut off" because of some handler, go and modify
2521           your program so that it explicitly calls BTrace.trigger
2522         - recompile (still instrumented), and rerun; look at the full history
2523    
2524    ----------------------------------------------------------------------
2525    Name: Matthias Blume
2526    Date: 2000/07/03 15:36:00 JST
2527    Tag: blume-20000702-manual
2528    Description:
2529    
2530    Small corrections and updates to CM manual.
2531    
2532    ----------------------------------------------------------------------
2533    Name: Matthias Blume
2534    Date: 2000/06/29 16:04:00 JST
2535    Tag: blume-20000629-yacctool
2536    Description:
2537    
2538    Changes:
2539    
2540    1. Class "mlyacc" now takes separate arguments to pass options to
2541       generated .sml- and .sig-files independently.
2542    2. Corresponding CM manual updates.
2543    3. BTrace module now also reports call sites.  (However, for loop clusters
2544       it only shows from where the cluster was entered.)  There are associated
2545       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
2546    
2547    ----------------------------------------------------------------------
2548    Name: Matthias Blume
2549    Date: 2000/06/27 16:51:00 JST
2550    Tag: blume-20000627-noweb
2551    Description:
2552    
2553    Changes:
2554    
2555     1. Implemented "subdir" and "witness" options for noweb tool.
2556        This caused some slight internal changes in CM's tool implementation.
2557     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
2558        some remaining issues from earlier path anchor changes.
2559     3. Updated CM manual accordingly.
2560    
2561     4. Changed implementation of back-tracing so that I now consider it
2562        ready for prime-time.
2563    
2564        In particular, you don't have to explicitly trigger the back-trace
2565        anymore.  Instead, if you are running BTrace-instrumented code and
2566        there is an uncaught exception (regardless of whether or not it was
2567        raised in instrumented code), the top-level evalloop will print
2568        the back-trace.
2569    
2570        Features:
2571    
2572          - Instrumented and uninstrumented code work together seemlessly.
2573            (Of course, uninstrumented code is never mentioned in actual
2574             back-traces.)
2575    
2576          - Asymptotic time- and space-complexity of instrumented code is
2577            equal to that of uninstrumented code.  (This means that
2578            tail-recursion is preserved by the instrumentation phase.)
2579    
2580          - Modules whose code has been instrumented in different sessions
2581            work together without problem.
2582    
2583          - There is no penalty whatsoever on uninstrumented code.
2584    
2585          - There is no penalty on "raise" expressions, even in
2586            instrumented code.
2587    
2588        A potential bug (or perhaps it is a feature, too):
2589    
2590          A back-trace reaches no further than the outermost instrumented
2591          non-trivial "raise".  Here, a "trivial" raise is one that is the
2592          sole RHS of a "handle" rule.  Thus, back-traces reach trough
2593    
2594               <exp> handle e => raise e
2595    
2596          and even
2597    
2598               <exp> handle Foo => raise Bar
2599    
2600          and, of course, through
2601    
2602               <exp> handle Foo => ...
2603    
2604         if the exception was not Foo.
2605    
2606         Back-traces always reach right through any un-instrumented code
2607         including any of its "handle" expressions, trivial or not.
2608    
2609       To try this out, do the following:
2610    
2611         - Erase all existing binfiles for your program.
2612           (You may keep binfiles for those modules where you think you
2613            definitely don't need back-tracing.)
2614         - Turn on back-trace instrumentation:
2615              SMLofNJ.Internals.BTrace.mode (SOME true);
2616         - Recompile your program.  (I.e., run "CM.make" or "use".)
2617         - You may now turn instrumentation off again (if you want):
2618              SMLofNJ.Internals.BTrace.mode (SOME false);
2619         - Run your program as usual.  If it raises an exception that
2620           reaches the interactive toplevel, then a back-trace will
2621           automatically be printed.  After that, the toplevel loop
2622           will print the exception history as usual.
2623    
2624    ----------------------------------------------------------------------
2625    Name: Matthias Blume
2626    Date: 2000/06/26 09:56:46 JST
2627    Tag: blume-20000626-setup
2628    Description:
2629    
2630    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
2631          ML code before and after compiling a file (e.g., to set compiler
2632          flags)
2633    
2634    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
2635              - associated changes to btrace.sml (BTrace instrumentation pass)
2636              - cleaner implementation of btimp.sml (BTrace tracing and report
2637                module)
2638    
2639    CM manual: * new path encoding documented
2640               * description of setup-parameter to "sml" added
2641    
2642    The biggest user-visible change to back-tracing is that it is no
2643    longer necessary to compile all traced modules within the same
2644    session.  (This was a real limitation.)
