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

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