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 572, Thu Mar 9 02:43:06 2000 UTC revision 932, Tue Sep 18 19:35:32 2001 UTC
# Line 14  Line 14 
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16  Name: Matthias Blume  Name: Matthias Blume
17    Date: 2001/09/18 15:35:00 EDT
18    Tag: blume-20010918-readme11036
19    Description:
20    
21    Added README files.
22    
23    ----------------------------------------------------------------------
24    Name: Matthias Blume
25    Date: 2001/09/18 11:45:00 EDT
26    Tag: Release_110_36 (retag)
27    Description:
28    
29    Fixed mistake in config/preloads. Retagged as 110.36.
30    
31    ----------------------------------------------------------------------
32    Name: Matthias Blume
33    Date: 2001/09/18 09:40:00 EDT
34    Tag: Release_110_36_orig (tag changed)
35    Description:
36    
37    New version (110.36).  New bootfiles.
38    
39    ----------------------------------------------------------------------
40    Name: Matthias Blume
41    Date: 2001/09/14 16:15:00 EDT
42    Tag: blume-20010914-x86fastfp
43    Description:
44    
45    John committed some changes that Allen made, in particular a (hopefully)
46    correctly working version of the x86-fp module.
47    
48    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
49    flag to "true".  Everything seems to compile to a fixpoint ok, and
50    "mandelbrot" speeds up by about 15%.
51    
52    ----------------------------------------------------------------------
53    Name: Matthias Blume
54    Date: 2001/09/13 11:20:00 EDT
55    Tag: blume-20010913-minimal
56    Description:
57    
58    1. Stefan Monnier's patch to fix a miscompilation problem that
59       was brought to light by John Reppy's work on Moby.
60    
61    2. Implemented a minimal "structure Compiler" that contains just
62       "version" and "architecture".  The minimal version will be
63       available when the full version is not.  This is for backward-
64       compatibility with code that wants to test Compiler.version.
65    
66    ----------------------------------------------------------------------
67    Name: Matthias Blume
68    Date: 2001/08/28 14:03:00 EDT
69    Tag: blume-20010828-ml-lex
70    Description:
71    
72    Fix for bug 1581, received from Neophytos Michael.
73    
74    ----------------------------------------------------------------------
75    Name: Matthias Blume
76    Date: 2001/08/27 11:20:00 EDT
77    Tag: blume-20010827-readme11035
78    Description:
79    
80    Fleshed out the README file for 110.35.
81    
82    ----------------------------------------------------------------------
83    Name: Matthias Blume
84    Date: 2001/08/24 17:10:00 EDT
85    Tag: Release_110_35
86    Description:
87    
88    New version number (110.35).  New bootfiles.
89    
90    ----------------------------------------------------------------------
91    Name: Lal George
92    Date: 2001/08/24 13:47:18 EDT 2001
93    Tag: george-20010824-MLRISC-graphs
94    Description:
95    
96     removed clusters from MLRISC completely and replaced with graphs.
97    
98    ----------------------------------------------------------------------
99    Name: Matthias Blume
100    Date: 2001/08/23 17:50:00 EDT
101    Tag: blume-20010823-toplevel
102    Description:
103    
104    - some reorganization of the code that implements various kinds of
105      environments in the compiler (static, dynamic, symbolic, combined)
106    - re-implemented the EnvRef module so that evalStream works properly
107      (if the stream contains references to "use", "CM.make", etc.)
108    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
109    
110    ----------------------------------------------------------------------
111    Name: Matthias Blume
112    Date: 2001/08/20 15:50 EDT
113    Tag: blume20010820-slipup
114    Description:
115    
116    I forgot to commit a few files.  Here they are...
117    
118    ----------------------------------------------------------------------
119    Name: Matthias Blume
120    Date: 2001/08/20 15:35:00 EDT
121    Tag: blume-20010820-debugprof
122    Description:
123    
124    !!!! NEW BOOTFILES !!!!
125    
126    This is another round of reorganizing the compiler sources.  This
127    time the main goal was to factor out all the "instrumentation"
128    passes (for profiling and backtracing) into their own library.
129    The difficulty was to do it in such a way that it does not depend
130    on elaborate.cm but only on elabdata.cm.
131    
132    Therefore there have been further changes to both elaborate.cm and
133    elabdata.cm -- more "generic" things have been moved from the former
134    to the latter.  As a result, I was forced to split the assignment
135    of numbers indicating "primtyc"s into two portions: SML-generic and
136    SML/NJ-specific.  Since it would have been awkward to maintain,
137    I bit the bullet and actually _changed_ the mapping between these
138    numbers and primtycs.  The bottom line of this is that you need
139    a new set of bin- and bootfiles.
140    
141    I have built new bootfiles for all architectures, so doing a fresh
142    checkout and config/install.sh should be all you need.
143    
144    The newly created library's name is
145    
146        $smlnj/viscomp/debugprof.cm
147    
148    and its sources live under
149    
150        src/compiler/DebugProf
151    
152    ----------------------------------------------------------------------
153    Name: Matthias Blume
154    Date: 2001/08/15 17:15:00 EDT
155    Tag: blume-20010815-compreorg
156    Description:
157    
158    This is a first cut at reorganizing the CM libraries that make up the
159    core of the compiler.  The idea is to separate out pieces that could
160    be used independently by tools, e.g., the parser, the typechecker, etc.
161    
162    The current status is a step in this direction, but it is not quite
163    satisfactory yet.  Expect more changes in the future.
164    
165    Here is the current (new) organization...
166    
167        What used to be $smlnj/viscomp/core.cm is now divided into
168        six CM libraries:
169    
170             $smlnj/viscomp/basics.cm
171                           /parser.cm
172                           /elabdata.cm
173                           /elaborate.cm
174                           /execute.cm
175                           /core.cm
176    
177        The CM files for these libraries live under src/system/smlnj/viscomp.
178        All these libraries are proxy libraries that contain precisely
179        one CM library component.  Here are the locations of the components
180        (all within the src/compiler tree):
181    
182             Basics/basics.cm
183             Parse/parser.cm
184             ElabData/elabdata.cm
185             Elaborator/elaborate.cm
186             Execution/execute.cm
187             core.cm
188    
189         [This organization is the same that has been used already
190         for a while for the architecture-specific parts of the visible
191         compiler and for the old version of core.cm.]
192    
193         As you will notice, many source files have been moved from their
194         respective original locations to a new home in one of the above
195         subtrees.
196    
197         The division of labor between the new libraries is the following:
198    
199             basics.cm:
200                - Simple, basic definitions that pertain to many (or all) of
201                  the other libraries.
202             parser.cm:
203                - The SML parser, producing output of type Ast.dec.
204                - The type family for Ast is also defined and exported here.
205             elabdata.cm:
206                - The datatypes that describe input and output of the elaborator.
207                  This includes types, absyn, and static environments.
208             elaborator.cm:
209                - The SML/NJ type checker and elaborator.
210                  This maps an Ast.dec (with a given static environment) to
211                  an Absyn.dec (with a new static environment).
212                - This libraries implements certain modules that used to be
213                  structures as functors (to remove dependencies on FLINT).
214             execute.cm:
215                - Everything having to do with executing binary code objects.
216                - Dynamic environments.
217             core.cm:
218                - SML/NJ-specific instantiations of the elaborator and MLRISC.
219                - Top-level modules.
220                - FLINT (this should eventually become its own library)
221    
222    Notes:
223    
224    I am not 100% happy with the way I separated the elaborator (and its
225    data structures) from FLINT.  Two instances of the same problem:
226    
227        1. Data structures contain certain fields that carry FLINT-specific
228           information.  I hacked around this using exn and the property list
229           module from smlnj-lib.  But the fact that there are middle-end
230           specific fields around at all is a bit annoying.
231    
232        2. The elaborator calculates certain FLINT-related information.  I tried
233           to make this as abstract as I could using functorization, but, again,
234           the fact that the elaborator has to perform calculations on behalf
235           of the middle-end at all is not nice.
236    
237        3. Having to used exn and property lists is unfortunate because it
238           weakens type checking.  The other alternative (parameterizing
239           nearly *everything*) is not appealing, though.
240    
241    I removed the "rebinding =" warning hack because due to the new organization
242    it was awkward to maintain it.  As a result, the compiler now issues some of
243    these warnings when compiling init.cmi during bootstrap compilation. On
244    the plus side, you also get a warning when you do, for example:
245       val op = = Int32.+
246    which was not the case up to now.
247    
248    I placed "assign" and "deref" into the _Core structure so that the
249    code that deals with the "lazy" keyword can find them there.  This
250    removes the need for having access to the primitive environment
251    during elaboration.
252    
253    ----------------------------------------------------------------------
254    Name: Matthias Blume
255    Date: 2001/08/13
256    Tag: blume-20010813-closures
257    Description:
258    
259    This fix was sent to us by Zhong Shao.  It is supposed to improve the
260    performance of certain loops by avoiding needless closure allocation.
261    
262    ----------------------------------------------------------------------
263    Name: Lal George
264    Date: 2001/07/31 10:03:23 EDT 2001
265    Tag: george-20010731-x86-fmalloc
266    Description: Fixed bug in x86 calls
267    
268        There was a bug where call instructions would mysteriously
269        vanish. The call instruction had to be one that returned
270        a floating point value.
271    
272    ----------------------------------------------------------------------
273    Name: Lal George
274    Date: 2001/07/19 16:36:29 EDT 2001
275    Tag: george-20010719-simple-cells
276    Description:
277    
278    I have dramatically simplified the interface for CELLS in MLRISC.
279    
280    In summary, the cells interface is broken up into three parts:
281    
282      1. CellsBasis : CELLS_BASIS
283    
284            CellsBasis is a top level structure and common for all
285            architectures.  it contains the definitions of basic datatypes
286            and utility  functions over these types.
287    
288      2. functor Cells() : CELLS
289    
290            Cells generates an interface for CELLS that incorporates the
291            specific resources on the target architecture, such as the
292            presence of special register classes, their number and size,
293            and various useful substructures.
294    
295      3. <ARCH>CELLS
296    
297            e.g. SparcCells: SPARCCELLS
298    
299            <ARCH>CELLS usually contains additional bindings for special
300            registers  on the architecture, such as:
301    
302                    val r0 : cell           (* register zero *)
303                    val y : cell            (* Y register *)
304                    val psr : cell          (* processor status register *)
305                    ...
306    
307            The structure returned by applying the Cells functor is opened
308            in this interface.
309    
310    The main implication of all this is that the datatypes for cells is
311    split between CellsBasis and CELLS -- a fairly simple change for user
312    code.
313    
314    In the old scheme the CELLS interface had a definitional binding of
315    the form:
316    
317            signature CELLS = sig
318    
319               structure CellsBasis = CellsBasis
320    
321               ...
322    
323            end
324    
325    With all the sharing constraints that goes on in MLRISC, this old
326    design  quickly leads to errors such as:
327    
328            "structure definition spec inside of sharing ... "
329    
330    
331    and appears to require an unacceptable amount of sharing and where
332    constraint hackery.
333    
334    I think this error message (the interaction of definitional specs and
335    sharing) requires more explanation on our web page.
336    
337    ----------------------------------------------------------------------
338    Name: Matthias Blume
339    Date: 2001/07/19 15:00:00 EDT
340    Tag: blume-20010719-libreorg
341    Description:
342    
343    This update puts together a fairly extensive but straightforward change
344    to the way the libraries that implement the interactive system are
345    organized:
346    
347       The biggest change is the elimination of structure Compiler.  As a
348       replacement for this structure, there is now a CM library
349       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
350       that exports all the substructures of the original structure Compiler
351       directly.  So instead of saying Compiler.Foo.bar one now simply
352       says Foo.bar.  (The CM libraries actually export a collection of
353       structures that is richer than the collection of substructures of
354       structure Compiler.)
355    
356       To make the transition smooth, there is a separate library called
357       $smlnj/compiler/compiler.cm which puts together and exports the
358       original structure Compiler (or at least something very close to it).
359    
360       There are five members of the original structure Compiler
361       that are not exported directly but which instead became members
362       of a new structure Backend (described by signature BACKEND).  These are:
363       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
364       Interact (: INTERACT), structure Machine (: MACHINE), and val
365       architecture (: string).
366    
367       Structure Compiler.Version has become structure CompilerVersion.
368    
369       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
370       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
371       or x86, respectively.
372       Each of these exports the same frontend structures that
373       $smlnj/compiler.cm exports.  But they do not have a structure Backend
374       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
375       Hppa, PPC, Sparc, or X86, respectively.
376    
377       Library $smlnj/compiler/all.cm exports the union of the exports of
378       $smlnj/compiler/<arch>.cm
379    
380       There are no structures <Arch>Compiler anymore, use
381       $smlnj/compiler/<arch>.cm instead.
382    
383       Library host-compiler-0.cm is gone.  Instead, the internal library
384       that instantiates CM is now called cm0.cm.  Selection of the host
385       compiler (backend) is no longer done here but. (Responsibility for it
386       now lies with $smlnj/compiler/current.cm.  This seems to be more
387       logical.)
388    
389       Many individual files have been moved or renamed.  Some files have
390       been split into multiple files, and some "dead" files have been deleted.
391    
392    Aside from these changes to library organization, there are also changes
393    to the way the code itself is organized:
394    
395       Structure Binfile has been re-implemented in such a way that it no
396       longer needs any knowledge of the compiler.  It exclusively deals
397       with the details of binfile layout.  It no longer invokes the
398       compiler (for the purpose of creating new prospective binfile
399       content), and it no longer has any knowledge of how to interpret
400       pickles.
401    
402       Structure Compile (: COMPILE) has been stripped down to the bare
403       essentials of compilation.  It no longer deals with linking/execution.
404       The interface has been cleaned up considerably.
405    
406       Utility routines for dealing with linking and execution have been
407       moved into their own substructures.
408    
409       (The ultimate goal of these changes is to provide a light-weight
410       binfile loader/linker (at least for, e.g., stable libraries) that
411       does not require CM or the compiler to be present.)
412    
413    CM documentation has been updated to reflect the changes to library
414    organization.
415    
416    ----------------------------------------------------------------------
417    Name: Matthias Blume
418    Date: 2001/07/10 17:30:00 EDT
419    Tag: Release_110_34
420    Description:
421    
422    Minor tweak to 110.34 (re-tagged):
423    
424      - README.html file added to CVS repository
425      - runtime compiles properly under FreeBSD 3.X and 4.X
426    
427    ----------------------------------------------------------------------
428    Name: Matthias Blume
429    Date: 2001/07/10 17:30:00 EDT
430    Tag: Release_110_34
431    Description:
432    
433    New version number (110.34). New bootfiles.
434    
435    ----------------------------------------------------------------------
436    Name: Matthias Blume
437    Date: 2001/07/09 16:00:00 EDT
438    Tag: blume-20010709-more-varargs
439    Description:
440    
441    I changed the handling of varargs in ml-nlffigen again:
442    The ellipsis ... will now simply be ignored (with an accompanying warning).
443    
444    The immediate effect is that you can actually call a varargs function
445    from ML -- but you can't actually supply any arguments beyond the ones
446    specified explicitly.  (For example, you can call printf with its format
447    string, but you cannot pass additional arguments.)
448    
449    This behavior is only marginally more useful than the one before, but
450    it has the advantage that a function or, more importantly, a function
451    type never gets dropped on the floor, thus avoiding follow-up problems with
452    other types that refer to the offending one.
453    
454    ----------------------------------------------------------------------
455    Name: Matthias Blume
456    Date: 2001/07/09 11:25:00 EDT
457    Tag: blume-20010709-varargs
458    Description:
459    
460    1. ckit-lib.cm now exports structure Error
461    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
462       with a warning accompanied by a source location.  Moreover, it
463       merely skips the offending function or type and proceeds with the
464       rest of its work.u  As a result, one can safely feed C code containing
465       "..." to ml-nlffigen.
