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 648, Thu May 25 21:28:21 2000 UTC revision 937, Thu Sep 20 21:20:28 2001 UTC
# Line 11  Line 11 
11  Date:  Date:
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15    ----------------------------------------------------------------------
16    Name: Matthias Blume
17    Date: 2001/09/20 17:20:00 EDT
18    Tag: blume-20010920-canonicalpaths
19    Description:
20    
21    0. Updated the BOOT file (something that I forgot to do earlier).
22    
23    1. Small internal change to CM so that it avoids "/../" in filenames
24       as much as possible (but only where it is safe).
25    
26    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
27       that arguments that contain delimiters are passed through correctly.
28       This change also means that all "special" arguments of the form
29       @SMLxxx... must come first.
30    
31    3. Changed install script to put relative anchor names for tool commands
32       into pathconfig.
33    
34    ----------------------------------------------------------------------
35    Name: Matthias Blume
36    Date: 2001/09/18 15:35:00 EDT
37    Tag: blume-20010918-readme11036
38    Description:
39    
40    Added README files.
41    
42    ----------------------------------------------------------------------
43    Name: Matthias Blume
44    Date: 2001/09/18 11:45:00 EDT
45    Tag: Release_110_36 (retag)
46    Description:
47    
48    Fixed mistake in config/preloads. Retagged as 110.36.
49    
50    ----------------------------------------------------------------------
51    Name: Matthias Blume
52    Date: 2001/09/18 09:40:00 EDT
53    Tag: Release_110_36_orig (tag changed)
54    Description:
55    
56    New version (110.36).  New bootfiles.
57    
58    ----------------------------------------------------------------------
59    Name: Matthias Blume
60    Date: 2001/09/14 16:15:00 EDT
61    Tag: blume-20010914-x86fastfp
62    Description:
63    
64    John committed some changes that Allen made, in particular a (hopefully)
65    correctly working version of the x86-fp module.
66    
67    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
68    flag to "true".  Everything seems to compile to a fixpoint ok, and
69    "mandelbrot" speeds up by about 15%.
70    
71    ----------------------------------------------------------------------
72    Name: Matthias Blume
73    Date: 2001/09/13 11:20:00 EDT
74    Tag: blume-20010913-minimal
75    Description:
76    
77    1. Stefan Monnier's patch to fix a miscompilation problem that
78       was brought to light by John Reppy's work on Moby.
79    
80    2. Implemented a minimal "structure Compiler" that contains just
81       "version" and "architecture".  The minimal version will be
82       available when the full version is not.  This is for backward-
83       compatibility with code that wants to test Compiler.version.
84    
85    ----------------------------------------------------------------------
86    Name: Matthias Blume
87    Date: 2001/08/28 14:03:00 EDT
88    Tag: blume-20010828-ml-lex
89    Description:
90    
91    Fix for bug 1581, received from Neophytos Michael.
92    
93    ----------------------------------------------------------------------
94    Name: Matthias Blume
95    Date: 2001/08/27 11:20:00 EDT
96    Tag: blume-20010827-readme11035
97    Description:
98    
99    Fleshed out the README file for 110.35.
100    
101    ----------------------------------------------------------------------
102    Name: Matthias Blume
103    Date: 2001/08/24 17:10:00 EDT
104    Tag: Release_110_35
105    Description:
106    
107    New version number (110.35).  New bootfiles.
108    
109    ----------------------------------------------------------------------
110    Name: Lal George
111    Date: 2001/08/24 13:47:18 EDT 2001
112    Tag: george-20010824-MLRISC-graphs
113    Description:
114    
115     removed clusters from MLRISC completely and replaced with graphs.
116    
117    ----------------------------------------------------------------------
118    Name: Matthias Blume
119    Date: 2001/08/23 17:50:00 EDT
120    Tag: blume-20010823-toplevel
121    Description:
122    
123    - some reorganization of the code that implements various kinds of
124      environments in the compiler (static, dynamic, symbolic, combined)
125    - re-implemented the EnvRef module so that evalStream works properly
126      (if the stream contains references to "use", "CM.make", etc.)
127    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
128    
129    ----------------------------------------------------------------------
130    Name: Matthias Blume
131    Date: 2001/08/20 15:50 EDT
132    Tag: blume20010820-slipup
133    Description:
134    
135    I forgot to commit a few files.  Here they are...
136    
137    ----------------------------------------------------------------------
138    Name: Matthias Blume
139    Date: 2001/08/20 15:35:00 EDT
140    Tag: blume-20010820-debugprof
141    Description:
142    
143    !!!! NEW BOOTFILES !!!!
144    
145    This is another round of reorganizing the compiler sources.  This
146    time the main goal was to factor out all the "instrumentation"
147    passes (for profiling and backtracing) into their own library.
148    The difficulty was to do it in such a way that it does not depend
149    on elaborate.cm but only on elabdata.cm.
150    
151    Therefore there have been further changes to both elaborate.cm and
152    elabdata.cm -- more "generic" things have been moved from the former
153    to the latter.  As a result, I was forced to split the assignment
154    of numbers indicating "primtyc"s into two portions: SML-generic and
155    SML/NJ-specific.  Since it would have been awkward to maintain,
156    I bit the bullet and actually _changed_ the mapping between these
157    numbers and primtycs.  The bottom line of this is that you need
158    a new set of bin- and bootfiles.
159    
160    I have built new bootfiles for all architectures, so doing a fresh
161    checkout and config/install.sh should be all you need.
162    
163    The newly created library's name is
164    
165        $smlnj/viscomp/debugprof.cm
166    
167    and its sources live under
168    
169        src/compiler/DebugProf
170    
171    ----------------------------------------------------------------------
172    Name: Matthias Blume
173    Date: 2001/08/15 17:15:00 EDT
174    Tag: blume-20010815-compreorg
175    Description:
176    
177    This is a first cut at reorganizing the CM libraries that make up the
178    core of the compiler.  The idea is to separate out pieces that could
179    be used independently by tools, e.g., the parser, the typechecker, etc.
