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

Legend:
Removed from v.575  
changed lines
  Added in v.923

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