466    3. There are some internal improvements to CM, providing slightly
467       more general string substitutions in the tools subsystem.
468    
469    ----------------------------------------------------------------------
470    Name: Matthias Blume
471    Date: 2001/06/27 15:10:00 EDT
472    Tag: blume-20010627-concur
473    Description:
474    
475    Fixed a small bug in CM's handling of parallel compilation.
476    (You could observe the bug by Control-C-interrupting an ordinary
477    CMB.make or CM.stabilize and then attaching some compile servers.
478    The result was that all of a sudden the previously interrupted
479    compilation would continue on its own.  This was because of
480    an over-optimization: CM did not bother to clean out certain queues
481    when no servers were attached "anyway", resulting in the contents
482    of these queues to grab control when new servers did get attached.)
483    
484    There is also another minor update to the CM manual.
485    
486    ----------------------------------------------------------------------
487    Name: Matthias Blume
488    Date: 2001/06/26 16:15:00 EDT
489    Tag: blume-20010626-cmdoc
490    Description:
491    
492    Minor typo fixed in CM manual (syntax diagram for libraries).
493    
494    ----------------------------------------------------------------------
495    Name: Matthias Blume
496    Date: 2001/06/25 22:55:00 EDT
497    Tag: blume-20010625-x86pc
498    Description:
499    
500    Fixed a nasty bug in the X86 assembly code that caused signal
501    handlers to fail (crash) randomly.
502    
503    ----------------------------------------------------------------------
504    Name: Matthias Blume
505    Date: 2001/06/25 12:05:00 EDT
506    Tag: blume-20010625-nlffigen
507    Description:
508    
509    This update fixes a number of minor bugs in ml-nlffigen as reported by
510    Nick Carter <nbc@andrew.cmu.edu>.
511    
512      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
513      2. Default names for generated files are now derived from the name of
514         the C file *without its directory*.  In particular, this causes generated
515         files to be placed locally even if the C file is in some system directory.
516      3. Default names for generated signatures and structures are also derived
517         from the C file name without its directory.  This avoids silly things
518         like "structure GL/GL".
519         (Other silly names are still possible because ml-nlffigen does not do
520          a thorough check of whether generated names are legal ML identifiers.
521          When in doubt, use command line arguments to force particular names.)
522    
523    ----------------------------------------------------------------------
524    Name: Matthias Blume
525    Date: 2001/06/21 12:25:00 EDT
526    Tag: blume-20010621-eXene
527    Description:
528    
529    eXene now compiles and (sort of) works again.
530    
531    The library name (for version > 110.33) is $/eXene.cm.
532    
533    I also added an new example in src/eXene/examples/nbody.  See the
534    README file there for details.
535    
536    ----------------------------------------------------------------------
537    Name: Matthias Blume
538    Date: 2001/06/20 16:40:00 EDT
539    Tag: blume-20010620-cml
540    Description:
541    
542    CML now compiles and works again.
543    
544    Libraries (for version > 110.33):
545    
546      $cml/cml.cm            Main CML library.
547      $cml/basis.cm          CML's version of $/basis.cm.
548      $cml/cml-internal.cm   Internal helper library.
549      $cml/core-cml.cm       Internal helper library.
550      $cml-lib/trace-cml.cm  Tracing facility.
551      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
552    
553    The installer (config/install.sh) has been taught how to properly
554    install this stuff.
555    
556    ----------------------------------------------------------------------
557    Name: Matthias Blume
558    Date: 2001/06/19 17:55:00 EDT
559    Tag: blume-20010619-instantiate
560    Description:
561    
562    This un-breaks the fix for bug 1432.
563    (The bug was originally fixed in 110.9 but I broke it again some
564    time after that.)
565    
566    ----------------------------------------------------------------------
567    Name: Matthias Blume
568    Date: 2001/06/19 17:25:00 EDT
569    Tag: blume-20010619-signals
570    Description:
571    
572    This should (hopefully) fix the long-standing signal handling bug.
573    (The runtime system was constructing a continuation record with an
574    incorrect descriptor which would cause the GC to drop data on the floor...)
575    
576    ----------------------------------------------------------------------
577    Name: Matthias Blume
578    Date: 2001/06/15 15:05:00 EDT
579    Tag: blume-20010615-moresparc
580    Description:
581    
582    Here is a short late-hour update related to Sparc c-calls:
583    
584     -- made handling of double-word arguments a bit smarter
585    
586     -- instruction selection phase tries to collapse certain clumsily
587        constructed ML-Trees; typical example:
588    
589            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
590    
591        This currently has no further impact on SML/NJ since mlriscGen does
592        not seem to generate such patterns in the first place, and c-calls
593        (which did generate them in the beginning) has meanwhile been fixed
594        so as to avoid them as well.
595    
596    ----------------------------------------------------------------------
597    Name: Matthias Blume
598    Date: 2001/06/15 15:05:00 EDT
599    Tag: blume-20010615-sparc
600    Description:
601    
602    The purpose of this update is to provide an implementation of NLFFI
603    on Sparc machines.
604    
605    Here are the changes in detail:
606    
607       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
608       the Sparc implementation of the c-calls API.
609       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
610       for accessing the ML frame.  Thus, we have a real frame pointer and
611       can freely modify %sp without need for an omit-frame-ptr phase.
612       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
613       this case.
614       * ml-nlffigen has been taught to produce code for different architectures
615       and calling conventions.
616       * In a way similar to what was done in the x86 case, the Sparc
617       backend uses its own specific extension to mltree.  (For example,
618       it needs to be able to generate UNIMP instructions which are part
619       of the calling convention.)
620       * ml-nlffi-lib was reorganized to make it more modular (in particular,
621       to make it easier to plug in new machine- and os-dependent parts).
622    
623    There are some other fairly unrelated bug fixes and cleanups as well:
624    
625       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
626       that they properly share their libraries with existing SML/NJ libraries.
627       * I fixed a minor cosmetic bug in CM, supressing certain spurious
628       follow-up error messages.
629       * Updates to CM/CMB documentation.
630    
631    TODO items:
632    
633       * MLRISC should use a different register as its asmTemp on the Sparc.
634         (The current %o2 is a really bad choice because it is part of the
635         calling conventions, so things might interfere in unexpected ways.)
636    
637    ----------------------------------------------------------------------
638    Name: Matthias Blume
639    Date: 2001/06/07
640    Tag: blume-20010607-calls
641    Description:
642    
643    A number of internal changes related to C calls and calling conventions:
644    
645    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
646       number of bytes popped implicitly (by the callee).  In most cases
647       this field is 0 but on x86/win32 it is some non-zero value.  This
648       is information provided for the benefit of the "omit-frameptr" pass.
649    2. The CALL instruction on the x86 carries a similar "pops" field.
650       The instruction selection phase copies its value from the ML-Tree
651       CALL statement.
652    3. On all other architectures, the instruction selection phase checks
653       whether "pops=0" and complains if not.
654    4. The c-calls implementation for x86 now accepts two calling conventions:
655       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
656       up after the call and pops is set to 0.  For "stdcall", the caller
657       does nothing, leaving the cleanup to the callee; pops is set to
658       the number of bytes that were pushed onto the stack.
659    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
660       distinguish between "ccall" and "stdcall".
661    6. The UNIMP instruction has been added to the supported Sparc instruction
662       set. (This is needed for implementing the official C calling convention
663       on this architecture.)
664    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
665       work with the latest CM.
666    
667    ----------------------------------------------------------------------
668    Name: Matthias Blume
669    Date: 2001/06/05 15:10:00 EDT
670    Tag: blume-20010605-cm-index
671    Description:
672    
673    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
674    
675    1. CM can now generate "index files".  These are human-readable files
676       that list on a per-.cm-file basis each toplevel symbol defined or
677       imported.  The location of the index file for
678       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
679       To enable index-file generation, set CM.Control.generate_index to true
680       or export an environment-symbol: export CM_GENERATE_INDEX=true.
681    
682       The CM manual has been updated accordingly.
683    
684    2. I made some slight modifications to the c-calls API in MLRISC.
685    
686         a) There is now a callback to support saving/restoring of
687            dedicated but caller-save registers around the actual call
688            instruction.
689         b) One can optionally specify a comment-annotation for the
690            call instruction.
691    
692    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
693       (For example, the comment annotation shows the C prototype of
694        the function being called.)
695    
696    ----------------------------------------------------------------------
697    Name: Matthias Blume
698    Date: 2001/06/01 13:30:00 EDT
699    Tag: blume-20010601-nlffi-cleanup
700    Description:
701    
702    This is mostly a cleanup of MLFFI stuff:
703    
704       - some signature files have been put into a more exposed place
705       - the ugly 'f type parameter is gone (simplifies types tremendously!)
706       - ml-nlffigen changed accordingly
707       - tutorial updated
708    
709    Other changes:
710    
711       - author's affiliation in CM manual(s) updated
712       - some more recognized keywords added to Allen's sml.sty
713    
714    ----------------------------------------------------------------------
715    Name: Matthias Blume
716    Date: 2001/05/25 15:30:00 EDT
717    Tag: blume-20010525-iptr
718    Description:
719    
720      - put the official 110.33-README (as it appears on the ftp server) under
721        CVS
722      - fixed a small bug related to incomplete pointer types in
723        ml-nlffigen
724      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
725        (it does not need the 'f type parameter)
726    
727    ----------------------------------------------------------------------
728    Name: Matthias Blume
729    Date: 2001/05/23 14:30:00 EDT
730    Tag: Release_110_33
731    Description:
732    
733    New version number (110.33).  New bootfiles.
734    
735    ----------------------------------------------------------------------
736    Name: Matthias Blume
737    Date: 2001/05/22 18:06:00 EDT
738    Tag: blume-20010522-targets
739    Description:
740    
741    Made install.sh use file config/targets.customized if it exists, falling
742    back to config/targets if it doesn't.  This way one can have a customized
743    version of the targets file without touching the "real thing", thus
744    eliminating the constant fear of accidentally checking something bogus
745    back into the CVS repository...  (File config/targets.customized must
746    not be added to the repository!)
747    
748    ----------------------------------------------------------------------
749    Name: Matthias Blume
750    Date: 2001/05/22 16:30:00 EDT
751    Tag: blume-20010522-minitut
752    Description:
753    
754    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
755       struct returns.
756    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
757       incomplete, preliminary documentation for NLFFI.
758    
759    ----------------------------------------------------------------------
760    Name: Matthias Blume
761    Date: 2001/05/14 11:30:00 EDT
762    Tag: blume-20010514-script
763    Description:
764    
765    Some bugs in install script fixed.
766    
767    In addition to that I also made a slight change to the NLFFI API:
768    Functors generated by ml-nlffigen now take the dynamic library as a
769    straight functor argument, not as a suspended one.  (The original
770    functor code used to force the suspension right away anyway, so there
771    was nothing gained by this complication of the interface.)
772    
773    ----------------------------------------------------------------------
774    Name: Matthias Blume
775    Date: 2001/05/11 14:35:00 EDT
776    Tag: blume-20010511-ml-nlffi
777    Description:
778    
779    I finally took the plunge and added my new FFI code to the main
780    repository.  For x86-linux it is now ready for prime-time.
781    
782    There are two new subdirectories of "src":
783    
784      - ml-nlffi-lib:
785           The utility library for programs using the FFI interface.
786           Here is the implementation of $/c.cm and its associated low-level
787           partners $/c-int.cm and $/memory.cm.
788      - ml-nlffigen:
789           A stand-alone program for generating ML glue code from C source
790           code.
791    
792    Building ml-nlffigen requires $/ckit-lib.cm.
793    
794    The config/install.sh script has been updates to do the Right Thing
795    (hopefully).
796    
797    Notice that the source tree for the C-Kit will not be put under "src"
798    but directly under the installation root directory.  (This is the
799    structure that currently exists on the CVS server when you check out
800    module "sml".)  Fortunately, config/install.sh knows about this oddity.
801    
802    Bugs: No documentation yet.
803    
804    ----------------------------------------------------------------------
805    Name: Matthias Blume
806    Date: 2001/05/09 16:35:00 EDT
807    Tag: blume-20010509-cpscontract
808    Description:
809    
810    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
811    wrapper/unwrapper elimination did not decrement usage counts and some
812    dead variables got overlooked by the dead-up logic.)
813    
814    ----------------------------------------------------------------------
815    Name: Lal George
816    Date: 2001/05/08  17:26:09 EDT
817    Tag: george-20010508-omit-frameptr
818    Description:
819    
820    Changes to implement the omit-frame-pointer optimization to support
821    raw C calls. For now, there is only support on the Intel x86, but
822    other architectures will follow as more experience is gained with this.
823    
824    
825    ----------------------------------------------------------------------
826    Name: Matthias Blume
827    Date: 2001/05/07 14:40:00 EDT
828    Tag: blume-20010507-proxies
829    Description:
830    
831    I made into "proxy libraries" all libraries that qualify for such a
832    change.  (A qualifying library is a library that has another library or
833    groups as its sole member and repeats that member's export list
834    verbatim.  A proxy library avoids this repetition by omitting its export
835    list, effectively inheriting the list that its (only) member exports.
836    See the CM manual for more explanation.)
837    The main effect is that explicit export lists for these libraries
838    do not have to be kepts in sync, making maintenance a bit easier.
839    
840    I also added copyright notices to many .cm-files.
841    
842    Last but not least, I made a new set of bootfiles.
843    
844    ----------------------------------------------------------------------
845    Name: Matthias Blume
846    Date: 2001/05/04 17:00:00 EDT
847    Tag: blume-20010504-cm-lsplit
848    Description:
849    
850    0. John merged pending changes to $/smlnj-lib.cm
851    
852    1. Allen's previous change accidentally backed out of one of Lal's
853       earlier changes.  I undid this mistake (re-introducing Lal's change).
854    
855    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
856       within the compiler where applicable.  There is some code simplification
857       because of that.
858    
859    3. The "split" phase (in FLINT) is now part of the default list of phases.
860       Compiler.Control.LambdaSplitting.* can be used to globally control the
861       lambda-splitting (cross-module-inlining) engine.  In addition to that,
862       it can now also be controlled on a per-source basis: CM has been taught
863       a new tool parameter applicable to ML source files.
864    
865       - To turn lambda-splitting off completely:
866            local open Compiler.Control.LambdaSplitting in
867                val _ = set Off
868            end
869       - To make "no lambda-splitting" the global default (but allow per-source
870         overriding); this is the initial setting:
871            local open Compiler.Control.LambdaSplitting in
872                val _ = set (Default NONE)
873            end
874       - To make "lambda-splitting with aggressiveness a" the global default
875         (and allow per-source overriding):
876            local open Compiler.Control.LambdaSplitting in
877                val _ = set (Default (SOME a))
878            end
879    
880       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
881         write (in the respective .cm-file):
882            a.sml (lambdasplitting:off)
883       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
884            a.sml (lambdasplitting:on)
885       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
886         <a> is a decimal non-negative integer):
887            a.sml (lambdasplitting:<a>)
888       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
889            a.sml (lambdasplitting:infinity)
890       - To use the global default for a.sml:
891            a.sml (lambdasplitting:default)
892         or simply
893            a.sml
894    
895    ----------------------------------------------------------------------
896    Name: Allen Leung
897    Date: 2001/05/04 01:57:00 EDT
898    Tag: leunga-20010504-sync
899    Description:
900    
901      MLRISC features.
902    
903      1. Fix to CMPXCHG instructions.
904      2. Changed RA interface to allow annotations in callbacks.
905      3. Added a new method to the stream interface to allow annotations updates.