2645    
2646    ----------------------------------------------------------------------
2647    Name: Matthias Blume
2648    Date: 2000/06/24 12:40:00 JST
2649    Tag: blume-20000624-startup
2650    Description:
2651    
2652    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
2653    bit too often -- to put it mildly. :)
2654    
2655    ----------------------------------------------------------------------
2656    Name: Matthias Blume
2657    Date: 2000/06/23 18:20:00 JST
2658    Tag: blume-20000623-btrace
2659    Description:
2660    
2661    This updates adds a backtrace facility to aid programmers in debugging
2662    their programs.  This involves the following changes:
2663    
2664    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
2665       keeping track of the current call stack.  When programs are compiled
2666       in a special mode, the compiler will insert calls to these hooks
2667       into the user program.
2668       "Hook" means that it is possible for different implementations of
2669       back-tracing to register themselves (at different times).
2670    
2671    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
2672       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
2673       It can be turned on using this call:
2674         SMLofNJ.Internals.BTrace.mode (SOME true);
2675       Turning it off again:
2676         SMLofNJ.Internals.BTrace.mode (SOME false);
2677       Querying the current status:
2678         SMLofNJ.Internals.BTrace.mode NONE;
2679       Annotated programs are about twice as big as normal ones, and they
2680       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
2681       where all hooks do nothing).  The slowdown with a plugin that is
2682       actually useful (such as the one supplied by default) is even greater,
2683       but in the case of the default plugin it is still only an constant
2684       factor (amortized).
2685    
2686    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
2687       with a sub-structure BTrace for controlling back-tracing.  In particular,
2688       the above-mentioned function "mode" controls whether the annotation
2689       phase is invoked by the compiler.  Another important function is
2690       "trigger": when called it aborts the current execution and causes
2691       the top-level loop to print a full back-trace.
2692    
2693    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
2694       for back-tracing.  It keeps track of the dynamic call stack and in
2695       addition to that it keeps a partial history at each "level" of that
2696       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
2697       d and b (at separate times, dynamically), then the report will show:
2698    
2699       GOTO   d
2700             /c
2701       GOTO  \b
2702       CALL   a
2703    
2704       This shows that there was an initial non-tail call of a, then a
2705       tail-call to b or c, looping behavior in a cluster of functions that
2706       consist of b and c, and then a goto from that cluster (i.e., either from
2707       b or from c) to d.
2708    
2709       Note that (depending on the user program) the amount of information
2710       that the back-trace module has to keep track of at each level is bounded
2711       by a constant.  Thus, the whole implementation has the same asymptotical
2712       complexity as the original program (both in space and in time).
2713    
2714    5. compiler/TopLevel/interact/evalloop.sml has been modified to
2715       handle the special exception SMLofNJ.Internals.BTrace.BTrace
2716       which is raised by the "trigger" function mentioned above.
2717    
2718    Notes on usage:
2719    
2720    - Annotated code works well together with unannotated code:
2721    Unannotated calls simply do not show up at all in the backtrace.
2722    
2723    - It is not a good idea to let modules that were annotated during
2724    different sessions run at the same time.  This is because the compiler
2725    chooses small integers to identify individual functions, and there
2726    will be clashes if different modules were compiled in separate sessions.
2727    (Nothing will crash, and you will even be told about the clashes, but
2728    back-trace information will in general not be useful.)
2729    
2730    - Back-tracing can be confused by callcc and capture.
2731    
2732    - The only way of getting a back-trace right now is to explicitly
2733    invoke the "trigger" function from your user program.  Eventually, we
2734    should make every exception carry back-trace information (if
2735    available).  But since this creates more overhead at "raise"-time
2736    (similar to the current exnHistory overhead), I have not yet
2737    implemented this.  (The implementation will be rather easy.)  With
2738    exceptions carrying back-trace information, this facility will be even
2739    more useful because users don't need to modify their programs...
2740    
2741    - While it is possible to compile the compiler with back-trace
2742    annotations turned on (I did it to get some confidence in
2743    correctness), you must make absolutely sure that core.sml and
2744    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
2745    be compiled with annotation because there is no core access yet, but
2746    if you compile btimp.sml with annotation, then the system will go into
2747    an infinite recursion and crash.)
2748    Since CM currently does not know about BTrace, the only way to turn
2749    annotations on and off for different modules of the compiler is to
2750    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
2751    this is awkward and clumsy.
