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

Legend:
Removed from v.587  
changed lines
  Added in v.905

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