906    
907    ----------------------------------------------------------------------
908    Name: Matthias Blume
909    Date: 2001/05/01 11:45:00 EDT
910    Tag: blume-20010501-pcedittmp
911    Description:
912    
913    Changed install.sh to use the current working directory instead of
914    /usr/tmp for a temporary file (pcedittmp).  The previous choice
915    of /usr/tmp caused trouble with MacOS X because of file premission
916    problems.
917    
918    ----------------------------------------------------------------------
919    Name: Matthias Blume
920    Date: 2001/04/20 11:10:00 EDT
921    Tag: blume-20010420-inMLflag
922    Description:
923    
924     - added vp_limitPtrMask to vproc-state.h
925       (for use by the raw-C-calls mechanism to implement proper interrupt
926        handling)
927     - made the ML compiler aware of various data-structure offsets so it
928       can generate code for accessing the vp_inML flag and vp_limitPtrMask
929     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
930       raw C-calls
931    
932    ----------------------------------------------------------------------
933    Name: Lal George
934    Date: 2001/04/20 09:15:28 EDT
935    Tag: george-20010420-macosX
936    Description:
937    
938     - Changes to port to Mac OS X; Darwin.
939    
940     - In the process I found that sqrt was broken on the PPC, because the
941       fsqrt instruction is not implemented.
942    
943    ----------------------------------------------------------------------
944    Name: Matthias Blume
945    Date: 2001/04/18 12:45:00 EDT
946    Tag: blume-20010418-ccalls
947    Description:
948    
949     - fixed two off-by-4 errors in the x86-specific c-calls implementation
950       (this bug prevented structure arguments containing pointers from being
951        passed correctly)
952     - changed the raw-C-call code in mlriscGen.sml in such a way that
953       structure arguments are represented as a pointer to the beginning
954       of the structure (instead of having a series of synthesized arguments,
955       one for each structure member)
956    
957     - made makeml script's verbosity level configurable via environment
958       variable (MAKEML_VERBOSITY)
959    
960     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
961       in rawmem-x86.sml; we are now using the real thing
962    
963    ----------------------------------------------------------------------
964    Name: Matthias Blume
965    Date: 2001/03/22 16:25:00 EST
966    Tag: blume-20010322-bootfiles
967    Description:
968    
969    Created a new set of bootfiles (for your automatic installation convenience).
970    
971    ----------------------------------------------------------------------
972    Name: Matthias Blume
973    Date: 2001/03/22 15:10:00 EST
974    Tag: blume-20010322-rawmem-parcm
975    Description:
976    
977    1. All "raw memory access" primitives for the new FFI are implemented now
978       (at least on the x86).
979    2. Some further cleanup of CM's parallel make mechanism.
980    
981    ----------------------------------------------------------------------
982    Name: Matthias Blume
983    Date: 2001/03/19 17:53:00 EST
984    Tag: blume-20010319-parallel
985    Description:
986    
987    Parallel make (using compile servers) now works again.
988    
989    To this end, CM.stabilize and CMB.make have been modified to work in
990    two passes when compile servers are attached:
991       1. Compile everything, do not perform stabilization; this pass
992          uses compile servers
993       2. Stabilize everything; this pass does not use compile servers
994    If there are no compile servers, the two passes are combined into one
995    (as before).  Splitting the passes increases the inherent parallelism
996    in the dependency graph because the entire graph including all
997    libraries is available at the same time.  This, in turn, improves
998    server utilization.  The downside is that the master process will
999    have to do some extra work after compilation is done (because for
1000    technical reasons it must re-read all the binfiles during stabilization).
1001    
1002    ----------------------------------------------------------------------
1003    Name: Matthias Blume
1004    Date: 2001/03/16 12:22:00 EST
1005    Tag: blume-20010316-bootfiles
1006    Description:
1007    
1008    Created a new set of bootfiles (for your automatic installation convenience).
1009    
1010    ----------------------------------------------------------------------
1011    Name: Matthias Blume
1012    Date: 2001/03/16 11:00:00 EST
1013    Tag: blume-20010316-MLTREE-fixup
1014    Description:
1015    
1016    This is a minor fixup for an (untagged) earlier commit by Allen.
1017    (A file was missing).
1018    
1019    ----------------------------------------------------------------------
1020    Name: Allen Leung
1021    Date: Mon Mar  5 18:54:57 EST 2001
1022    Tag: leunga-20010305-cut-support
1023    
1024    1. New support for alternative control-flow in MLTREE.
1025       Currently we support
1026    
1027          FLOW_TO(CALL ...., [k1,...,kn])
1028    
1029       This is needed for 'cuts to' in C-- and try/handle-like constructs
1030       in Moby
1031    
1032       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
1033    
1034    2. Register Allocator
1035    
1036       Changes in interface [from Fermin, John]
1037    
1038    3. Alpha 8-bit SLL support [Fermin]
1039    
1040    4. All architectures
1041    
1042       A new module (ClusterExpandCopies) for expanding parallel copies.
1043    
1044    ----------------------------------------------------------------------
1045    Name: Allen Leung
1046    Date: 2001/02/27 23:07:00 EST
1047    Tag: leunga-20010227-minor-stuff
1048    
1049    1. Alpha bug fix for CMOVNE
1050    2. Handle mltree COND(..,FCMP ...,...)
1051    3. Bug fix in simplifier
1052    
1053    ----------------------------------------------------------------------
1054    Name: Matthias Blume
1055    Date: 2001/01/30 17:50:00 EST
1056    Tag: blume-20010130-sync
1057    Description:
1058    
1059    This is just a minor update to sync my devel branch with the main brach.
1060    The only visible change is the addition of some README files.
1061    
1062    ----------------------------------------------------------------------
1063    Name: Matthias Blume
1064    Date: 2001/01/12 23:30:00 JST
1065    Tag: blume-20010112-bootfiles
1066    Description:
1067    
1068    Made a new set of bootfiles that goes with the current state of the
1069    repository.
1070    
1071    ----------------------------------------------------------------------
1072    Name: Matthias Blume
1073    Date: 2001/01/12 21:20:00 JST
1074    Tag: blume-20010112-sync
1075    Description:
1076    
1077    I am just flushing out some minor changes that had accumulated in
1078    my private branch in order to sync with the main tree.  (This is
1079    mainly because I had CVS trouble when trying to merge _into_ my
1080    private branch.)
1081    
1082    Most people should be completely unaffected by this.
1083    
1084    ----------------------------------------------------------------------
1085    Name: Allen Leung
1086    Date: Thu Jan 11 21:03:00 EST 2001
1087    Tag: leunga-20010111-labexp=mltree
1088    Description:
1089    
1090    1.  Removed the type LabelExp and replace it by MLTree.
1091    2.  Rewritten mltree-simplify with the pattern matcher tool.
1092    3.  There were some bugs in alpha code generator which would break
1093        64-bit code generation.
1094    4.  Redo the tools to generate code with the
1095    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
1096        are now generated by perl scripts.
1097    
1098    ----------------------------------------------------------------------
1099    Name: Matthias Blume
1100    Date: 2001/01/10 21:55:00 JST
1101    Tag: blume-20010110-rcc
1102    Description:
1103    
1104    The RCC stuff now seems to work (but only on the x86).
1105    This required hacking of the c-calls interface (and -implementation) in
1106    MLRISC.
1107    
1108    Normal compiler users should be unaffected.
1109    
1110    ----------------------------------------------------------------------
1111    Name: Matthias Blume
1112    Date: 2001/01/09 01:20:00 JST
1113    Tag: blume-20010109-rcc
1114    Description:
1115    
1116    This is a fairly big patch, flushing out a large number of pending
1117    changes that I made to my development copy over the last couple of days.
1118    
1119    Of practical relevance at this moment is a workaround for a pickling
1120    bug that Allen ran into the other day.  The cause of the bug itself is
1121    still unknown and it might be hard to fix it properly, but the
1122    workaround has some merits of its own (namely somewhat reducing pickling
1123    overhead for certain libraries).  Therefore, I think this solution should
1124    be satisfactory at this time.
1125    
1126    The rest of the changes (i.e., the vast majority) has to do with my
1127    ongoing efforts of providing direct support for C function calls from
1128    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
1129    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
1130    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
1131    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
1132    handle RCC.
1133    
1134    The last part (the changes to mlriscGen.sml) are still known to be
1135    wrong on the x86 and not implemented on all other architectures.  But
1136    the infrastructure is in place. I had to change a few functor
1137    signatures in the backend to be able to route the CCalls interface
1138    from MLRISC there, and I had to specialize the mltree type (on the
1139    x86) to include the necessary extensions. (The extensions themselves
1140    were already there and redy to go in MLRISC/x86).
1141    
1142    Everything should be very happy as soon as someone helps me with
1143    mlriscGen.sml...
1144    
1145    In any case, nothing of this should matter to anyone as long as the
1146    new primop is not being used (which is going to be the case unless you
1147    find it where I hid it :). The rest of the compiler is completely
1148    unaffected.
1149    
1150    ----------------------------------------------------------------------
1151    Name: Matthias Blume
1152    Date: 2001/01/05 00:30:00 JST
1153    Tag: blume-20010105-primops
1154    Description:
1155    
1156    Added some experimental support for work that I am doing right now.
1157    These changes mostly concern added primops, but there is also a new
1158    experimental C library in the runtime system (but currently not enabled
1159    anywhere except on Linux/X86).
1160    
1161    In the course of adding primops (and playing with them), I discovered that
1162    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
1163    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
1164    To recover, I made type information in INL_PRIM mandatory and changed
1165    prim.sml as well as built-in.sml accordingly.  The InLine structure now
1166    has complete, correct type information (i.e., no bottom types).
1167    
1168    Since all these changes mean that we need new binfiles, I also bumped the
1169    version number to 110.32.1.
1170    
1171    ----------------------------------------------------------------------
1172    Name: Matthias Blume
1173    Date: 2000/12/30 22:10:00 JST
1174    Tag: blume-20001230-various
1175    Description:
1176    
1177    Added proxy libraries for MLRISC and let MLRISC libraries refer
1178    to each other using path anchors.  (See CM manual for explanation.)
1179    
1180    Updated CM documentation.
1181    
1182    Fixed some bugs in CM.
1183    
1184    Implemented "proxy" libraries (= syntactic sugar for CM).
1185    
1186    Added "-quiet" option to makeml and changed runtime system accordingly.
1187    
1188    Added cleanup handler for exportML to reset timers and compiler stats.
1189    
1190    ----------------------------------------------------------------------
1191    Name: Lal George
1192    Date: 2000/12/22 22:22:58 EST 2000
1193    Tag: Release_110_32
1194    Description:
1195    
1196            Infinite precision used throughout MLRISC.
1197            see MLRISC/mltree/machine-int.sig
1198    
1199    ----------------------------------------------------------------------
1200    Name: Matthias Blume
1201    Date: 2000/12/22 23:16:00 JST
1202    Tag: blume-20001222-warn
1203    Description:
1204    
1205    Corrected wording and formatting of some CM warning message which I
1206    broke in my previous patch.
1207    
1208    ----------------------------------------------------------------------
1209    Name: Matthias Blume
1210    Date: 2000/12/22 21:20:00 JST
1211    Tag: blume-20001222-anchorenv
1212    Description:
1213    
1214    Fixed CM's handling of anchor environments in connection with CMB.make.
1215    
1216    ----------------------------------------------------------------------
1217    Name: Matthias Blume
1218    Date: 2000/12/22 13:15:00 JST
1219    Tag: blume-20001222-cleanup
1220    Description:
1221    
1222    Removed src/cm/ffi which does not (and did not) belong here.
1223    
1224    ----------------------------------------------------------------------
1225    Name: Matthias Blume
1226    Date: 2000/12/21 23:55:00 JST
1227    Tag: blume-20001221-exn
1228    Description:
1229    
1230    Probably most important: CM no longer silently swallows all exceptions
1231    in the compiler.
1232    Plus: some other minor CM changes.  For example, CM now reports some
1233    sizes for generated binfiles (code, data, envpickle, lambdapickle).
1234    
1235    ----------------------------------------------------------------------
1236    Name: Matthias Blume
1237    Date: 2000/12/15 00:01:05 JST
1238    Tag: blume-20001215-dirtool
1239    Description:
1240    
1241    - "dir" tool added.
1242    - improvements and cleanup to Tools structure
1243    - documentation updates
1244    
1245    ----------------------------------------------------------------------
1246    Name: Allen Leung
1247    Date: Thu Dec 14 03:45:24 EST 2000
1248    Description:
1249    Tag:  leunga-20001214-int-inf
1250    Description:
1251    
1252       In IntInf, added these standard functions, which are missing from our
1253    implementation:
1254    
1255        andb : int * int -> int
1256        xorb : int * int -> int
1257        orb  : int * int -> int
1258        notb : int -> int
1259         <<   : int * word -> int
1260        ~>>  : int * word -> int
1261    
1262       Not tested, I hope they are correct.
1263    
1264    ----------------------------------------------------------------------
1265    Name: Allen Leung
1266    Date: Fri Dec  8 19:23:26 EST 2000
1267    Description:
1268    Tag:  leunga-20001208-nowhere
1269    Description:
1270    
1271      Slight improvements to the 'nowhere' tool to handle OR-patterns,
1272    to generate better error messages etc.  Plus a brief manual.
1273    
1274    ----------------------------------------------------------------------
1275    Name: Lal George
1276    Date: 2000/12/08 09:54:02 EST 2000
1277    Tag: Release_110_31
1278    Description:
1279    
1280    - Version 110.31
1281    ----------------------------------------------------------------------
1282    Name: Allen Leung
1283    Date: Thu Dec  7 22:01:04 EST 2000
1284    Tag:  leunga-20001207-cell-monster-hack
1285    Description:
1286    
1287    Major MLRISC internal changes.  Affect all clients.
1288    Summary:
1289    
1290    1.  Type CELLS.cell = int is now replaced by a datatype.
1291        As a result, the old regmap is now gone.  Almost all interfaces
1292        in MLRISC change as a consequence.
1293    
1294    2.  A new brand version of machine description tool (v3.0) that generates
1295        modules expecting the new interface.  The old version is removed.
1296    
1297    3.  The RA interface has been further abstracted into two new functors.
1298        RISC_RA and X86RA.  These functors have much simpler interfaces.
1299        [See also directory MLRISC/demo.]
1300    
1301    4.  Some other new source->source code generation tools are available:
1302    
1303        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
1304        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
1305           I use this tool to generate the peephole optimizers---with the new
1306           cell type changes, peephole rules are becoming difficult to write
1307           without conditional pattern matching.
1308    
1309    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
1310        cover the entire MLRISC source tree so many things broke.
1311    
1312    6.  CM files have been moved to the subdirectory MLRISC/cm.
1313        They are moved because there are a lot of them and they clutter up the
1314        root dir.
1315    
1316    7.  More detailed documentation to come...
1317    
1318        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
1319        a makeml -rebuild first.  This is because of other other
1320        changes that Matthias has made (see below).
1321    
1322    
1323    ----------------------------------------------------------------------
1324    Name: Matthias Blume
1325    Date: 2000/11/30 23:12:00 JST
1326    Tag: blume-20001130-filereorg
1327    Description:
1328    
1329    Some manual updates and some file reorganizations in CM.
1330    
1331    ----------------------------------------------------------------------
1332    Name: Matthias Blume
1333    Date: 2000/11/24 17:45:00 JST
1334    Tag: blume-20001124-link
1335    Description:
1336    
1337    Drastically improved link traversal code for the case that the dynamic
1338    value was already loaded at bootstrap time.  As a result, CM and CMB
1339    now both load blazingly fast -- even on a very slow machine.  Also,
1340    memory consumption has been further reduced by this.