2752    
2753    Sample sessions:
2754    
2755    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
2756    - SMLofNJ.Internals.BTrace.mode (SOME true);
2757    [autoloading]
2758    [autoloading done]
2759    val it = false : bool
2760    - structure X = struct
2761    -     fun main n = let
2762    -         fun a (x, 0) = d x
2763    -           | a (x, n) = b (x, n - 1)
2764    -         and b (x, n) = c (x, n)
2765    -         and c (x, n) = a (x, n)
2766    -         and d x = e (x, 3)
2767    -         and e (x, 0) = f x
2768    -           | e (x, n) = e (x, n - 1)
2769    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
2770    -           | f n = n * g (n - 1)
2771    -         and g n = a (n, 3)
2772    -     in
2773    -         f n
2774    -     end
2775    - end;
2776    structure X : sig val main : int -> int end
2777    - X.main 3;
2778    *** BACK-TRACE ***
2779    GOTO   stdIn:4.2-13.20: X.main[2].f
2780    GOTO-( stdIn:4.2-13.20: X.main[2].e
2781    GOTO   stdIn:4.2-13.20: X.main[2].d
2782         / stdIn:4.2-13.20: X.main[2].a
2783         | stdIn:4.2-13.20: X.main[2].b
2784    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2785    CALL   stdIn:4.2-13.20: X.main[2].g
2786    GOTO   stdIn:4.2-13.20: X.main[2].f
2787    GOTO-( stdIn:4.2-13.20: X.main[2].e
2788    GOTO   stdIn:4.2-13.20: X.main[2].d
2789         / stdIn:4.2-13.20: X.main[2].a
2790         | stdIn:4.2-13.20: X.main[2].b
2791    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2792    CALL   stdIn:4.2-13.20: X.main[2].g
2793    GOTO   stdIn:4.2-13.20: X.main[2].f
2794    GOTO-( stdIn:4.2-13.20: X.main[2].e
2795    GOTO   stdIn:4.2-13.20: X.main[2].d
2796         / stdIn:4.2-13.20: X.main[2].a
2797         | stdIn:4.2-13.20: X.main[2].b
2798    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2799    CALL   stdIn:4.2-13.20: X.main[2].g
2800    GOTO   stdIn:4.2-13.20: X.main[2].f
2801    CALL   stdIn:2.15-17.4: X.main[2]
2802    -
2803    
2804    (Note that because of a FLINt bug the above code currently does not
2805    compile without BTrace turned on.)
2806    
2807    Here is another example, using my modified Tiger compiler:
2808    
2809    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
2810    - SMLofNJ.Internals.BTrace.mode (SOME true);
2811    [autoloading]
2812    [autoloading done]
2813    val it = false : bool
2814    - CM.make "sources.cm";
2815    [autoloading]
2816    ...
2817    [autoloading done]
2818    [scanning sources.cm]
2819    [parsing (sources.cm):parse.sml]
2820    [creating directory CM/SKEL ...]
2821    [parsing (sources.cm):tiger.lex.sml]
2822    ...
2823    [wrote CM/sparc-unix/semant.sml]
2824    [compiling (sources.cm):main.sml]
2825    [wrote CM/sparc-unix/main.sml]
2826    [New bindings added.]
2827    val it = true : bool
2828    - Main.compile ("../testcases/merge.tig", "foo.out");
2829    *** BACK-TRACE ***
2830    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
2831    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2832    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2833    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2834    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2835    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2836    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
2837         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
2838    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
2839    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2840    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
2841    CALL   main.sml:1.18-118.4: Main.compile[2]
2842    -
2843    
2844    ----------------------------------------------------------------------
2845    Name: Matthias Blumen
2846    Date: 2000/06/21 18:00:00 JST
2847    Tag: blume-20000621-manual
2848    Description:
2849    
2850    CM manual update: Path environments documented.
2851    
2852    ----------------------------------------------------------------------
2853    Name: Matthias Blume
2854    Date: 2000/06/19 13:40:00
2855    Tag: blume-20000619-manual
2856    Description:
2857    
2858    CM manual and system/README update.  This only covers the fact that
2859    there are no more implicit anchors.  (Path environments and the "bind"
2860    option to "cm" have yet to be documented.)
2861    
2862    ----------------------------------------------------------------------
2863    Name: Matthias Blume
2864    Date: 2000/06/19 11:05:00 JST
2865    Tag: blume-20000619-chdir-bugfix
2866    Description:
2867    
2868    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
2869    
2870  ----------------------------------------------------------------------  ----------------------------------------------------------------------
2871  Name: Matthias Blume  Name: Matthias Blume
2872  Date: 2000/06/18 22:00:10 JST  Date: 2000/06/18 22:00:10 JST
# Line 942  Line 3798 
3798       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3799       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3800       Please, don't be too harsh on me for messing with this code a bit more       Please, don't be too harsh on me for messing with this code a bit more
3801       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3802       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3803       Dave's hands. :)       Dave's hands. :)
3804    

Legend:
Removed from v.670  
changed lines
  Added in v.1066

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