180    
181    The current status is a step in this direction, but it is not quite
182    satisfactory yet.  Expect more changes in the future.
183    
184    Here is the current (new) organization...
185    
186        What used to be $smlnj/viscomp/core.cm is now divided into
187        six CM libraries:
188    
189             $smlnj/viscomp/basics.cm
190                           /parser.cm
191                           /elabdata.cm
192                           /elaborate.cm
193                           /execute.cm
194                           /core.cm
195    
196        The CM files for these libraries live under src/system/smlnj/viscomp.
197        All these libraries are proxy libraries that contain precisely
198        one CM library component.  Here are the locations of the components
199        (all within the src/compiler tree):
200    
201             Basics/basics.cm
202             Parse/parser.cm
203             ElabData/elabdata.cm
204             Elaborator/elaborate.cm
205             Execution/execute.cm
206             core.cm
207    
208         [This organization is the same that has been used already
209         for a while for the architecture-specific parts of the visible
210         compiler and for the old version of core.cm.]
211    
212         As you will notice, many source files have been moved from their
213         respective original locations to a new home in one of the above
214         subtrees.
215    
216         The division of labor between the new libraries is the following:
217    
218             basics.cm:
219                - Simple, basic definitions that pertain to many (or all) of
220                  the other libraries.
221             parser.cm:
222                - The SML parser, producing output of type Ast.dec.
223                - The type family for Ast is also defined and exported here.
224             elabdata.cm:
225                - The datatypes that describe input and output of the elaborator.
226                  This includes types, absyn, and static environments.
227             elaborator.cm:
228                - The SML/NJ type checker and elaborator.
229                  This maps an Ast.dec (with a given static environment) to
230                  an Absyn.dec (with a new static environment).
231                - This libraries implements certain modules that used to be
232                  structures as functors (to remove dependencies on FLINT).
233             execute.cm:
234                - Everything having to do with executing binary code objects.
235                - Dynamic environments.
236             core.cm:
237                - SML/NJ-specific instantiations of the elaborator and MLRISC.
238                - Top-level modules.
239                - FLINT (this should eventually become its own library)
240    
241    Notes:
242    
243    I am not 100% happy with the way I separated the elaborator (and its
244    data structures) from FLINT.  Two instances of the same problem:
245    
246        1. Data structures contain certain fields that carry FLINT-specific
247           information.  I hacked around this using exn and the property list
248           module from smlnj-lib.  But the fact that there are middle-end
249           specific fields around at all is a bit annoying.
250    
251        2. The elaborator calculates certain FLINT-related information.  I tried
252           to make this as abstract as I could using functorization, but, again,
253           the fact that the elaborator has to perform calculations on behalf
254           of the middle-end at all is not nice.
255    
256        3. Having to used exn and property lists is unfortunate because it
257           weakens type checking.  The other alternative (parameterizing
258           nearly *everything*) is not appealing, though.
259    
260    I removed the "rebinding =" warning hack because due to the new organization
261    it was awkward to maintain it.  As a result, the compiler now issues some of
262    these warnings when compiling init.cmi during bootstrap compilation. On
263    the plus side, you also get a warning when you do, for example:
264       val op = = Int32.+
265    which was not the case up to now.
266    
267    I placed "assign" and "deref" into the _Core structure so that the
268    code that deals with the "lazy" keyword can find them there.  This
269    removes the need for having access to the primitive environment
270    during elaboration.
271    
272    ----------------------------------------------------------------------
273    Name: Matthias Blume
274    Date: 2001/08/13
275    Tag: blume-20010813-closures
276    Description:
277    
278    This fix was sent to us by Zhong Shao.  It is supposed to improve the
279    performance of certain loops by avoiding needless closure allocation.
280    
281    ----------------------------------------------------------------------
282    Name: Lal George
283    Date: 2001/07/31 10:03:23 EDT 2001
284    Tag: george-20010731-x86-fmalloc
285    Description: Fixed bug in x86 calls
286    
287        There was a bug where call instructions would mysteriously
288        vanish. The call instruction had to be one that returned
289        a floating point value.
290    
291    ----------------------------------------------------------------------
292    Name: Lal George
293    Date: 2001/07/19 16:36:29 EDT 2001
294    Tag: george-20010719-simple-cells
295    Description:
296    
297    I have dramatically simplified the interface for CELLS in MLRISC.
298    
299    In summary, the cells interface is broken up into three parts:
300    
301      1. CellsBasis : CELLS_BASIS
302    
303            CellsBasis is a top level structure and common for all
304            architectures.  it contains the definitions of basic datatypes
305            and utility  functions over these types.
306    
307      2. functor Cells() : CELLS
308    
309            Cells generates an interface for CELLS that incorporates the
310            specific resources on the target architecture, such as the
311            presence of special register classes, their number and size,
312            and various useful substructures.
313    
314      3. <ARCH>CELLS
315    
316            e.g. SparcCells: SPARCCELLS
317    
318            <ARCH>CELLS usually contains additional bindings for special
319            registers  on the architecture, such as:
320    
321                    val r0 : cell           (* register zero *)
322                    val y : cell            (* Y register *)
323                    val psr : cell          (* processor status register *)
324                    ...
325    
326            The structure returned by applying the Cells functor is opened
327            in this interface.
328    
329    The main implication of all this is that the datatypes for cells is
330    split between CellsBasis and CELLS -- a fairly simple change for user
331    code.
332    
333    In the old scheme the CELLS interface had a definitional binding of
334    the form:
335    
336            signature CELLS = sig
337    
338               structure CellsBasis = CellsBasis
339    
340               ...
341    
342            end
343    
344    With all the sharing constraints that goes on in MLRISC, this old
345    design  quickly leads to errors such as:
346    
347            "structure definition spec inside of sharing ... "
348    
349    
350    and appears to require an unacceptable amount of sharing and where
351    constraint hackery.