1341    
1342    Warning: The format of the PIDMAP file has changed.  THerefore, to
1343    bootstrap you have to do this:
1344    
1345    1. Run CMB.make
1346    2. Make a symbolic link for the boot directory:
1347         ln -s sml.boot.ARCH-OS xxx
1348    3. "Rebuild" the boot directory:
1349         ./makeml -boot xxx -rebuild sml ; rm xxx
1350    4. Boot normally:
1351          ./makeml
1352    
1353    ----------------------------------------------------------------------
1354    Name: Matthias Blume
1355    Date: 2000/11/21 21:20:00 JST
1356    Tag: blume-20001121-tools
1357    Description:
1358    
1359    Continued hacking on autoloading problem -- with success this time.
1360    Also changed tool-plugin mechanism.  See new CM manual.
1361    
1362    ----------------------------------------------------------------------
1363    Name: Matthias Blume
1364    Date: 2000/11/19 14:30:00 JST
1365    Tag:  blume-20001119-autoload
1366    Description:
1367    
1368    Some hacking to make autoloading faster.  Success for CMB, no success
1369    so far for CM.  There is a reduced structure CM' that autoloads faster.
1370    (This is a temporary, non-documented hack to be eliminated again when
1371    the general problem is solved.)
1372    
1373    ----------------------------------------------------------------------
1374    Name: Matthias Blume
1375    Date: 2000/11/17 14:10:00 JST
1376    Tag: blume-20001117-pickle-lib
1377    Description:
1378    
1379    1. Eliminated comp-lib.cm
1380    2. Made pickle-lib.cm
1381    3. Eliminated all uses of intset.sml (from comp-lib.cm)
1382    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
1383       equivalent constructs from smlnj-lib.cm (INtHashTable).
1384    5. Point 4. also goes for those uses of intmap.* in MLRISC.
1385       Duplicated intmap modules thrown out.
1386    6. Hunted down all duplicated SCC code and replaced it with
1387       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
1388    7. Rewrote Feedback module.
1389    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
1390       should be thrown out and equivalent modules from smlnj-lib.cm
1391       should be used (IntRedBlackSet, IntListSet, ...).
1392    
1393    Confirmed that compiler compiles to fixpoint.
1394    
1395    ----------------------------------------------------------------------
1396    Name: Allen Leung
1397    Date: 2000/11/10 18:00:00
1398    Tag: leunga-20001110-new-x86-fp
1399    
1400    A new x86 floating point code generator has been added.
1401    By default this is turned off.  To turn this on, do:
1402    
1403        CM.autoload "$smlnj/compiler.cm";
1404        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
1405    
1406    Changes:
1407    
1408    1.  Changed FTAN to FPTAN so that the assembly output is correct.
1409    2.  Changed the extension callback for FTANGENT to generate:
1410    
1411              fptan
1412              fstp  %st(0)
1413        instead of
1414              fptan
1415              fstpl ftempmem
1416    
1417    3.  Numerous assembly fixes for x86.
1418    
1419    5.  Cleaned up the machine code output module x86/x86MC.sml and added
1420        support for a whole bunch of instructions and addressing modes:
1421    
1422          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
1423          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
1424          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
1425          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
1426          fxch %st(n)
1427          fld %st(n)
1428          fst %st(n)
1429          fst mem
1430          fstp %st(n)
1431          fucom %st(n)
1432          fucomp %st(n)
1433    
1434        All these are now generated when the fast fp mode is turned on.
1435    
1436    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
1437    
1438    ----------------------------------------------------------------------
1439    Name: Matthias Blume
1440    Date: 2000/11/09 11:20:00 JST
1441    Tag: blume-20001109-scc
1442    Description:
1443    
1444    Eliminated some code duplication:
1445    
1446    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
1447       (Otherwise the functor is useless.)
1448    2. Used GraphSCCFn where SCCUtilFun was used previously.
1449    3. Got rid of SCCUtilFun (in comp-lib.cm).
1450    
1451    ----------------------------------------------------------------------
1452    Name: Lal George
1453    Date: 2000/11/06 09:02:21 EST 2000
1454    Tag: Release_110_30
1455    Description:
1456    
1457    - Version 110.30
1458    ----------------------------------------------------------------------
1459    Name: Matthias Blume
1460    Date: 2000/11/04 14:45:00
1461    Tag: blume-20001104-mlbuild
1462    Description:
1463    
1464    - Made ml-build faster on startup.
1465    - Documentation fixes.
1466    
1467    ----------------------------------------------------------------------
1468    Name: Matthias Blume
1469    Date: 2000/11/02 17:00:00 JST
1470    Tag: blume-20001102-condcomp
1471    Description:
1472    
1473    - Small tweaks to pickler -- new BOOTFILES!
1474    - Version bumped to 110.29.2.
1475    - Added conditional compilation facility to init.cmi (see comment there).
1476    ----------------------------------------------------------------------
1477    Name: Allen Leung
1478    Date: 2000/10/23 19:31:00
1479    Tag: leunga-20001023-demo-ra
1480    
1481    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
1482    2. Test programs for the graph library updated
1483    3. Some new MLRISC demo programs added
1484    
1485    ----------------------------------------------------------------------
1486    Name: Matthias Blume
1487    Date: 2000/08/31 22:15:00 JST
1488    Tag: blume-20001017-errmsg
1489    Description:
1490    
1491    More error message grief: Where there used to be no messages, there
1492    now were some that had bogus error regions.  Fixed.
1493    
1494    ----------------------------------------------------------------------
1495    Name: Matthias Blume
1496    Date: 2000/08/31 17:30:00 JST
1497    Tag: blume-20001017-v110p29p1
1498    Description:
1499    
1500    I made a version 110.29.1 with new bootfiles.
1501    
1502    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
1503              CM documentation changes and a small bugfix in CM's error reporting.
1504    
1505    ----------------------------------------------------------------------
1506    Name: Lal George
1507    Date: 2000/09/27 14:42:35 EDT
1508    Tag: george-20000927-nodestatus
1509    Description:
1510    
1511    Changed the type of the nodestatus, so that:
1512    
1513            SPILLED(~1)             is now SPILLED
1514            SPILLED(m) where m>=0   is now MEMREG(m)
1515            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
1516    
1517    ----------------------------------------------------------------------
1518    Name: Matthias Blume
1519    Date: 2000/09/07 14:45:00 JST
1520    Tag: blume-20000907-cmerrmsg
1521    Description:
1522    
1523    Small tweak to CM to avoid getting ML syntax error messages twice.
1524    
1525    ----------------------------------------------------------------------
1526    Name: Matthias Blume
1527    Date: 2000/08/31 18:00:00 JST
1528    Tag: blume-20000831-cvsbootfiles
1529    Description:
1530    
1531    New URL for boot files (because the 110.29 files on the BL server do
1532    now work correctly with my updated install scripts for yacc and lex).
1533    
1534    ----------------------------------------------------------------------
1535    Name: Matthias Blume
1536    Date: 2000/08/08 12:33:00 JST
1537    Tag: blume-20000808-manual
1538    Description:
1539    
1540    Tiny update to CM manual.
1541    
1542    ----------------------------------------------------------------------
1543    Name: Allen Leung
1544    Date: 2000/08/7 19:31:00
1545    Tag: leunga-20000807-a-whole-bunch-of-stuff
1546    
1547      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
1548    and Mobdy.
1549    
1550    1.  x86
1551    
1552       a.  Fixes to peephole module by John and Dan.
1553       b.  Assembly fix to SETcc by Allen.
1554       c.  Fix to c-call by John.
1555       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
1556       e.  Instruction selection optimization to SETcc as suggested by John.
1557    
1558           For example,
1559    
1560            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
1561    
1562           should generate:
1563    
1564            MOVL a, x
1565            SUBL b, x
1566            SHRL 31, x
1567    
1568    2.  IR stuff
1569    
1570         A bunch of new DJ-graph related algorithms added.  These
1571         speed up SSA construction.
1572    
1573    3.  SSA + Scheduling
1574    
1575         Added code for SSA and scheduling to the repository
1576    
1577    ----------------------------------------------------------------------
1578    Name: Lal George
1579    Date: 2000/07/27 11:53:14 EDT
1580    
1581    Tag: lal-20000727-linux-ppc
1582    Description:
1583    
1584     Made changes to support Linux PPC.
1585     p.s. I have confirmation that the 110.29 boot files work fine.
1586    
1587    ----------------------------------------------------------------------
1588    Name: Matthias Blume
1589    Date: 2000/07/27 17:40:00 JST
1590    Tag: blume-20000727-scripts
1591    Description:
1592    
1593    !!!! WARNING !!!!
1594    You must recompile the runtime system!
1595    !!!! WARNING !!!!
1596    
1597    This is basically another round of script-enhancements:
1598    
1599    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
1600       and undefine CM preprocessor symbols.
1601    
1602    2. ml-build avoids generating a new heap image if it finds that the
1603       existing one is still ok.  (The condition is that no ML file had to
1604       be recompiled and all ML files are found to be older that the heap
1605       file.)
1606    
1607       To make this work smoothly, I also hacked the runtime system as
1608       well as SMLofNJ.SysInfo to get access to the heap image suffix
1609       (.sparc-solaris, ...) that is currently being used.
1610    
1611       Moreover, the signature of CM.mk_standalone has changed.  See the
1612       CM manual.
1613    
1614    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
1615       CM manual for details.)
1616    
1617    4. More CM manual updates:
1618        - all of the above has been documented.
1619        - there is now a section describing the (CM-related) command line
1620          arguments that are accepted by the "sml" command
1621    
1622    ----------------------------------------------------------------------
1623    Name: Matthias Blume
1624    Date: 2000/07/25 16:20:00 JST
1625    Tag: blume-20000725-makedepend
1626    Description:
1627    
1628    Added a script called ml-makedepend.  This can be used in makefiles
1629    for Unix' make in a way very similar to the "makedepend" command for
1630    C.
1631    
1632    The script internally uses function CM.sources.
1633    
1634    Synopsis:
1635    
1636        ml-makedepend [-f makefile] cmfile targetname
1637    
1638    The default for the makefile is "makefile" (or "Makefile" should
1639    "makefile" not exist).
1640    
1641    ml-makedepend adds a cmfile/targetname-specific section to this
1642    makefile (after removing the previous version of this section).  The
1643    section contains a single dependency specification with targetname on
1644    the LHS (targetname is an arbitrary name), and a list of files derived
1645    from the cmfile on the RHS.  Some of the files on the RHS are
1646    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
1647    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
1648    path names.  The makefile writer is responsible for making sure that
1649    these variables have correct at the time "make" is invoked.
1650    
1651    ----------------------------------------------------------------------
1652    Name: Matthias Blume
1653    Date: 2000/07/22 23:30:00 JST
1654    Tag: blume-20000722-urlupdate
1655    Description:
1656    
1657    Changed BOOT and config/srcarchiveurl to point to BL server:
1658    
1659        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
1660    
1661    ----------------------------------------------------------------------
1662    Name: Matthias Blume
1663    Date: 2000/07/18 18:00:00 JST
1664    Tag: blume-20000718-Version_110_29
1665    Description:
1666    
1667    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
1668    
1669    2. Updated config/version to 110.29
1670    
1671    3. Updated config/srcarchiveurl
1672    
1673    3. New boot files!
1674       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
1675    
1676    ----------------------------------------------------------------------
1677    Name: Matthias Blume
1678    Date: 2000/07/11 13:58:00 JST
1679    Tag: blume-20000711-doctypo
1680    Description:
1681    
1682    Fixed a few typos in CM manual.
1683    
1684    ----------------------------------------------------------------------
1685    Name: Allen Leung
1686    Date: 2000/06/15 00:38:00
1687    Tag: leunga-20000704-sparc-x86
1688    
1689    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
1690    2. fix to x86 RET bug [found by Dan Grossman]
1691    3. sparc assembly bug fix for ticc instructions [found by Fermin]
1692    
1693       Affects c-- and moby only
1694    
1695    ----------------------------------------------------------------------
1696    Name: Matthias Blume
1697    Date: 2000/07/04 15:26:00
1698    Tag: blume-20000704-trigger
1699    Description:
1700    
1701    1. Improvements to CM manual.
1702    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
1703       of getting a back-trace.  The function, when called, raises an
1704       internal exception which explicitly carries the full back-trace history,
1705       so it is unaffected by any intervening handle-raise pairs ("trivial"
1706       or not).  The interactive loop will print that history once it arrives
1707       at top level.
1708       Short of having all exceptions implicitly carry the full history, the
1709       recommended way of using this facility is:
1710         - compile your program with instrumentation "on"
1711         - run it, when it raises an exception, look at the history
1712         - if the history is "cut off" because of some handler, go and modify
1713           your program so that it explicitly calls BTrace.trigger
1714         - recompile (still instrumented), and rerun; look at the full history
1715    
1716    ----------------------------------------------------------------------
1717    Name: Matthias Blume
1718    Date: 2000/07/03 15:36:00 JST
1719    Tag: blume-20000702-manual
1720    Description:
1721    
1722    Small corrections and updates to CM manual.
1723    
1724    ----------------------------------------------------------------------
1725    Name: Matthias Blume
1726    Date: 2000/06/29 16:04:00 JST
1727    Tag: blume-20000629-yacctool
1728    Description:
1729    
1730    Changes:
1731    
1732    1. Class "mlyacc" now takes separate arguments to pass options to
1733       generated .sml- and .sig-files independently.
1734    2. Corresponding CM manual updates.
1735    3. BTrace module now also reports call sites.  (However, for loop clusters
1736       it only shows from where the cluster was entered.)  There are associated
1737       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1738    
1739    ----------------------------------------------------------------------
1740    Name: Matthias Blume
1741    Date: 2000/06/27 16:51:00 JST
1742    Tag: blume-20000627-noweb
1743    Description:
1744    
1745    Changes:
1746    
1747     1. Implemented "subdir" and "witness" options for noweb tool.
1748        This caused some slight internal changes in CM's tool implementation.
1749     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
1750        some remaining issues from earlier path anchor changes.
1751     3. Updated CM manual accordingly.
1752    
1753     4. Changed implementation of back-tracing so that I now consider it
1754        ready for prime-time.
1755    
1756        In particular, you don't have to explicitly trigger the back-trace
1757        anymore.  Instead, if you are running BTrace-instrumented code and
1758        there is an uncaught exception (regardless of whether or not it was
1759        raised in instrumented code), the top-level evalloop will print
1760        the back-trace.
1761    
1762        Features:
1763    
1764          - Instrumented and uninstrumented code work together seemlessly.
1765            (Of course, uninstrumented code is never mentioned in actual
1766             back-traces.)
1767    
1768          - Asymptotic time- and space-complexity of instrumented code is
1769            equal to that of uninstrumented code.  (This means that
1770            tail-recursion is preserved by the instrumentation phase.)
1771    
1772          - Modules whose code has been instrumented in different sessions
1773            work together without problem.
1774    
1775          - There is no penalty whatsoever on uninstrumented code.
1776    
1777          - There is no penalty on "raise" expressions, even in
1778            instrumented code.
1779    
1780        A potential bug (or perhaps it is a feature, too):
1781    
1782          A back-trace reaches no further than the outermost instrumented
1783          non-trivial "raise".  Here, a "trivial" raise is one that is the
1784          sole RHS of a "handle" rule.  Thus, back-traces reach trough
1785    
1786               <exp> handle e => raise e
1787    
1788          and even
1789    
1790               <exp> handle Foo => raise Bar
1791    
1792          and, of course, through
1793    
1794               <exp> handle Foo => ...
1795    
1796         if the exception was not Foo.
1797    
1798         Back-traces always reach right through any un-instrumented code
1799         including any of its "handle" expressions, trivial or not.
1800    
1801       To try this out, do the following:
1802    
1803         - Erase all existing binfiles for your program.
1804           (You may keep binfiles for those modules where you think you
1805            definitely don't need back-tracing.)
1806         - Turn on back-trace instrumentation:
1807              SMLofNJ.Internals.BTrace.mode (SOME true);
1808         - Recompile your program.  (I.e., run "CM.make" or "use".)
1809         - You may now turn instrumentation off again (if you want):
1810              SMLofNJ.Internals.BTrace.mode (SOME false);
1811         - Run your program as usual.  If it raises an exception that
1812           reaches the interactive toplevel, then a back-trace will
1813           automatically be printed.  After that, the toplevel loop
1814           will print the exception history as usual.
1815    
1816    ----------------------------------------------------------------------
1817    Name: Matthias Blume
1818    Date: 2000/06/26 09:56:46 JST
1819    Tag: blume-20000626-setup
1820    Description:
1821    
1822    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1823          ML code before and after compiling a file (e.g., to set compiler
1824          flags)
1825    
1826    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1827              - associated changes to btrace.sml (BTrace instrumentation pass)
1828              - cleaner implementation of btimp.sml (BTrace tracing and report
1829                module)
1830    
1831    CM manual: * new path encoding documented
1832               * description of setup-parameter to "sml" added
1833    
1834    The biggest user-visible change to back-tracing is that it is no
1835    longer necessary to compile all traced modules within the same
1836    session.  (This was a real limitation.)
1837    
1838    ----------------------------------------------------------------------
1839    Name: Matthias Blume
1840    Date: 2000/06/24 12:40:00 JST
1841    Tag: blume-20000624-startup
1842    Description:
1843    
1844    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1845    bit too often -- to put it mildly. :)
1846    
1847    ----------------------------------------------------------------------
1848    Name: Matthias Blume
1849    Date: 2000/06/23 18:20:00 JST
1850    Tag: blume-20000623-btrace
1851    Description:
1852    
1853    This updates adds a backtrace facility to aid programmers in debugging
1854    their programs.  This involves the following changes:
1855    
1856    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1857       keeping track of the current call stack.  When programs are compiled
1858       in a special mode, the compiler will insert calls to these hooks
1859       into the user program.
1860       "Hook" means that it is possible for different implementations of
1861       back-tracing to register themselves (at different times).
1862    
1863    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1864       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1865       It can be turned on using this call:
1866         SMLofNJ.Internals.BTrace.mode (SOME true);
1867       Turning it off again:
1868         SMLofNJ.Internals.BTrace.mode (SOME false);
1869       Querying the current status:
1870         SMLofNJ.Internals.BTrace.mode NONE;
1871       Annotated programs are about twice as big as normal ones, and they
1872       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1873       where all hooks do nothing).  The slowdown with a plugin that is
1874       actually useful (such as the one supplied by default) is even greater,
1875       but in the case of the default plugin it is still only an constant
1876       factor (amortized).
1877    
1878    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1879       with a sub-structure BTrace for controlling back-tracing.  In particular,
1880       the above-mentioned function "mode" controls whether the annotation
1881       phase is invoked by the compiler.  Another important function is
1882       "trigger": when called it aborts the current execution and causes
1883       the top-level loop to print a full back-trace.
1884    
1885    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1886       for back-tracing.  It keeps track of the dynamic call stack and in
1887       addition to that it keeps a partial history at each "level" of that
1888       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1889       d and b (at separate times, dynamically), then the report will show:
1890    
1891       GOTO   d
1892             /c
1893       GOTO  \b
1894       CALL   a
1895    
1896       This shows that there was an initial non-tail call of a, then a
1897       tail-call to b or c, looping behavior in a cluster of functions that
1898       consist of b and c, and then a goto from that cluster (i.e., either from
1899       b or from c) to d.
1900    
1901       Note that (depending on the user program) the amount of information
1902       that the back-trace module has to keep track of at each level is bounded
1903       by a constant.  Thus, the whole implementation has the same asymptotical
1904       complexity as the original program (both in space and in time).
1905    
1906    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1907       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1908       which is raised by the "trigger" function mentioned above.
1909    
1910    Notes on usage:
1911    
1912    - Annotated code works well together with unannotated code:
1913    Unannotated calls simply do not show up at all in the backtrace.
1914    
1915    - It is not a good idea to let modules that were annotated during
1916    different sessions run at the same time.  This is because the compiler
1917    chooses small integers to identify individual functions, and there
1918    will be clashes if different modules were compiled in separate sessions.
1919    (Nothing will crash, and you will even be told about the clashes, but
1920    back-trace information will in general not be useful.)
1921    
1922    - Back-tracing can be confused by callcc and capture.
1923    
1924    - The only way of getting a back-trace right now is to explicitly
1925    invoke the "trigger" function from your user program.  Eventually, we
1926    should make every exception carry back-trace information (if
1927    available).  But since this creates more overhead at "raise"-time
1928    (similar to the current exnHistory overhead), I have not yet
1929    implemented this.  (The implementation will be rather easy.)  With
1930    exceptions carrying back-trace information, this facility will be even
1931    more useful because users don't need to modify their programs...
1932    
1933    - While it is possible to compile the compiler with back-trace
1934    annotations turned on (I did it to get some confidence in
1935    correctness), you must make absolutely sure that core.sml and
1936    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1937    be compiled with annotation because there is no core access yet, but
1938    if you compile btimp.sml with annotation, then the system will go into
1939    an infinite recursion and crash.)
1940    Since CM currently does not know about BTrace, the only way to turn
1941    annotations on and off for different modules of the compiler is to
1942    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1943    this is awkward and clumsy.
1944    
1945    Sample sessions:
1946    
1947    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1948    - SMLofNJ.Internals.BTrace.mode (SOME true);
1949    [autoloading]
1950    [autoloading done]
1951    val it = false : bool
1952    - structure X = struct
1953    -     fun main n = let
1954    -         fun a (x, 0) = d x
1955    -           | a (x, n) = b (x, n - 1)
1956    -         and b (x, n) = c (x, n)
1957    -         and c (x, n) = a (x, n)
1958    -         and d x = e (x, 3)
1959    -         and e (x, 0) = f x
1960    -           | e (x, n) = e (x, n - 1)
1961    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1962    -           | f n = n * g (n - 1)
1963    -         and g n = a (n, 3)
1964    -     in
1965    -         f n
1966    -     end
1967    - end;
1968    structure X : sig val main : int -> int end
1969    - X.main 3;
1970    *** BACK-TRACE ***
1971    GOTO   stdIn:4.2-13.20: X.main[2].f
1972    GOTO-( stdIn:4.2-13.20: X.main[2].e
1973    GOTO   stdIn:4.2-13.20: X.main[2].d
1974         / stdIn:4.2-13.20: X.main[2].a
1975         | stdIn:4.2-13.20: X.main[2].b
1976    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1977    CALL   stdIn:4.2-13.20: X.main[2].g
1978    GOTO   stdIn:4.2-13.20: X.main[2].f
1979    GOTO-( stdIn:4.2-13.20: X.main[2].e
1980    GOTO   stdIn:4.2-13.20: X.main[2].d
1981         / stdIn:4.2-13.20: X.main[2].a
1982         | stdIn:4.2-13.20: X.main[2].b
1983    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1984    CALL   stdIn:4.2-13.20: X.main[2].g
1985    GOTO   stdIn:4.2-13.20: X.main[2].f
1986    GOTO-( stdIn:4.2-13.20: X.main[2].e
1987    GOTO   stdIn:4.2-13.20: X.main[2].d
1988         / stdIn:4.2-13.20: X.main[2].a
1989         | stdIn:4.2-13.20: X.main[2].b
1990    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1991    CALL   stdIn:4.2-13.20: X.main[2].g
1992    GOTO   stdIn:4.2-13.20: X.main[2].f
1993    CALL   stdIn:2.15-17.4: X.main[2]
1994    -
1995    
1996    (Note that because of a FLINt bug the above code currently does not
1997    compile without BTrace turned on.)
1998    
1999    Here is another example, using my modified Tiger compiler:
2000    
2001    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
2002    - SMLofNJ.Internals.BTrace.mode (SOME true);
2003    [autoloading]
2004    [autoloading done]
2005    val it = false : bool
2006    - CM.make "sources.cm";
2007    [autoloading]
2008    ...
2009    [autoloading done]
2010    [scanning sources.cm]
2011    [parsing (sources.cm):parse.sml]
2012    [creating directory CM/SKEL ...]
2013    [parsing (sources.cm):tiger.lex.sml]
2014    ...
2015    [wrote CM/sparc-unix/semant.sml]
2016    [compiling (sources.cm):main.sml]
2017    [wrote CM/sparc-unix/main.sml]
2018    [New bindings added.]
2019    val it = true : bool
2020    - Main.compile ("../testcases/merge.tig", "foo.out");
2021    *** BACK-TRACE ***
2022    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
2023    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2024    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2025    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2026    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2027    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2028    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
2029         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
2030    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
2031    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2032    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
2033    CALL   main.sml:1.18-118.4: Main.compile[2]
2034    -
2035    
2036    ----------------------------------------------------------------------
2037    Name: Matthias Blumen
2038    Date: 2000/06/21 18:00:00 JST
2039    Tag: blume-20000621-manual
2040    Description:
2041    
2042    CM manual update: Path environments documented.
2043    
2044    ----------------------------------------------------------------------
2045    Name: Matthias Blume
2046    Date: 2000/06/19 13:40:00
2047    Tag: blume-20000619-manual
2048    Description:
2049    
2050    CM manual and system/README update.  This only covers the fact that
2051    there are no more implicit anchors.  (Path environments and the "bind"
2052    option to "cm" have yet to be documented.)
2053    
2054    ----------------------------------------------------------------------
2055    Name: Matthias Blume
2056    Date: 2000/06/19 11:05:00 JST
2057    Tag: blume-20000619-chdir-bugfix
2058    Description:
2059    
2060    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
2061    
2062    ----------------------------------------------------------------------
2063    Name: Matthias Blume
2064    Date: 2000/06/18 22:00:10 JST
2065    Tag: blume-20000618-implicit-anchors-really-gone
2066    Description:
2067    
2068    I updates the previous HISTORY entry where I forgot to mention that
2069    implicit anchors are no longer with us.
2070    
2071    The current update also gets rid of the (now useless) controller
2072    CM.Control.implicit_anchors.
2073    
2074    ----------------------------------------------------------------------
2075    Name: Matthias Blume
2076    Date: 2000/06/16 17:30:00 JST
2077    Tag: blume-20000616-anchorenv
2078    Description:
2079    
2080    This patch implements the long anticipated (just kidding :) "anchor
2081    environment" mechanism.  In the course of doing this, I also
2082    re-implemented CM's internal "SrcPath" module from scratch.  The new
2083    one should be more robust in certain boundary cases.  In any case, it
2084    is a lot cleaner than its predecessor (IMHO).
2085    
2086    This time, although there is yet another boot file format change, I
2087    kept the unpickler backward-compatible.  As a result, no new bootfiles
2088    are necessary and bootstrapping is straightforward.  (You cannot read
2089    new bootfiles into an old system, but the other way around is no
2090    problem.)
2091    
2092    Visible changes:
2093    
2094    ** 0. Implicit path anchors (without the leading $-symbol) are no
2095    longer recognized at all. This means that such path names are not
2096    illegal either.  For example, the name basis.cm simply refers to a
2097    local file called "basis.cm" (i.e, the name is an ordinary path
2098    relative to .cm-files directory).  Or, to put it differently, only
2099    names that start with $ are anchored paths.
2100    
2101    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
2102    vanished.
2103    
2104    John (Reppy) had critizised this as soon as I originally proposed and
2105    implemented it, but at that time I did not really deeply believe
2106    him. :) Now I came full-circle because I need the $<singlearc> syntax
2107    in another place where it cannot be seen as an abbreviation for
2108    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
2109    seems to mean (i.e., it "expands" into the corresponding anchor
2110    value).
2111    
2112    However, when paths are used as members in CM description files, it
2113    continues to be true that there must be at least another arc after the
2114    anchor.  This is now enforced separately during semantic analysis
2115    (i.e., from a lexical/syntactical point of view, the notation is ok.)
2116    
2117    ** 2. The "cm" class now accepts an option "bind".  The option's value
2118    is a sub-option list of precisely two items -- one labeled "anchor"
2119    and the other one labeled "value".  As you might expect, "anchor" is
2120    used to specify an anchor name to be bound, and "value" specifies what
2121    the anchor is being bound to.
2122    
2123    The value must be a directory name and can be given in either standard
2124    syntax (including the possibility that it is itself an anchored path)
2125    or native syntax.
2126    
2127    Examples:
2128    
2129       foo.cm (bind:(anchor:bar value:$mystuff/bar))
2130       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
2131    
2132    and so on.
2133    
2134    The meaning of this is that the .cm-file will be processed with an
2135    augmented anchor environment where the given anchor(s) is/are bound to
2136    the given values(s).
2137    
2138    The rationale for having this feature is this: Suppose you are trying
2139    to use two different (already stable) libraries a.cm and b.cm (that
2140    you perhaps didn't write yourself).  Further, suppose each of these
2141    two libraries internally uses its own auxiliary library $aux/lib.cm.
2142    Normally you would now have a problem because the anchor "lib" can not
2143    be bound to more than one value globally.  Therefore, the project that
2144    uses both a.cm and b.cm must locally redirect the anchor to some other
2145    place:
2146    
2147       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
2148       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
2149    
2150    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
2151    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
2152    
2153    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
2154    you when you do so at the time of CM.stabilize).  Therefore, you can
2155    also use an anchored path as the value:
2156    
2157      a.cm (bind:(anchor:lib value:$a-lib))
2158      b.cm (bind:(anchor:lib value:$b-lib))
2159    
2160    Now you can globally configure (using the usual CM.Anchor.anchor or
2161    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
2162    itself is always locally bound, setting it globally is no longer
2163    meaningful or necessary (but it does not hurt either).  In fact, "lib"
2164    can still be used as a global anchor for separate purposes.  As a
2165    matter of fact, one can locally define "lib" in terms of a global
2166    "lib":
2167    
2168      a.cm (bind:(anchor:lib value:$lib/a))
2169      b.cm (bind:(anchor:lib value:$lib/b))
2170    
2171    ** 3: The encoding of path names has changed.  This affects the way
2172    path names are shown in CM's progress report and also the internal
2173    protocol encoding used for parallel make.
2174    
2175    The encoding now uses one or more ':'-separated segments.  Each
2176    segments corresponds to a file that has been specified relative to the
2177    file given by its preceding segment.  The first segment is either
2178    relative to the CWD, absolute, or anchored.  Each segment itself is
2179    basically a Unix pathname; all segments but the first are relative.
2180    
2181    Example:
2182    
2183       $foo/bar/baz.cm:a/b/c.sml
2184    
2185    This path denotes the file bar/a/b/c.sml relative to the directory
2186    denoted by anchor "foo".  Notice that the encoding also includes
2187    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
2188    paths are resolved relative to the .cm-files directory, so baz.cm must
2189    be ignored to get the "real" pathname.
2190    
2191    To make this fact more obvious, CM puts the names of such "virtual
2192    arcs" into parentheses when they appear in progress reports. (No
2193    parentheses will appear in the internal protocol encoding.)  Thus,
2194    what you really see is:
2195    
2196      $foo/bar/(baz.cm):a/b/c.sml
2197    
2198    I find this notation to be much more informative than before.
2199    
2200    Another new feature of the encoding is that special characters
2201    including parentheses, colons, (back)slashes, and white space are
2202    written as \ddd (where ddd is the decimal encoding of the character).
2203    
2204    *** The CM manual still needs to be updated.