352    
353    I think this error message (the interaction of definitional specs and
354    sharing) requires more explanation on our web page.
355    
356    ----------------------------------------------------------------------
357    Name: Matthias Blume
358    Date: 2001/07/19 15:00:00 EDT
359    Tag: blume-20010719-libreorg
360    Description:
361    
362    This update puts together a fairly extensive but straightforward change
363    to the way the libraries that implement the interactive system are
364    organized:
365    
366       The biggest change is the elimination of structure Compiler.  As a
367       replacement for this structure, there is now a CM library
368       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
369       that exports all the substructures of the original structure Compiler
370       directly.  So instead of saying Compiler.Foo.bar one now simply
371       says Foo.bar.  (The CM libraries actually export a collection of
372       structures that is richer than the collection of substructures of
373       structure Compiler.)
374    
375       To make the transition smooth, there is a separate library called
376       $smlnj/compiler/compiler.cm which puts together and exports the
377       original structure Compiler (or at least something very close to it).
378    
379       There are five members of the original structure Compiler
380       that are not exported directly but which instead became members
381       of a new structure Backend (described by signature BACKEND).  These are:
382       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
383       Interact (: INTERACT), structure Machine (: MACHINE), and val
384       architecture (: string).
385    
386       Structure Compiler.Version has become structure CompilerVersion.
387    
388       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
389       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
390       or x86, respectively.
391       Each of these exports the same frontend structures that
392       $smlnj/compiler.cm exports.  But they do not have a structure Backend
393       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
394       Hppa, PPC, Sparc, or X86, respectively.
395    
396       Library $smlnj/compiler/all.cm exports the union of the exports of
397       $smlnj/compiler/<arch>.cm
398    
399       There are no structures <Arch>Compiler anymore, use
400       $smlnj/compiler/<arch>.cm instead.
401    
402       Library host-compiler-0.cm is gone.  Instead, the internal library
403       that instantiates CM is now called cm0.cm.  Selection of the host
404       compiler (backend) is no longer done here but. (Responsibility for it
405       now lies with $smlnj/compiler/current.cm.  This seems to be more
406       logical.)
407    
408       Many individual files have been moved or renamed.  Some files have
409       been split into multiple files, and some "dead" files have been deleted.
410    
411    Aside from these changes to library organization, there are also changes
412    to the way the code itself is organized:
413    
414       Structure Binfile has been re-implemented in such a way that it no
415       longer needs any knowledge of the compiler.  It exclusively deals
416       with the details of binfile layout.  It no longer invokes the
417       compiler (for the purpose of creating new prospective binfile
418       content), and it no longer has any knowledge of how to interpret
419       pickles.
420    
421       Structure Compile (: COMPILE) has been stripped down to the bare
422       essentials of compilation.  It no longer deals with linking/execution.
423       The interface has been cleaned up considerably.
424    
425       Utility routines for dealing with linking and execution have been
426       moved into their own substructures.
427    
428       (The ultimate goal of these changes is to provide a light-weight
429       binfile loader/linker (at least for, e.g., stable libraries) that
430       does not require CM or the compiler to be present.)
431    
432    CM documentation has been updated to reflect the changes to library
433    organization.
434    
435    ----------------------------------------------------------------------
436    Name: Matthias Blume
437    Date: 2001/07/10 17:30:00 EDT
438    Tag: Release_110_34
439    Description:
440    
441    Minor tweak to 110.34 (re-tagged):
442    
443      - README.html file added to CVS repository
444      - runtime compiles properly under FreeBSD 3.X and 4.X
445    
446    ----------------------------------------------------------------------
447    Name: Matthias Blume
448    Date: 2001/07/10 17:30:00 EDT
449    Tag: Release_110_34
450    Description:
451    
452    New version number (110.34). New bootfiles.
453    
454    ----------------------------------------------------------------------
455    Name: Matthias Blume
456    Date: 2001/07/09 16:00:00 EDT
457    Tag: blume-20010709-more-varargs
458    Description:
459    
460    I changed the handling of varargs in ml-nlffigen again:
461    The ellipsis ... will now simply be ignored (with an accompanying warning).
462    
463    The immediate effect is that you can actually call a varargs function
464    from ML -- but you can't actually supply any arguments beyond the ones
465    specified explicitly.  (For example, you can call printf with its format
466    string, but you cannot pass additional arguments.)
467    
468    This behavior is only marginally more useful than the one before, but
469    it has the advantage that a function or, more importantly, a function
470    type never gets dropped on the floor, thus avoiding follow-up problems with
471    other types that refer to the offending one.
472    
473    ----------------------------------------------------------------------
474    Name: Matthias Blume
475    Date: 2001/07/09 11:25:00 EDT
476    Tag: blume-20010709-varargs
477    Description:
478    
479    1. ckit-lib.cm now exports structure Error
480    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
481       with a warning accompanied by a source location.  Moreover, it
482       merely skips the offending function or type and proceeds with the
483       rest of its work.u  As a result, one can safely feed C code containing
484       "..." to ml-nlffigen.
485    3. There are some internal improvements to CM, providing slightly
486       more general string substitutions in the tools subsystem.
487    
488    ----------------------------------------------------------------------
489    Name: Matthias Blume
490    Date: 2001/06/27 15:10:00 EDT
491    Tag: blume-20010627-concur
492    Description:
493    
494    Fixed a small bug in CM's handling of parallel compilation.
495    (You could observe the bug by Control-C-interrupting an ordinary
496    CMB.make or CM.stabilize and then attaching some compile servers.
497    The result was that all of a sudden the previously interrupted
498    compilation would continue on its own.  This was because of
499    an over-optimization: CM did not bother to clean out certain queues
500    when no servers were attached "anyway", resulting in the contents
501    of these queues to grab control when new servers did get attached.)