2205    
2206    ----------------------------------------------------------------------
2207    Name: Allen Leung
2208    Date: 2000/06/15 00:38:00
2209    Tag: leunga-20000615-x86-peephole
2210    
2211    x86 Peephole fix by Fermin.  Affects c-- and moby only.
2212    
2213    ----------------------------------------------------------------------
2214    Name: Matthias Blume
2215    Date: 2000/06/12 11:40:00
2216    Tag: blume-20000612-parmakefix
2217    Description:
2218    
2219    More cleanup after changing the file naming scheme: This time I
2220    repaired the parallel make mechanism for CMB.make which I broke earlier.
2221    
2222    ----------------------------------------------------------------------
2223    Name: Allen Leung
2224    Date: 2000/06/09 01:25:00
2225    Tag: leunga-20000609-various
2226    
2227    None of these things should affect normal SML/NJ operations
2228    
2229    1. Peephole improvements provided by Fermin (c--)
2230    2. New annotation DEFUSE for adding extra dependence (moby)
2231    3. New X86 LOCK instructions (moby)
2232    4. New machine description language for reservation tables (scheduling)
2233    5. Fixes to various optimization/analysis modules (branch chaining, dominator
2234       trees etc.)
2235    6. I've changed the CM files so that they can work with versions
2236       110.0.6, 110.25 and 110.28
2237    
2238    ----------------------------------------------------------------------
2239    Name: Matthias Blume
2240    Date: 2000/06/09 12:40:00
2241    Tag: blume-20000609-log
2242    Description:
2243    
2244    - Removed all(?) remaining RCS Log entries from sources.
2245    
2246    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
2247      anchored paths).
2248    
2249    ----------------------------------------------------------------------
2250    Name: Matthias Blume
2251    Date: 2000/06/07 17:00:00 JST
2252    Tag: blume-20000607-no-implicit-anchors
2253    Description:
2254    
2255    1. This update changes the default setting for
2256    CM.Control.implicit_anchors from true to false.  This means that
2257    implicit anchors are no longer permitted by default.  I also tried to
2258    make sure that nothing else still relies on implicit anchors.
2259    (This is the next step on the schedule towards a CM that does not even
2260    have the notion of implicit anchors anymore.)
2261    
2262    2. More CM manual updates.
2263    
2264    3. I managed to track down and fix the pickling bug I mentioned last
2265    time.  Because of the previously existing workaround, this entails no
2266    immediate practical changes.
2267    
2268    ----------------------------------------------------------------------
2269    Name: Matthias Blume
2270    Date: 2000/06/06 11:15:00 JST
2271    Tag: blume-20000606-lazierpickle
2272    Description:
2273    
2274    !!!! NEW BOOT FILES !!!!
2275    
2276    * The main purpose of this update is to make library pickles lazier in
2277    order to reduce the initial space penalty for autoloading a library.
2278    As a result, it is now possible to have $smlnj/compiler.cm
2279    pre-registered.  This should take care of the many complaints or
2280    inquiries about missing structure Compiler.  This required changes to
2281    CM's internal data structures and small tweaks to some algorithms.
2282    
2283    As a neat additional effect, it is no longer necessary (for the sake
2284    of lean heap image files) to distinguish between a "minimal" CM and a
2285    "full" CM.  Now, there is only one CM (i.e., the "full" version:
2286    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
2287    interactive top level. ($smlnj/cm/minimal.cm is gone.)
2288    
2289    To make the life of compiler-hackers easier, "makeml" now also
2290    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
2291    words, after you bootstrap a new sml for the first time, you will not
2292    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
2293    around you will still have to do it, though.)
2294    
2295    * A second change consists of major updates to the CM manual.  There
2296    are now several appendices with summary information and also a full
2297    specification of the CM description file syntax.
2298    
2299    * In directory src/system I added the script "allcross".  This script
2300    invokes sml and cross-compiles the compiler for all supported
2301    architectures.  (Useful when providing a new set of boot files.)
2302    
2303    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
2304    added a small tweak to pickle-util.sml to work around this problem,
2305    but it is not a proper fix yet.  I will investigate further.  (The
2306    effect of the bug was an inflation of library pickle size.)
2307    
2308    * Version number increased to 110.28.1 (to avoid compatibility problems).
2309    
2310    ----------------------------------------------------------------------
2311    Name: Allen Leung
2312    Date: 2000/05/25 17:28 EDT
2313    Tag: leunga-20000525-ra
2314    Description:
2315    
2316      Fixed a bug in freezing phase of the register allocator.
2317    
2318    ----------------------------------------------------------------------
2319    Name: Allen Leung
2320    Date: 2000/05/15 22:53 EDT
2321    Tag: leunga-20000515-alpha-x86-ra
2322    Description:
2323    
2324      1. Alpha
2325    
2326          Slight cleanup.  Removed the instruction SGNXL
2327    
2328      2. X86
2329    
2330          Added the following instructions to the instruction set:
2331    
2332            ROLx, RORx,
2333            BTx, BTSx, BTLx, BTRx,
2334            XCHGx, and variants with the LOCK prefix
2335    
2336      3. Register Allocation
2337    
2338          The module ra-rewrite-with-renaming has been improved.
2339    
2340      These have no effect on SML/NJ.
2341    
2342    ----------------------------------------------------------------------
2343    Name: Matthias Blume
2344    Date: 2000/05/15 16:20:00 JST
2345    Tag: blume-20000515-lightrebuild
2346    Description:
2347    
2348    1. I added an alternative to "-rebuild" to "makeml".  The difference is
2349       that prior to calling CMB.make' the CM-variable "LIGHT" will be
2350       defined.  In effect, the command will not build any cross-compiler
2351       backends and therefore finish more quickly.
2352    
2353       The "fixpt" script also takes a "-light" switch to be able to use
2354       this new facility while compiling for a fixpoint.
2355    
2356    2. I replaced all mentions of anchored paths in group owner specifications
2357       with simple relative paths (usually starting with "..").
2358       The rationale is that a library's internal workings should not be
2359       compromised by the lack of some anchor.  (An anchor is necessary
2360       for someone who wants to refer to the library by an anchored path,
2361       but it should not be necessary to build the same library in the first
2362       place.)
2363    
2364    3. I changed the way CM's tool mechanism determines the shell command
2365       string used for things like ml-yacc etc. so that it does not break
2366       when CM.Control.implicit_anchors is turned off.
2367    
2368    ----------------------------------------------------------------------
2369    Name: Matthias Blume
2370    Date: 2000/05/12 18:20:00 JST
2371    Tag: blume-20000512-ml-build
2372    Description:
2373    
2374    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
2375    getting installed properly (by config/install.sh).
2376    
2377    ----------------------------------------------------------------------
2378    Name: Matthias Blume
2379    Date: 2000/05/12 17:30:00 JST
2380    Tag: blume-20000512-anchors
2381    Description:
2382    
2383    !!! NEW BOOT FILES !!!
2384    
2385    This change is in preparation of fading out support for "implicitly
2386    anchored path names".  I went through all sources and used the
2387    explicit (and relatively new) $-notation.  See system/README and the
2388    CM manual for more info on this.
2389    
2390    I also modified the anchoring scheme for some things such as "smlnj",
2391    "MLRISC", "cm", etc. to take advantage of the fact that explicit
2392    anchors are more expressive: anchor name and first arc do not have to
2393    coincide.  This entails the following user-visible change:
2394    
2395    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
2396    particular, when you fire up sml with a command-line argument, say,
2397    e.g.:
2398    
2399       sml '$smlnj/cmb.cm'
2400    
2401    At the ML toplevel prompt:
2402    
2403       CM.autoload "$smlnj/cmb.cm";
2404    
2405    There is also a new controller in CM.Control that can be used to turn
2406    off all remaining support for implicit anchors by saying:
2407    
2408        CM.autoload "$smlnj/
2409        #set CM.Control.implicit_anchors false;
2410    
2411    This causes CM to reject implicitly anchored paths.  This is (for the
2412    time being) less permissive than the "final" version where there will
2413    be no more such implicit anchors and relative paths will be just that:
2414    relative.
2415    
2416    The next step (version after next version?) will be to make the
2417    default for CM.Control.implicit_anchors false.  After the dust has
2418    settled, I can then produce the "final" version of this...
2419    
2420    Note: Since bootstrapping is a bit tricky, I provided new boot files.
2421    
2422    ----------------------------------------------------------------------
2423    Name: Matthias Blume
2424    Date: 2000/05/11 16:30:00 JST
2425    Tag: blume-20000511-sources
2426    Description:
2427    
2428    The main change is that I added function CM.sources as a generalized
2429    version of the earlier CM.makedepend.  This entails the following
2430    additional changes:
2431    
2432      - CM.makedepend has been dropped.
2433    
2434      - CM manual has been updated.
2435    
2436      - TOOLS signature and API have been changed.
2437    
2438    ----------------------------------------------------------------------
2439    Name: Allen Leung
2440    Date: 2000/05/10 21:17 EDT
2441    Tag: leunga-20000510-moby-c--ssa
2442    Description:
2443    
2444      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
2445    None of these affect SML/NJ.
2446    
2447    1. Register Allocation
2448    
2449        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
2450           This module tries to remove local (i.e. basic block level) redundancies
2451           during spilling.
2452    
2453        b. A new framework for performing region based register allocation.
2454           Not yet entirely functional.
2455    
2456    2. X86
2457    
2458       a. DefUse for POP was missing the stack pointer [found by Lal]
2459       b. Reload for CALL was incorrect in X86Spill [found by John]
2460       c. Various fixes in X86Spill so that it can be used correctly for
2461          the new spilling module.
2462    
2463    3. SSA/IR
2464    
2465       a. New module ir/dj-dataflow.sml implements elimination based
2466          data flow analysis.
2467    
2468    4. MLRiscGen
2469    
2470       a. Fix for gc type annotation
2471    
2472    5. MDGen
2473    
2474       Various fixes for machine description -> ml code translation.  For ssa
2475       only.
2476    
2477    ----------------------------------------------------------------------
2478    Name: Allen Leung
2479    Date: 2000/05/08 22:17 EDT
2480    Tag: leunga-20000508-labexp
2481    Description:
2482    
2483      Fermin has found a few assembly problems with constant expressions
2484      generated in LabelExp.  Mostly, the problems involve extra parentheses,
2485      which choke on dumb assemblers.  This is his fix.
2486    
2487    ----------------------------------------------------------------------
2488    Name: Dave MacQueen
2489    Date: 2000/04/09 14:00 EDT
2490    Tag: dbm-20000502-Version_110_28
2491    Description:
2492    
2493    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
2494    
2495    2. Updated config/version to 110.28
2496    
2497    3. Updated config/srcarchiveurl
2498    
2499    3. New boot files!
2500       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
2501    
2502    ----------------------------------------------------------------------
2503    Name: Matthias Blume
2504    Date: 2000/05/01 19:05:00 JST
2505    Tag: blume-20000501-noweb
2506    Description:
2507    
2508    A new noweb tool has been added.  The existing system is entirely
2509    unaffected by this, but some CM users have asked for renewed noweb
2510    support.  Everything is documented in the CM manual.
2511    
2512    New (plugin) libraries:
2513    
2514       noweb-tool.cm
2515       nw-ext.cm
2516    
2517    ----------------------------------------------------------------------
2518    Name: Dave MacQueen
2519    Date: 2000/04/30 12:40PM EDT
2520    Tag: dbm-20000430-bug_fixes
2521    Description:
2522    
2523    1. Fix for bug 1498
2524       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
2525       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
2526         added toRealArray function
2527       smlnj/src/compiler/MiscUtil/print/ppobj.sml
2528         added check for tag Obj.RealArray to array printing case in ppObj
2529    
2530    2. Fix for bug 1510
2531       smlnj/src/compiler/Semant/types/typesutil.sml
2532         fixed definition of dummyargs (used by equalTycon) so that
2533         dummy args are distinct types
2534    
2535    ----------------------------------------------------------------------
2536    Name: Matthias Blume
2537    Date: 2000/04/30 01:00:00 JST
2538    Tag: blume-20000430-versions
2539    Description:
2540    
2541    1. CM version numbering added.  This is an implementation of Lal's
2542       proposal for adding version numbers and version checking to .cm
2543       files.  Lal said that his proposal was just that -- a proposal.
2544       For the time being I went ahead and implemented it so that people
2545       can comment on it.  Everything is completely backward-compatible
2546       (except for the stable library format, i.e., new bootfiles!).
2547    
2548       As usual, see the CM manual for details.
2549    
2550    2. An alternative syntax for anchored paths has been implemented.
2551       Dave has recently voiced the same concerns that I had when I did
2552       this, so there should be some support.  My take is that eventually
2553       I will let support for the current syntax (where anchors are
2554       "implicit") fade out in favor of the new, explicit syntax.
2555       In order to be backward-compatible, both old and new syntax are
2556       currently supported.
2557    
2558       Again, see the CM manual for details.
2559    
2560    3. Parallel make is trying to be slightly smarter:  When the master
2561       process finds a "bottleneck", i.e., when there is only one
2562       compilation unit that can be compiled and everybody else is
2563       waiting on it, then it will simply compile it directly instead
2564       of clumsily telling one of the slaves to do it.
2565    
2566    4. Support for "unsharing" added.  This is necessary in order to be
2567       able to have two different versions of the same library running
2568       at the same time (e.g., for trying out a new MLRISC while still
2569       having the old MLRISC linked into the current compiler, etc.)
2570       See the CM manual.
2571    
2572    5. Simple "makedepend" functionality added for generating Makefile
2573       dependency information.  (This is rather crude at the moment.
2574       Expect some changes here in the future.)
2575    
2576    6. ".fun" added as a recognized suffix for ML files. Also documented
2577       explicitly in the manual that the fallback behavior (unknown suffix
2578       -> ML file) is not an official feature!
2579    
2580    7. Small changes to the pickler for stable libraries.
2581    
2582    8. Several internal changes to CM (for cleanup/improvement).
2583    
2584    
2585    !!!! NEW BINFILES !!!!
2586    
2587    ----------------------------------------------------------------------
2588    Name: Matthias Blume
2589    Date: 2000/04/28 17:30:00 JST
2590    Tag: blume-20000428-pathconfig
2591    Description:
2592    
2593    1. I changed config/install.sh to remove duplicate entries from the
2594       lib/pathconfig file at the end.  Moreover, the final version of
2595       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
2596       in src/system/installml.
2597    
2598    2. The config/install.sh script now consistently uses relative
2599       pathnames in lib/pathconfig whenever the anchor is in the lib
2600       directory.  (So far this was true for the libraries that come
2601       pre-compiled and bundled as part of the bootfiles but not for
2602       libraries that are compiled by the script itself.)
2603    
2604    ----------------------------------------------------------------------
2605    Name: Matthias Blume
2606    Date: 2000/04/26 13:10:00 JST
2607    Tag: blume-20000426-fun_suffix
2608    Description:
2609    
2610    Added ".fun" as a recognized file name suffix (for ML code).
2611    
2612    ----------------------------------------------------------------------
2613    Name: Allen Leung
2614    Date: 2000/04/25 17:00:00 EST
2615    Tag: leunga-20000425-alpha-ra
2616    Description:
2617    
2618    1. Alpha
2619    
2620        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
2621    in C--.
2622    
2623    2. RA
2624    
2625       Added a flag "ra-dump-size" to print out the size of the flowgraph
2626       and the interference graph.
2627    
2628    ----------------------------------------------------------------------
2629    Name: Dave MacQueen
2630    Date: 2000/04/25/
2631    Tag: dbm-20000425-mlyacc_doc_examples
2632    Description:
2633      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
2634      Updated all three examples in src/ml-yacc/examples to run
2635      under 110.* using CM.make.
2636    
2637    ----------------------------------------------------------------------
2638    Name: Allen Leung
2639    Date: 2000/04/20 23:04:00 EST
2640    Tag: leunga-20000420-ssa-c---stuff
2641    Description:
2642    
2643      This update synchronizes my repository with Yale's.  Most of these
2644    changes, however, do not affect SML/NJ at all (the RA is an exception).
2645    
2646    1. Register Allocator
2647    
2648       a. An improvement in the interference graph construction:
2649          Given a copy
2650    
2651                s <- t
2652    
2653          no interference edge between s and t is added for this definition of s.
2654    
2655       b. I've added two new spill heuristic modules that Fermin and I developed
2656          (in the new library RA.cm). These are unused in SML/NJ but maybe
2657          useful for others (Moby?)
2658    
2659    2. X86
2660    
2661       a. Various fixes in the backend provided by Fermin [C--] and Lal.
2662    
2663    3. Alpha
2664    
2665       a. Added the BSR instruction and code generation that goes with it [C--]
2666       b. Other fixes too numerous to recount provided by Fermin [C--]
2667    
2668    4. Regmaps
2669    
2670       a. The regmaps are not initialized with the identity physical bindings
2671          at creation time.  This is unneeded.
2672    
2673    5. MLRISC Optimizations
2674    
2675       a. The DJ-Graph module can now compute the iterated dominance frontiers
2676          intersects with liveness incrementally in linear time! Woohoo!
2677          This is now used in my new SSA construction algorithm.
2678    
2679       b. THe branch reorganization module is now smarter about linear chains of
2680          basic blocks.
2681    
2682    
2683    ----------------------------------------------------------------------
2684    Name: Matthias Blume
2685    Date: 2000/04/12 13:52:00 JST
2686    Tag: blume_main_v110p27_1
2687    Description:
2688    
2689    Changed install.sh script to handle archive files without version number
2690    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
2691    name of the boot file archive.
2692    
2693    ----------------------------------------------------------------------
2694    Name: Dave MacQueen
2695    Date: 2000/04/09 14:00 EDT
2696    Tag: dbm-20000410-Version_110_27
2697    Description:
2698    
2699    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
2700    
2701    2. Updated src/config/version to 110.27
2702    
2703    3. New boot files!
2704    
2705    ----------------------------------------------------------------------
2706    Name: Allen Leung
2707    Date: 2000/04/09 19:09:00 EST
2708    Tag: leunga-20000409-misc
2709    Description:
2710    
2711    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
2712    
2713    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
2714    
2715    ----------------------------------------------------------------------
2716    Name: Stefan
2717    Date: 2000/04/07 10:00:00 EDT
2718    Tag: monnier-20000406-branch-handling
2719    Description:
2720    
2721    Improved handling of branches (mostly those generated from
2722    polymorphic equality), removed switchoff and changed the
2723    default optimization settings (more cpsopt and less flintopt).
2724    
2725    ----------------------------------------------------------------------
2726    Name: Allen Leung
2727    Date: 2000/04/06 01:30:00 EST
2728    Tag: leunga-20000406-peephole-x86-SSA-2
2729    Description:
2730    
2731       Forgot a few files.
2732    
2733    ----------------------------------------------------------------------
2734    Name: Allen Leung
2735    Date: 2000/04/06 00:36:00 EST
2736    Tag: leunga-20000406-peephole-x86-SSA
2737    Description:
2738    
2739    1.  New Peephole code
2740    
2741    2.  Minor improvement to X86 instruction selection
2742    
2743    3.  Various fixes to SSA and machine description -> code translator
2744    
2745    ----------------------------------------------------------------------
2746    Name: Matthias Blume
2747    Date: 2000/04/05 12:30:00 JST
2748    Tag: blume_main_v110p26p2_3
2749    Description:
2750    
2751    This update just merges three minor cosmetic updates to CM's sources
2752    to get ready for the 110.27 code freeze on Friday.  No functionality
2753    has changed.
2754    
2755    ----------------------------------------------------------------------
2756    Name: Allen Leung
2757    Date: 2000/04/04 19:39:00 EST
2758    Tag: leunga-20000404-x86-asm
2759    Description:
2760    
2761    1.  Fixed a problem in X86 assembly.
2762    
2763        Things like
2764    
2765           jmp %eax
2766           jmp (%eax)
2767    
2768        should be output as
2769    
2770           jmp *%eax
2771           jmp *(%eax)
2772    
2773    2.  Assembly output
2774    
2775          Added a new flag
2776    
2777              "asm-indent-copies" (default to false)
2778    
2779          When this flag is on, parallel copies will be indented an extra level.
2780    
2781    ----------------------------------------------------------------------
2782    Name: Allen Leung
2783    Date: 2000/04/04 03:18:00 EST
2784    Tag: leunga-20000404-C--Moby
2785    Description:
2786    
2787        All of these fixes are related to C--, Moby, and my own optimization
2788        stuff; so they shouldn't affect SML/NJ.
2789    
2790    1.  X86
2791    
2792        Various fixes related floating point, and extensions.
2793    
2794    2.  Alpha
2795    
2796        Some extra patterns related to loads with signed/zero extension
2797        provided by Fermin.
2798    
2799    3.  Assembly
2800    
2801        When generating assembly, resolve the value of client defined constants,
2802        instead of generating symbolic values.  This is controlled by the
2803        new flag "asm-resolve-constants", which is default to true.
2804    
2805    4.  Machine Descriptions
2806    
2807        a. The precedence parser was slightly broken when parsing infixr symbols.
2808        b. The type generalizing code had the bound variables reversed, resulting
2809           in a problem during arity raising.
2810        c. Various fixes in machine descriptions.
2811    
2812    ----------------------------------------------------------------------
2813    Name: Matthias Blume
2814    Date: 2000/04/03 16:05:00 JST
2815    Tag: blume_main_v110p26p2_2
2816    Description:
2817    
2818    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
2819    now done via the ordinary static environment that is context to each
2820    compilation unit.
2821    
2822    To this end, I arranged that instead of "structure Core" as "structure
2823    _Core" is bound in the pervasive environment.  Core access is done via
2824    _Core (which can never be accidentally rebound because _Core is not a
2825    legal surface-syntax symbol).
2826    
2827    The current solution is much cleaner because the core environment is
2828    now simply part of the pervasive environment which is part of every
2829    compilation unit's context anyway.  In particular, this eliminates all
2830    special-case handling that was necessary until now in order to deal
2831    with dynamic and symbolic parts of the core environment.
2832    
2833    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2834    compilation manager's bootstrap compiler (actually: in the "init group"
2835    handling).  See the comments in src/system/smlnj/init/init.cmi for
2836    more details.
2837    
2838    I also tried to track down all mentions of "Core" (as string argument
2839    to Symbol.strSymbol) in the compiler and replaced them with a
2840    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
2841    name appears in one place only.
2842    
2843    Binfile and bootfile format have not changed, but the switchover from
2844    the old "init.cmi" to the new one is a bit tricky, so I supplied new
2845    bootfiles anyway.
2846    
2847    ----------------------------------------------------------------------
2848    Name: Allen Leung
2849    Date: 2000/04/02 21:17:00 EST
2850    Tag: leunga-20000402-mltree
2851    Description:
2852    
2853       1. Renamed the constructor CALL in MLTREE by popular demand.
2854       2. Added a bunch of files from my repository.  These are currently
2855          used by other non-SMLNJ backends.
2856    
2857    ----------------------------------------------------------------------
2858    Name: Allen Leung
2859    Date: 2000/03/31 21:15:00 EST
2860    Tag: leunga-20000331-aliasing
2861    Description:
2862    
2863    This update contains a rewritten (and hopefully more correct) module
2864    for extracting aliasing information from CPS.
2865    
2866       To turn on this feature:
2867    
2868            Compiler.Control.CG.memDisambiguate := true
2869    
2870       To pretty print the region information with assembly
2871    
2872           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
2873    
2874       To control how many levels of aliasing information are printed, use:
2875    
2876           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
2877    
2878       The default of n is 3.
2879    
2880    ----------------------------------------------------------------------
2881    Name: David MacQueen
2882    Date: 2000/03/31 11:15:00 EST
2883    Tag: dbm-20000331-runtime_fix
2884    Description:
2885    
2886    This update contains:
2887    
2888    1. runtime/c-lib/c-libraries.c
2889       includes added in revision 1.2 caused compilation errors on hppa-hpux
2890    
2891    2. fix for bug 1556
2892       system/Basis/Implementation/NJ/internal-signals.sml
2893    
2894    ----------------------------------------------------------------------
2895    Name: Matthias Blume
2896    Date: 2000/03/31 18:00:00 JST
2897    Tag: blume_main_v110p26p2_1
2898    Description:
2899    
2900    This update contains:
2901    
2902    1. A small change to CM's handling of stable libraries:
2903       CM now maintains one "global" modmap that is used for all stable
2904       libraries.  The use of such a global modmap maximizes sharing and
2905       minimizes the need for re-traversing parts of environments during
2906       modmap construction.  (However, this has minor impact since modmap
2907       construction seems to account for just one percent or less of total
2908       compile time.)
2909    
2910    2. I added a "genmap" phase to the statistics.  This is where I got the
2911       "one percent" number (see above).
2912    
2913    3. CM's new tool parameter mechanism just became _even_ better. :)
2914       - The parser understands named parameters and recursive options.
2915       - The "make" and "shell" tools use these new features.
2916         (This makes it a lot easier to cascade these tools.)
2917       - There is a small syntax change: named parameters use a
2918    
2919           <name> : ( <option> ... )            or
2920           <name> : <string>
2921    
2922         syntax.  Previously, named parameters were implemented in an
2923         ad-hoc fashion by each tool individually (by parsing strings)
2924         and had the form
2925    
2926           <name>=<string>
2927    
2928       See the CM manual for a full description of these issues.
2929    
2930    ----------------------------------------------------------------------
2931    Name: Matthias Blume
2932    Date: 2000/03/30 18:00:00 JST
2933    Tag: blume_main_v110p26p2_0
2934    Description:
2935    
2936    !!!!! WARNING !!!!!!
2937    !!  New binfiles  !!
2938    !!!!!!!!!!!!!!!!!!!!
2939    
2940    This update contains:
2941    
2942    1. Moderate changes to CM:
2943    
2944       - Changes to CM's tools mechanism.  In particular, it is now possible
2945       to have tools that accept additional "command line" parameters
2946       (specified in the .cm file at each instance where the tool's class is
2947       used).
2948    
2949       This was done to accommodate the new "make" and "shell" tools which
2950       facilitate fairly seamless hookup to portions of code managed using
2951       Makefiles or Shell scripts.
2952    
2953       There are no classes "shared" or "private" anymore.  Instead, the
2954       sharing annotation is now a parameter to the "sml" class.
2955    
2956       There is a bit of generic machinery for implementing one's own
2957       tools that accept command-line parameters.  However, I am not yet fully
2958       satisfied with that part, so expect changes here in the future.
2959    
2960       All existing tools are described in the CM manual.
2961    
2962       - Slightly better error handling.  (CM now suppresses many followup
2963       error messages that tended to be more annoying than helpful.)
2964    
2965    2. Major changes to the compiler's static environment data structures.
2966    
2967       - no CMStaticEnv anymore.
2968            - no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
2969              but it is called Environment), no conversions between different
2970              kinds of static environments
2971    
2972       - There is still a notion of a "modmap", but such modmaps are generated
2973         on demand at the time when they are needed.  This sounds slow, but I
2974         sped up the code that generates modmaps enough for this not to lead to
2975         a slowdown of the compiler (at least I didn't detect any).
2976    
2977       - To facilitate rapid modmap generation, static environments now
2978         contain an (optional) "modtree" structure.  Modtree annotations are
2979         constructed by the unpickler during unpickling.  (This means that
2980         the elaborator does not have to worry about modtrees at all.)
2981         Modtrees have the advantage that they are compositional in the same
2982         way as the environment data structure itself is compositional.
2983         As a result, modtrees never hang on to parts of an environment that
2984         has already been rendered "stale" by filtering or rebinding.
2985    
2986       - I went through many, many trials and errors before arriving at the
2987         current solution.  (The initial idea of "linkpaths" did not work.)
2988         But the result of all this is that I have touched a lot of files that
2989         depend on the "modules" and "types" data structures (most of the
2990         elaborator). There were a lot of changes during my "linkpath" trials
2991         that could have been reverted to their original state but weren't.
2992         Please, don't be too harsh on me for messing with this code a bit more
2993         than what was strictly necessary...  (I _did_ resist the temptation
2994         of doing any "global reformatting" to avoid an untimely death at
2995         Dave's hands. :)
2996    
2997       - One positive aspect of the previous point:  At least I made sure that
2998         all files that I touched now compile without warnings (other than
2999         "polyEqual").
3000    
3001       - compiler now tends to run "leaner" (i.e., ties up less memory in
3002         redundant modmaps)
3003    
3004    ----------------------------------------------------------------------
3005    Name: Allen Leung
3006    Date: 2000/03/29 18:00:00
3007    Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
3008    Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
3009    Description:
3010    
3011       This update contains *MAJOR* changes to the way code is generated from CPS
3012    in the module mlriscGen, and in various backend modules.
3013    
3014    CHANGES
3015    =======
3016    
3017    1. MLRiscGen: forward propagation fix.
3018    
3019       There was a bug in forward propagation introduced at about the same time
3020       as the MLRISC x86 backend, which prohibits coalescing to be
3021       performed effectively in loops.
3022    
3023       Effect: speed up of loops in RISC architectures.
3024               By itself, this actually slowed down certain benchmarks on the x86.
3025    
3026    2. MLRiscGen:  forward propagating addresses from consing.
3027    
3028       I've changed the way consing code is generated.  Basically I separated
3029       out the initialization part:
3030    
3031            store tag,   offset(allocptr)
3032            store elem1, offset+4(allocptr)
3033            store elem2, offset+8(allocptr)
3034            ...
3035            store elemn, offset+4n(allocptr)
3036    
3037       and the address computation part:
3038    
3039            celladdr <- offset+4+alloctpr
3040    
3041       and move the address computation part
3042    
3043       Effect:  register pressure is generally lower as a result.  This
3044                makes compilation of certain expressions much faster, such as
3045                long lists with non-trivial elements.
3046    
3047                 [(0,0), (0,0), .... (0,0)]
3048    
3049    3. MLRiscGen: base pointer elimination.
3050    
3051        As part of the linkage mechanism, we generate the sequence:
3052    
3053         L:  ...  <- start of the code fragment
3054    
3055         L1:
3056             base pointer <- linkreg - L1 + L
3057    
3058         The base pointer was then used for computing relocatable addresses
3059       in the code fragment.  Frequently (such as in lots of continuations)
3060       this is not needed.  We now eliminate this sequence whenever possible.
3061    
3062         For compile time efficiency, I'm using a very stupid local heuristic.
3063       But in general, this should be done as a control flow analysis.
3064    
3065       Effect:  Smaller code size.  Speed up of most programs.
3066    
3067    4. Hppa back end
3068    
3069         Long jumps in span dependence resolution used to depend on the existence
3070      of the base pointer.
3071    
3072         A jump to a long label L was expanded into the following sequence:
3073    
3074          LDIL %hi(L-8192), %r29
3075          LDO  %lo(L-8192)(%r29), %r29
3076          ADD  %r29, baseptr, %r29
3077          BV,n %r0(%r29)
3078    
3079         In the presence of change (3) above, this will not work.  I've changed
3080       it so that the following sequence of instructions are generated, which
3081       doesn't mention the base pointer at all:
3082    
3083             BL,n  L', %r29           /* branch and link, L' + 4 -> %r29 */
3084        L':  ADDIL L-(L'+4), %r29     /* Compute address of L */
3085             BV,n  %r0(%r29)          /* Jump */
3086    
3087    5. Alpha back end
3088    
3089          New alpha instructions LDB/LDW have been added, as per Fermin's
3090       suggestions.   This is unrelated to all other changes.