502    
503    There is also another minor update to the CM manual.
504    
505    ----------------------------------------------------------------------
506    Name: Matthias Blume
507    Date: 2001/06/26 16:15:00 EDT
508    Tag: blume-20010626-cmdoc
509    Description:
510    
511    Minor typo fixed in CM manual (syntax diagram for libraries).
512    
513    ----------------------------------------------------------------------
514    Name: Matthias Blume
515    Date: 2001/06/25 22:55:00 EDT
516    Tag: blume-20010625-x86pc
517    Description:
518    
519    Fixed a nasty bug in the X86 assembly code that caused signal
520    handlers to fail (crash) randomly.
521    
522    ----------------------------------------------------------------------
523    Name: Matthias Blume
524    Date: 2001/06/25 12:05:00 EDT
525    Tag: blume-20010625-nlffigen
526    Description:
527    
528    This update fixes a number of minor bugs in ml-nlffigen as reported by
529    Nick Carter <nbc@andrew.cmu.edu>.
530    
531      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
532      2. Default names for generated files are now derived from the name of
533         the C file *without its directory*.  In particular, this causes generated
534         files to be placed locally even if the C file is in some system directory.
535      3. Default names for generated signatures and structures are also derived
536         from the C file name without its directory.  This avoids silly things
537         like "structure GL/GL".
538         (Other silly names are still possible because ml-nlffigen does not do
539          a thorough check of whether generated names are legal ML identifiers.
540          When in doubt, use command line arguments to force particular names.)
541    
542    ----------------------------------------------------------------------
543    Name: Matthias Blume
544    Date: 2001/06/21 12:25:00 EDT
545    Tag: blume-20010621-eXene
546    Description:
547    
548    eXene now compiles and (sort of) works again.
549    
550    The library name (for version > 110.33) is $/eXene.cm.
551    
552    I also added an new example in src/eXene/examples/nbody.  See the
553    README file there for details.
554    
555    ----------------------------------------------------------------------
556    Name: Matthias Blume
557    Date: 2001/06/20 16:40:00 EDT
558    Tag: blume-20010620-cml
559    Description:
560    
561    CML now compiles and works again.
562    
563    Libraries (for version > 110.33):
564    
565      $cml/cml.cm            Main CML library.
566      $cml/basis.cm          CML's version of $/basis.cm.
567      $cml/cml-internal.cm   Internal helper library.
568      $cml/core-cml.cm       Internal helper library.
569      $cml-lib/trace-cml.cm  Tracing facility.
570      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
571    
572    The installer (config/install.sh) has been taught how to properly
573    install this stuff.
574    
575    ----------------------------------------------------------------------
576    Name: Matthias Blume
577    Date: 2001/06/19 17:55:00 EDT
578    Tag: blume-20010619-instantiate
579    Description:
580    
581    This un-breaks the fix for bug 1432.
582    (The bug was originally fixed in 110.9 but I broke it again some
583    time after that.)
584    
585    ----------------------------------------------------------------------
586    Name: Matthias Blume
587    Date: 2001/06/19 17:25:00 EDT
588    Tag: blume-20010619-signals
589    Description:
590    
591    This should (hopefully) fix the long-standing signal handling bug.
592    (The runtime system was constructing a continuation record with an
593    incorrect descriptor which would cause the GC to drop data on the floor...)
594    
595    ----------------------------------------------------------------------
596    Name: Matthias Blume
597    Date: 2001/06/15 15:05:00 EDT
598    Tag: blume-20010615-moresparc
599    Description:
600    
601    Here is a short late-hour update related to Sparc c-calls:
602    
603     -- made handling of double-word arguments a bit smarter
604    
605     -- instruction selection phase tries to collapse certain clumsily
606        constructed ML-Trees; typical example:
607    
608            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
609    
610        This currently has no further impact on SML/NJ since mlriscGen does
611        not seem to generate such patterns in the first place, and c-calls
612        (which did generate them in the beginning) has meanwhile been fixed
613        so as to avoid them as well.
614    
615    ----------------------------------------------------------------------
616    Name: Matthias Blume
617    Date: 2001/06/15 15:05:00 EDT
618    Tag: blume-20010615-sparc
619    Description:
620    
621    The purpose of this update is to provide an implementation of NLFFI
622    on Sparc machines.
623    
624    Here are the changes in detail:
625    
626       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
627       the Sparc implementation of the c-calls API.
628       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
629       for accessing the ML frame.  Thus, we have a real frame pointer and
630       can freely modify %sp without need for an omit-frame-ptr phase.
631       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
632       this case.
633       * ml-nlffigen has been taught to produce code for different architectures
634       and calling conventions.
635       * In a way similar to what was done in the x86 case, the Sparc
636       backend uses its own specific extension to mltree.  (For example,
637       it needs to be able to generate UNIMP instructions which are part
638       of the calling convention.)
639       * ml-nlffi-lib was reorganized to make it more modular (in particular,
640       to make it easier to plug in new machine- and os-dependent parts).
641    
642    There are some other fairly unrelated bug fixes and cleanups as well:
643    
644       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
645       that they properly share their libraries with existing SML/NJ libraries.
646       * I fixed a minor cosmetic bug in CM, supressing certain spurious
647       follow-up error messages.
648       * Updates to CM/CMB documentation.
649    
650    TODO items:
651    
652       * MLRISC should use a different register as its asmTemp on the Sparc.
653         (The current %o2 is a really bad choice because it is part of the
654         calling conventions, so things might interfere in unexpected ways.)
655    
656    ----------------------------------------------------------------------
657    Name: Matthias Blume
658    Date: 2001/06/07
659    Tag: blume-20010607-calls
660    Description:
661    
662    A number of internal changes related to C calls and calling conventions:
663    
664    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
665       number of bytes popped implicitly (by the callee).  In most cases
666       this field is 0 but on x86/win32 it is some non-zero value.  This
667       is information provided for the benefit of the "omit-frameptr" pass.