3091    
3092    6. X86 back end
3093    
3094         I've changed andl to testl in the floating point test sequence
3095         whenever appropriate.  The Intel optimization guide states that
3096         testl is preferable to andl.
3097    
3098    7. RA (x86 only)
3099    
3100         I've improved the spill propagation algorithm, using an approximation
3101       of maximal weighted independent sets.   This seems to be necessary to
3102       alleviate the negative effect in light of the slow down in (1).
3103    
3104         I'll write down the algorithm one of these days.
3105    
3106    8. MLRiscGen: frequencies
3107    
3108         I've added an annotation that states that all call gc blocks have zero
3109       execution frequencies.  This improves register allocation on the x86.
3110    
3111    BENCHMARKS
3112    ==========
3113    
3114       I've only perform the comparison on 110.25.
3115    
3116       The platforms are:
3117    
3118        HPPA  A four processor HP machine (E9000) with 5G of memory.
3119        X86   A 300Hhz Pentium II with 128M of memory, and
3120        SPARC An Ultra sparc 2 with 512M of memory.
3121    
3122       I used the following parameters for the SML benchmarks:
3123    
3124                 @SMLalloc
3125         HPPA    256k
3126         SPARC   512k
3127         X86     256k
3128    
3129    COMPILATION TIME
3130    ----------------
3131       Here are the numbers comparing the compilation times of the compilers.
3132       I've only compared 110.25 compiling the new sources versus
3133       a fixpoint version of the new compiler compiling the same.
3134    
3135                     110.25                                  New
3136               Total  Time in RA  Spill+Reload   Total  Time In RA Spill+Reload
3137         HPPA   627s    116s        2684+3584     599s    95s       1003+1879
3138         SPARC  892s    173s        2891+3870     708s    116s      1004+1880
3139         X86    999s    315s       94006+130691   987s    296s    108877+141957
3140    
3141                   110.25         New
3142                Code Size      Code Size
3143         HPPA   8596736         8561421
3144         SPARC  8974299         8785143
3145         X86    9029180         8716783
3146    
3147       So in summary, things are at least as good as before.   Dramatic
3148       reduction in compilation is obtained on the Sparc; I can't explain it,
3149       but it is reproducible.  Perhaps someone should try to reproduce this
3150       on their own machines.
3151    
3152    SML BENCHMARKS
3153    --------------
3154    
3155        On the average, all benchmarks perform at least as well as before.
3156    
3157          HPPA         Compilation Time     Spill+Reload      Run Time
3158                     110.25  New            110.25    New   110.25  New
3159    
3160          barnesHut  3.158  3.015  4.75%    1+1       0+0   2.980  2.922   2.00%
3161              boyer  6.152  5.708  7.77%    0+0       0+0   0.218  0.213   2.34%
3162       count-graphs  1.168  1.120  4.32%    0+0       0+0  22.705 23.073  -1.60%
3163                fft  0.877  0.792 10.74%    1+3       1+3   0.602  0.587   2.56%
3164        knuthBendix  3.180  2.857 11.32%    0+0       0+0   0.675  0.662   2.02%
3165             lexgen  6.190  5.290 17.01%    0+0       0+0   0.913  0.788  15.86%
3166               life  0.803  0.703 14.22%   25+25      0+0   0.153  0.140   9.52%
3167              logic  2.048  2.007  2.08%    6+6       1+1   4.133  4.008   3.12%
3168         mandelbrot  0.077  0.080 -4.17%    0+0       0+0   0.765  0.712   7.49%
3169             mlyacc 22.932 20.937  9.53%  154+181    32+57  0.468  0.430   8.91%
3170            nucleic  5.183  5.060  2.44%    2+2       0+0   0.125  0.120   4.17%
3171      ratio-regions  3.357  3.142  6.84%    0+0       0+0  116.225 113.173 2.70%
3172                ray  1.283  1.290 -0.52%    0+0       0+0   2.887  2.855   1.11%
3173             simple  6.307  6.032  4.56%   28+30      5+7   3.705  3.658   1.28%
3174                tsp  0.888  0.862  3.09%    0+0       0+0   7.040  6.893   2.13%
3175               vliw 24.378 23.455  3.94%  106+127    25+45  2.758  2.707   1.91%
3176      --------------------------------------------------------------------------
3177       Average                     6.12%                                   4.09%
3178    
3179          SPARC        Compilation Time     Spill+Reload      Run Time
3180                     110.25  New            110.25    New   110.25  New
3181    
3182          barnesHut  3.778  3.592  5.20%    2+2       0+0   3.648  3.453    5.65%
3183              boyer  6.632  6.110  8.54%    0+0       0+0   0.258  0.242    6.90%
3184       count-graphs  1.435  1.325  8.30%    0+0       0+0  33.672 34.737   -3.07%
3185                fft  0.980  0.940  4.26%    3+9       2+6   0.838  0.827    1.41%
3186        knuthBendix  3.590  3.138 14.39%    0+0       0+0   0.962  0.967   -0.52%
3187             lexgen  6.593  6.072  8.59%    1+1       0+0   1.077  1.078   -0.15%
3188               life  0.972  0.868 11.90%   26+26      0+0   0.143  0.140    2.38%
3189              logic  2.525  2.387  5.80%    7+7       1+1   5.625  5.158    9.05%
3190         mandelbrot  0.090  0.093 -3.57%    0+0       0+0   0.855  0.728   17.39%
3191             mlyacc 26.732 23.827 12.19%  162+189    32+57  0.550  0.560   -1.79%
3192            nucleic  6.233  6.197  0.59%    3+3       0+0   0.163  0.173   -5.77%
3193      ratio-regions  3.780  3.507  7.79%    0+0       0+0 133.993 131.035   2.26%
3194                ray  1.595  1.550  2.90%    1+1       0+0   3.440  3.418    0.63%
3195             simple  6.972  6.487  7.48%   29+32      5+7   3.523  3.525   -0.05%
3196                tsp  1.115  1.063  4.86%    0+0       0+0   7.393  7.265    1.77%
3197               vliw 27.765 24.818 11.87%  110+135    25+45  2.265  2.135    6.09%
3198      ----------------------------------------------------------------------------
3199       Average                     6.94%                                    2.64%
3200    
3201          X86          Compilation Time     Spill+Reload      Run Time
3202                     110.25  New            110.25    New   110.25  New
3203    
3204          barnesHut  5.530  5.420  2.03%  593+893   597+915   3.532  3.440   2.66%
3205              boyer  8.768  7.747 13.19%  493+199   301+289   0.327  0.297  10.11%
3206       count-graphs  2.040  2.010  1.49%  298+394   315+457  26.578 28.660  -7.26%
3207                fft  1.327  1.302  1.92%  112+209   115+210   1.055  0.962   9.71%
3208        knuthBendix  5.218  5.475 -4.69%  451+598   510+650   0.928  0.932  -0.36%
3209             lexgen  9.970  9.623  3.60% 1014+841  1157+885   0.947  0.928   1.97%
3210               life  1.183  1.183  0.00%  162+182   145+148   0.127  0.103  22.58%
3211              logic  3.285  3.512 -6.45%  514+684   591+836   5.682  5.577   1.88%
3212         mandelbrot  0.147  0.143  2.33%   38+41     33+54    0.703  0.690   1.93%
3213             mlyacc 35.457 32.763  8.22% 3496+4564 3611+4860  0.552  0.550   0.30%
3214            nucleic  7.100  6.888  3.07%  239+168   201+158   0.175  0.173   0.96%
3215      ratio-regions  6.388  6.843 -6.65% 1182+257   981+300  120.142 120.345 -0.17%
3216                ray  2.332  2.338 -0.29%  346+398   402+494   3.593  3.540   1.51%
3217             simple  9.912  9.903  0.08% 1475+941  1579+1168  3.057  3.178  -3.83%
3218                tsp  1.623  1.532  5.98%  266+200   250+211   8.045  7.878   2.12%
3219               vliw 33.947 35.470 -4.29% 2629+2774 2877+3171  2.072  1.890   9.61%
3220      ----------------------------------------------------------------------------
3221       Average                     1.22%                                     3.36%
3222    
3223    ----------------------------------------------------------------------
3224    Name: Allen Leung
3225    Date: 2000/03/23 16:25:00
3226    Tag: leunga-20000323-fix_x86_alpha
3227    Description:
3228    
3229    1. X86 fixes/changes
3230    
3231       a.  The old code generated for SETcc was completely wrong.
3232           The Intel optimization guide is VERY misleading.
3233    
3234    2. ALPHA fixes/changes
3235    
3236       a.  Added the instructions LDBU, LDWU, STB, STW as per Fermin's suggestion.
3237       b.  Added a new mode byteWordLoadStores to the functor parameter to Alpha()
3238       c.  Added reassociation code for address computation.
3239    
3240    ----------------------------------------------------------------------
3241    Name: Allen Leung
3242    Date: 2000/03/22 01:23:00
3243    Tag: leunga-20000322-fix_x86_hppa_ra
3244    Description:
3245    
3246    1. X86 fixes/changes
3247    
3248       a.  x86Rewrite bug with MUL3 (found by Lal)
3249       b.  Added the instructions FSTS, FSTL
3250    
3251    2. PA-RISC fixes/changes
3252    
3253       a.  B label should not be a delay slot candidate!  Why did this work?
3254       b.  ADDT(32, REG(32, r), LI n) now generates one instruction instead of two,
3255           as it should be.
3256       c.  The assembly syntax for fstds and fstdd was wrong.
3257       d.  Added the composite instruction COMICLR/LDO, which is the immediate
3258           operand variant of COMCLR/LDO.
3259    
3260    3. Generic MLRISC
3261    
3262       a.  shuffle.sml rewritten to be slightly more efficient
3263       b.  DIV bug in mltree-simplify fixed (found by Fermin)
3264    
3265    4. Register Allocator
3266    
3267       a.  I now release the interference graph earlier during spilling.
3268           May improve memory usage.
3269    
3270    ----------------------------------------------------------------------
3271    Name: Matthias Blume
3272    Date: 2000/03/14 14:15:32
3273    Tag: blume_main_v110p26p1_2
3274    Description:
3275    
3276    1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
3277    additional argument called "template" which is an optional string that
3278    specifies the layout of the tool command line.  See the CM manual for
3279    explanation.
3280    
3281    2. A special-purpose tool can be "registered" by simply dropping the
3282    corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
3283    directory where the .cm file lives that uses this tool.  (The
3284    behavior/misfeature until now was to look for the tool description
3285    files in the current working directory.)  As before, tool description
3286    files could also be anchored -- in which case they can live anywhere
3287    they like.  Following the recent e-mail discussion, this change should
3288    make it easier to have special-purpose tools that are shipped together
3289    with the sources of the program that uses them.
3290    
3291    ----------------------------------------------------------------------
3292    Name: Matthias Blume
3293    Date: 2000/03/10 07:48:34
3294    Tag: blume_main_v110p26p1_1
3295    Description:
3296    
3297    I added a re-written version of Dave's fixpt script to src/system.
3298    Changes relative to the original version:
3299      - sh-ified (not everybody has ksh)
3300      - automatically figures out which architecture it runs on
3301      - uses ./makeml a bit more cleverly
3302      - never invokes ./installml (and, thus, does not clobber your
3303        good and working installation of sml in case something goes wrong)
3304      - accepts max iteration count using option "-iter <n>"
3305      - accepts a "base" name using option "-base <base>"
3306    
3307    It does not build any extraneous heap images but directly rebuilds
3308    bin- and boot-hierarchies using makeml's "-rebuild" switch. Finally,
3309    it can incorporate existing bin- and boot- hierarchies.  For example,
3310    suppose the base is set to "sml" (which is the default).  Then it
3311    successively builds
3312    
3313            sml.bin.<arch>-unix and sml.boot.<arch>-unix
3314    then    sml1.bin.<arch>-unix and sml1.boot.<arch>-unix
3315    then    sml2.bin.<arch>-unix and sml2.boot.<arch>-unix
3316    ...
3317    then    sml<n>.bin.<arch>-unix and sml<n>.boot.<arch>-unix
3318    
3319    and so on.  If any of these already exist, it will just use what's
3320    there.  In particular, many people will have the initial set of bin
3321    and boot files around, so this saves time for at least one full
3322    rebuild.  Having sets of the form <base><k>.{bin,boot}.<arch>-unix for
3323    <k>=1,2,... is normally not a good idea when invoking fixpt.  However,
3324    they might be the result of an earlier partial run of fixpt (which
3325    perhaps got accidentally killed).  In this case, fixpt will quickly
3326    move through what exists before continuing where it left off earlier,
3327    and, thus, saves a lot of time.
3328    
3329    ----------------------------------------------------------------------
3330    Name: Allen Leung
3331    Date: 00/03/10 02:20:00
3332    Tag: leunga-20000310-fix_x86_asm_ra
3333    Description:
3334    
3335    More assembly output problems involving the indexed addressing mode
3336    on the x86 have been found and corrected. Thanks to Fermin Reig for the
3337    fix.
3338    
3339    The interface and implementation of the register allocator have been changed
3340    slightly to accommodate the possibility to skip the register allocation
3341    phases completely and go directly to memory allocation.  This is needed
3342    for C-- use.
3343    
3344    ----------------------------------------------------------------------
3345    Name: Matthias Blume
3346    Date: 00/03/09 10:23:53
3347    Tag: blume_main_v110p26p1_0
3348    Description:
3349    
3350    * Complete re-organization of library names.  Many libraries have been
3351    consolidated so that they share the same path anchor.  For example,
3352    all MLRISC-related libraries are anchored at MLRISC, most libraries that
3353    are SML/NJ-specific are under "smlnj".  Notice that names like
3354    host-cmb.cm or host-compiler.cm no longer exist.  See system/README
3355    for a complete description of the new naming scheme.  Quick reference:
3356    
3357       host-cmb.cm        -> smlnj/cmb.cm
3358       host-compiler.cm   -> smlnj/compiler.cm
3359       full-cm.cm         -> smlnj/cm.cm
3360       <arch>-<os>.cm     -> smlnj/cmb/<arch>-<os>.cm
3361       <arch>-compiler.cm -> smlnj/compiler/<arch>.cm
3362    
3363    * Bug fixes in CM.
3364        - exceptions in user code are being passed through (i.e., reach top level)
3365        - more bugs in paranoia mode fixed
3366        - bug related to checking group owners fixed
3367    
3368    * New install.sh script that automagically fetches archive files:
3369      The new file config/srcarchiveurl must contain the URL of the
3370      (remote) directory that contains bin files (or other source archives).
3371      If install.sh does not find the archive locally, it tries to get
3372      it from that remote directory.
3373      This should simplify installation further:  For machines that have
3374      access to the internet, just fetch <version>-config.tgz, unpack it,
3375      edit config/targets, and go (run config/install.sh).  The script will
3376      fetch everything else that it might need all by itself.
3377    
3378      For CVS users, this mechanism is not relevant for source archives, but
3379      it is convenient for getting new sets of binfiles.
3380    
3381      Archives should be tar files compressed with either gzip, compress, or
3382      bzip2.  The script recognizes .tgz, .tar, tar.gz, tz, .tar.Z, and .tar.bz2.
3383    
3384    ----------------------------------------------------------------------
3385    Name: Matthias Blume
3386  Date: 2000/03/07 04:01:04  Date: 2000/03/07 04:01:04
3387  Tag: blume_main_v110_26_2  Tag: blume_main_v110_26_2
3388  Description:  Description:

Legend:
Removed from v.572  
changed lines
  Added in v.932

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