668    2. The CALL instruction on the x86 carries a similar "pops" field.
669       The instruction selection phase copies its value from the ML-Tree
670       CALL statement.
671    3. On all other architectures, the instruction selection phase checks
672       whether "pops=0" and complains if not.
673    4. The c-calls implementation for x86 now accepts two calling conventions:
674       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
675       up after the call and pops is set to 0.  For "stdcall", the caller
676       does nothing, leaving the cleanup to the callee; pops is set to
677       the number of bytes that were pushed onto the stack.
678    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
679       distinguish between "ccall" and "stdcall".
680    6. The UNIMP instruction has been added to the supported Sparc instruction
681       set. (This is needed for implementing the official C calling convention
682       on this architecture.)
683    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
684       work with the latest CM.
685    
686    ----------------------------------------------------------------------
687    Name: Matthias Blume
688    Date: 2001/06/05 15:10:00 EDT
689    Tag: blume-20010605-cm-index
690    Description:
691    
692    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
693    
694    1. CM can now generate "index files".  These are human-readable files
695       that list on a per-.cm-file basis each toplevel symbol defined or
696       imported.  The location of the index file for
697       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
698       To enable index-file generation, set CM.Control.generate_index to true
699       or export an environment-symbol: export CM_GENERATE_INDEX=true.
700    
701       The CM manual has been updated accordingly.
702    
703    2. I made some slight modifications to the c-calls API in MLRISC.
704    
705         a) There is now a callback to support saving/restoring of
706            dedicated but caller-save registers around the actual call
707            instruction.
708         b) One can optionally specify a comment-annotation for the
709            call instruction.
710    
711    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
712       (For example, the comment annotation shows the C prototype of
713        the function being called.)
714    
715    ----------------------------------------------------------------------
716    Name: Matthias Blume
717    Date: 2001/06/01 13:30:00 EDT
718    Tag: blume-20010601-nlffi-cleanup
719    Description:
720    
721    This is mostly a cleanup of MLFFI stuff:
722    
723       - some signature files have been put into a more exposed place
724       - the ugly 'f type parameter is gone (simplifies types tremendously!)
725       - ml-nlffigen changed accordingly
726       - tutorial updated
727    
728    Other changes:
729    
730       - author's affiliation in CM manual(s) updated
731       - some more recognized keywords added to Allen's sml.sty
732    
733    ----------------------------------------------------------------------
734    Name: Matthias Blume
735    Date: 2001/05/25 15:30:00 EDT
736    Tag: blume-20010525-iptr
737    Description:
738    
739      - put the official 110.33-README (as it appears on the ftp server) under
740        CVS
741      - fixed a small bug related to incomplete pointer types in
742        ml-nlffigen
743      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
744        (it does not need the 'f type parameter)
745    
746    ----------------------------------------------------------------------
747    Name: Matthias Blume
748    Date: 2001/05/23 14:30:00 EDT
749    Tag: Release_110_33
750    Description:
751    
752    New version number (110.33).  New bootfiles.
753    
754    ----------------------------------------------------------------------
755    Name: Matthias Blume
756    Date: 2001/05/22 18:06:00 EDT
757    Tag: blume-20010522-targets
758    Description:
759    
760    Made install.sh use file config/targets.customized if it exists, falling
761    back to config/targets if it doesn't.  This way one can have a customized
762    version of the targets file without touching the "real thing", thus
763    eliminating the constant fear of accidentally checking something bogus
764    back into the CVS repository...  (File config/targets.customized must
765    not be added to the repository!)
766    
767    ----------------------------------------------------------------------
768    Name: Matthias Blume
769    Date: 2001/05/22 16:30:00 EDT
770    Tag: blume-20010522-minitut
771    Description:
772    
773    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
774       struct returns.
775    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
776       incomplete, preliminary documentation for NLFFI.
777    
778    ----------------------------------------------------------------------
779    Name: Matthias Blume
780    Date: 2001/05/14 11:30:00 EDT
781    Tag: blume-20010514-script
782    Description:
783    
784    Some bugs in install script fixed.
785    
786    In addition to that I also made a slight change to the NLFFI API:
787    Functors generated by ml-nlffigen now take the dynamic library as a
788    straight functor argument, not as a suspended one.  (The original
789    functor code used to force the suspension right away anyway, so there
790    was nothing gained by this complication of the interface.)
791    
792    ----------------------------------------------------------------------
793    Name: Matthias Blume
794    Date: 2001/05/11 14:35:00 EDT
795    Tag: blume-20010511-ml-nlffi
796    Description:
797    
798    I finally took the plunge and added my new FFI code to the main
799    repository.  For x86-linux it is now ready for prime-time.
800    
801    There are two new subdirectories of "src":
802    
803      - ml-nlffi-lib:
804           The utility library for programs using the FFI interface.
805           Here is the implementation of $/c.cm and its associated low-level
806           partners $/c-int.cm and $/memory.cm.
807      - ml-nlffigen:
808           A stand-alone program for generating ML glue code from C source
809           code.
810    
811    Building ml-nlffigen requires $/ckit-lib.cm.
812    
813    The config/install.sh script has been updates to do the Right Thing
814    (hopefully).
815    
816    Notice that the source tree for the C-Kit will not be put under "src"
817    but directly under the installation root directory.  (This is the
818    structure that currently exists on the CVS server when you check out
819    module "sml".)  Fortunately, config/install.sh knows about this oddity.
820    
821    Bugs: No documentation yet.
822    
823    ----------------------------------------------------------------------
824    Name: Matthias Blume
825    Date: 2001/05/09 16:35:00 EDT
826    Tag: blume-20010509-cpscontract
827    Description:
828    
829    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
830    wrapper/unwrapper elimination did not decrement usage counts and some
831    dead variables got overlooked by the dead-up logic.)
832    
833    ----------------------------------------------------------------------
834    Name: Lal George
835    Date: 2001/05/08  17:26:09 EDT
836    Tag: george-20010508-omit-frameptr
837    Description:
838    
839    Changes to implement the omit-frame-pointer optimization to support
840    raw C calls. For now, there is only support on the Intel x86, but
841    other architectures will follow as more experience is gained with this.
842    
843    
844    ----------------------------------------------------------------------
845    Name: Matthias Blume
846    Date: 2001/05/07 14:40:00 EDT
847    Tag: blume-20010507-proxies
848    Description:
849    
850    I made into "proxy libraries" all libraries that qualify for such a
851    change.  (A qualifying library is a library that has another library or
852    groups as its sole member and repeats that member's export list
853    verbatim.  A proxy library avoids this repetition by omitting its export
854    list, effectively inheriting the list that its (only) member exports.
855    See the CM manual for more explanation.)
856    The main effect is that explicit export lists for these libraries
857    do not have to be kepts in sync, making maintenance a bit easier.
858    
859    I also added copyright notices to many .cm-files.
860    
861    Last but not least, I made a new set of bootfiles.
862    
863    ----------------------------------------------------------------------
864    Name: Matthias Blume
865    Date: 2001/05/04 17:00:00 EDT
866    Tag: blume-20010504-cm-lsplit
867    Description:
868    
869    0. John merged pending changes to $/smlnj-lib.cm
870    
871    1. Allen's previous change accidentally backed out of one of Lal's
872       earlier changes.  I undid this mistake (re-introducing Lal's change).
873    
874    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
875       within the compiler where applicable.  There is some code simplification
876       because of that.
877    
878    3. The "split" phase (in FLINT) is now part of the default list of phases.
879       Compiler.Control.LambdaSplitting.* can be used to globally control the
880       lambda-splitting (cross-module-inlining) engine.  In addition to that,
881       it can now also be controlled on a per-source basis: CM has been taught
882       a new tool parameter applicable to ML source files.
883    
884       - To turn lambda-splitting off completely:
885            local open Compiler.Control.LambdaSplitting in
886                val _ = set Off
887            end
888       - To make "no lambda-splitting" the global default (but allow per-source
889         overriding); this is the initial setting:
890            local open Compiler.Control.LambdaSplitting in
891                val _ = set (Default NONE)
892            end
893       - To make "lambda-splitting with aggressiveness a" the global default
894         (and allow per-source overriding):
895            local open Compiler.Control.LambdaSplitting in
896                val _ = set (Default (SOME a))
897            end
898    
899       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
900         write (in the respective .cm-file):
901            a.sml (lambdasplitting:off)
902       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
903            a.sml (lambdasplitting:on)
904       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
905         <a> is a decimal non-negative integer):
906            a.sml (lambdasplitting:<a>)
907       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
908            a.sml (lambdasplitting:infinity)
909       - To use the global default for a.sml:
910            a.sml (lambdasplitting:default)
911         or simply
912            a.sml
913    
914    ----------------------------------------------------------------------
915    Name: Allen Leung
916    Date: 2001/05/04 01:57:00 EDT
917    Tag: leunga-20010504-sync
918    Description:
919    
920      MLRISC features.
921    
922      1. Fix to CMPXCHG instructions.
923      2. Changed RA interface to allow annotations in callbacks.
924      3. Added a new method to the stream interface to allow annotations updates.
925    
926    ----------------------------------------------------------------------
927    Name: Matthias Blume
928    Date: 2001/05/01 11:45:00 EDT
929    Tag: blume-20010501-pcedittmp
930    Description:
931    
932    Changed install.sh to use the current working directory instead of
933    /usr/tmp for a temporary file (pcedittmp).  The previous choice
934    of /usr/tmp caused trouble with MacOS X because of file premission
935    problems.
936    
937    ----------------------------------------------------------------------
938    Name: Matthias Blume
939    Date: 2001/04/20 11:10:00 EDT
940    Tag: blume-20010420-inMLflag
941    Description:
942    
943     - added vp_limitPtrMask to vproc-state.h
944       (for use by the raw-C-calls mechanism to implement proper interrupt
945        handling)
946     - made the ML compiler aware of various data-structure offsets so it
947       can generate code for accessing the vp_inML flag and vp_limitPtrMask
948     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
949       raw C-calls
950    
951    ----------------------------------------------------------------------
952    Name: Lal George
953    Date: 2001/04/20 09:15:28 EDT
954    Tag: george-20010420-macosX
955    Description:
956    
957     - Changes to port to Mac OS X; Darwin.
958    
959     - In the process I found that sqrt was broken on the PPC, because the
960       fsqrt instruction is not implemented.
961    
962    ----------------------------------------------------------------------
963    Name: Matthias Blume
964    Date: 2001/04/18 12:45:00 EDT
965    Tag: blume-20010418-ccalls
966    Description:
967    
968     - fixed two off-by-4 errors in the x86-specific c-calls implementation
969       (this bug prevented structure arguments containing pointers from being
970        passed correctly)
971     - changed the raw-C-call code in mlriscGen.sml in such a way that
972       structure arguments are represented as a pointer to the beginning
973       of the structure (instead of having a series of synthesized arguments,
974       one for each structure member)
975    
976     - made makeml script's verbosity level configurable via environment
977       variable (MAKEML_VERBOSITY)
978    
979     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
980       in rawmem-x86.sml; we are now using the real thing
981    
982    ----------------------------------------------------------------------
983    Name: Matthias Blume
984    Date: 2001/03/22 16:25:00 EST
985    Tag: blume-20010322-bootfiles
986    Description:
987    
988    Created a new set of bootfiles (for your automatic installation convenience).
989    
990    ----------------------------------------------------------------------
991    Name: Matthias Blume
992    Date: 2001/03/22 15:10:00 EST
993    Tag: blume-20010322-rawmem-parcm
994    Description:
995    
996    1. All "raw memory access" primitives for the new FFI are implemented now
997       (at least on the x86).
998    2. Some further cleanup of CM's parallel make mechanism.
999    
1000    ----------------------------------------------------------------------
1001    Name: Matthias Blume
1002    Date: 2001/03/19 17:53:00 EST
1003    Tag: blume-20010319-parallel
1004    Description:
1005    
1006    Parallel make (using compile servers) now works again.
1007    
1008    To this end, CM.stabilize and CMB.make have been modified to work in
1009    two passes when compile servers are attached:
1010       1. Compile everything, do not perform stabilization; this pass
1011          uses compile servers
1012       2. Stabilize everything; this pass does not use compile servers
1013    If there are no compile servers, the two passes are combined into one
1014    (as before).  Splitting the passes increases the inherent parallelism
1015    in the dependency graph because the entire graph including all
1016    libraries is available at the same time.  This, in turn, improves
1017    server utilization.  The downside is that the master process will
1018    have to do some extra work after compilation is done (because for
1019    technical reasons it must re-read all the binfiles during stabilization).
1020    
1021    ----------------------------------------------------------------------
1022    Name: Matthias Blume
1023    Date: 2001/03/16 12:22:00 EST
1024    Tag: blume-20010316-bootfiles
1025    Description:
1026    
1027    Created a new set of bootfiles (for your automatic installation convenience).
1028    
1029    ----------------------------------------------------------------------
1030    Name: Matthias Blume
1031    Date: 2001/03/16 11:00:00 EST
1032    Tag: blume-20010316-MLTREE-fixup
1033    Description:
1034    
1035    This is a minor fixup for an (untagged) earlier commit by Allen.
1036    (A file was missing).
1037    
1038    ----------------------------------------------------------------------
1039    Name: Allen Leung
1040    Date: Mon Mar  5 18:54:57 EST 2001
1041    Tag: leunga-20010305-cut-support
1042    
1043    1. New support for alternative control-flow in MLTREE.
1044       Currently we support
1045    
1046          FLOW_TO(CALL ...., [k1,...,kn])
1047    
1048       This is needed for 'cuts to' in C-- and try/handle-like constructs
1049       in Moby
1050    
1051       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
1052    
1053    2. Register Allocator
1054    
1055       Changes in interface [from Fermin, John]
1056    
1057    3. Alpha 8-bit SLL support [Fermin]
1058    
1059    4. All architectures
1060    
1061       A new module (ClusterExpandCopies) for expanding parallel copies.
1062    
1063    ----------------------------------------------------------------------
1064    Name: Allen Leung
1065    Date: 2001/02/27 23:07:00 EST
1066    Tag: leunga-20010227-minor-stuff
1067    
1068    1. Alpha bug fix for CMOVNE
1069    2. Handle mltree COND(..,FCMP ...,...)
1070    3. Bug fix in simplifier
1071    
1072    ----------------------------------------------------------------------
1073    Name: Matthias Blume
1074    Date: 2001/01/30 17:50:00 EST
1075    Tag: blume-20010130-sync
1076    Description:
1077    
1078    This is just a minor update to sync my devel branch with the main brach.
1079    The only visible change is the addition of some README files.
1080    
1081    ----------------------------------------------------------------------
1082    Name: Matthias Blume
1083    Date: 2001/01/12 23:30:00 JST
1084    Tag: blume-20010112-bootfiles
1085    Description:
1086    
1087    Made a new set of bootfiles that goes with the current state of the
1088    repository.
1089    
1090    ----------------------------------------------------------------------
1091    Name: Matthias Blume
1092    Date: 2001/01/12 21:20:00 JST
1093    Tag: blume-20010112-sync
1094    Description:
1095    
1096    I am just flushing out some minor changes that had accumulated in
1097    my private branch in order to sync with the main tree.  (This is
1098    mainly because I had CVS trouble when trying to merge _into_ my
1099    private branch.)
1100    
1101    Most people should be completely unaffected by this.
1102    
1103    ----------------------------------------------------------------------
1104    Name: Allen Leung
1105    Date: Thu Jan 11 21:03:00 EST 2001
1106    Tag: leunga-20010111-labexp=mltree
1107    Description:
1108    
1109    1.  Removed the type LabelExp and replace it by MLTree.
1110    2.  Rewritten mltree-simplify with the pattern matcher tool.
1111    3.  There were some bugs in alpha code generator which would break
1112        64-bit code generation.
1113    4.  Redo the tools to generate code with the
1114    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
1115        are now generated by perl scripts.
1116    
1117    ----------------------------------------------------------------------
1118    Name: Matthias Blume
1119    Date: 2001/01/10 21:55:00 JST
1120    Tag: blume-20010110-rcc
1121    Description:
1122    
1123    The RCC stuff now seems to work (but only on the x86).
1124    This required hacking of the c-calls interface (and -implementation) in
1125    MLRISC.
1126    
1127    Normal compiler users should be unaffected.
1128    
1129    ----------------------------------------------------------------------
1130    Name: Matthias Blume
1131    Date: 2001/01/09 01:20:00 JST
1132    Tag: blume-20010109-rcc
1133    Description:
1134    
1135    This is a fairly big patch, flushing out a large number of pending
1136    changes that I made to my development copy over the last couple of days.
1137    
1138    Of practical relevance at this moment is a workaround for a pickling
1139    bug that Allen ran into the other day.  The cause of the bug itself is
1140    still unknown and it might be hard to fix it properly, but the
1141    workaround has some merits of its own (namely somewhat reducing pickling
1142    overhead for certain libraries).  Therefore, I think this solution should
1143    be satisfactory at this time.
1144    
1145    The rest of the changes (i.e., the vast majority) has to do with my
1146    ongoing efforts of providing direct support for C function calls from
1147    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
1148    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
1149    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
1150    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
1151    handle RCC.
1152    
1153    The last part (the changes to mlriscGen.sml) are still known to be
1154    wrong on the x86 and not implemented on all other architectures.  But
1155    the infrastructure is in place. I had to change a few functor
1156    signatures in the backend to be able to route the CCalls interface
1157    from MLRISC there, and I had to specialize the mltree type (on the
1158    x86) to include the necessary extensions. (The extensions themselves
1159    were already there and redy to go in MLRISC/x86).
1160    
1161    Everything should be very happy as soon as someone helps me with
1162    mlriscGen.sml...
1163    
1164    In any case, nothing of this should matter to anyone as long as the
1165    new primop is not being used (which is going to be the case unless you
1166    find it where I hid it :). The rest of the compiler is completely
1167    unaffected.
1168    
1169    ----------------------------------------------------------------------
1170    Name: Matthias Blume
1171    Date: 2001/01/05 00:30:00 JST
1172    Tag: blume-20010105-primops
1173    Description:
1174    
1175    Added some experimental support for work that I am doing right now.
1176    These changes mostly concern added primops, but there is also a new
1177    experimental C library in the runtime system (but currently not enabled
1178    anywhere except on Linux/X86).
1179    
1180    In the course of adding primops (and playing with them), I discovered that
1181    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
1182    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
1183    To recover, I made type information in INL_PRIM mandatory and changed
1184    prim.sml as well as built-in.sml accordingly.  The InLine structure now
1185    has complete, correct type information (i.e., no bottom types).
1186    
1187    Since all these changes mean that we need new binfiles, I also bumped the
1188    version number to 110.32.1.
1189    
1190    ----------------------------------------------------------------------
1191    Name: Matthias Blume
1192    Date: 2000/12/30 22:10:00 JST
1193    Tag: blume-20001230-various
1194    Description:
1195    
1196    Added proxy libraries for MLRISC and let MLRISC libraries refer
1197    to each other using path anchors.  (See CM manual for explanation.)
1198    
1199    Updated CM documentation.
1200    
1201    Fixed some bugs in CM.
1202    
1203    Implemented "proxy" libraries (= syntactic sugar for CM).
1204    
1205    Added "-quiet" option to makeml and changed runtime system accordingly.
1206    
1207    Added cleanup handler for exportML to reset timers and compiler stats.
1208    
1209    ----------------------------------------------------------------------
1210    Name: Lal George
1211    Date: 2000/12/22 22:22:58 EST 2000
1212    Tag: Release_110_32
1213    Description:
1214    
1215            Infinite precision used throughout MLRISC.
1216            see MLRISC/mltree/machine-int.sig
1217    
1218    ----------------------------------------------------------------------
1219    Name: Matthias Blume
1220    Date: 2000/12/22 23:16:00 JST
1221    Tag: blume-20001222-warn
1222    Description:
1223    
1224    Corrected wording and formatting of some CM warning message which I
1225    broke in my previous patch.
1226    
1227    ----------------------------------------------------------------------
1228    Name: Matthias Blume
1229    Date: 2000/12/22 21:20:00 JST
1230    Tag: blume-20001222-anchorenv
1231    Description:
1232    
1233    Fixed CM's handling of anchor environments in connection with CMB.make.
1234    
1235    ----------------------------------------------------------------------
1236    Name: Matthias Blume
1237    Date: 2000/12/22 13:15:00 JST
1238    Tag: blume-20001222-cleanup
1239    Description:
1240    
1241    Removed src/cm/ffi which does not (and did not) belong here.
1242    
1243    ----------------------------------------------------------------------
1244    Name: Matthias Blume
1245    Date: 2000/12/21 23:55:00 JST
1246    Tag: blume-20001221-exn
1247    Description:
1248    
1249    Probably most important: CM no longer silently swallows all exceptions
1250    in the compiler.
1251    Plus: some other minor CM changes.  For example, CM now reports some
1252    sizes for generated binfiles (code, data, envpickle, lambdapickle).
1253    
1254    ----------------------------------------------------------------------
1255    Name: Matthias Blume
1256    Date: 2000/12/15 00:01:05 JST
1257    Tag: blume-20001215-dirtool
1258    Description:
1259    
1260    - "dir" tool added.
1261    - improvements and cleanup to Tools structure
1262    - documentation updates
1263    
1264    ----------------------------------------------------------------------
1265    Name: Allen Leung
1266    Date: Thu Dec 14 03:45:24 EST 2000
1267    Description:
1268    Tag:  leunga-20001214-int-inf
1269    Description:
1270    
1271       In IntInf, added these standard functions, which are missing from our
1272    implementation:
1273    
1274        andb : int * int -> int
1275        xorb : int * int -> int
1276        orb  : int * int -> int
1277        notb : int -> int
1278         <<   : int * word -> int
1279        ~>>  : int * word -> int
1280    
1281       Not tested, I hope they are correct.
1282    
1283    ----------------------------------------------------------------------
1284    Name: Allen Leung
1285    Date: Fri Dec  8 19:23:26 EST 2000
1286    Description:
1287    Tag:  leunga-20001208-nowhere
1288    Description:
1289    
1290      Slight improvements to the 'nowhere' tool to handle OR-patterns,
1291    to generate better error messages etc.  Plus a brief manual.
1292    
1293    ----------------------------------------------------------------------
1294    Name: Lal George
1295    Date: 2000/12/08 09:54:02 EST 2000
1296    Tag: Release_110_31
1297    Description:
1298    
1299    - Version 110.31
1300    ----------------------------------------------------------------------
1301    Name: Allen Leung
1302    Date: Thu Dec  7 22:01:04 EST 2000
1303    Tag:  leunga-20001207-cell-monster-hack
1304    Description:
1305    
1306    Major MLRISC internal changes.  Affect all clients.
1307    Summary:
1308    
1309    1.  Type CELLS.cell = int is now replaced by a datatype.
1310        As a result, the old regmap is now gone.  Almost all interfaces
1311        in MLRISC change as a consequence.
1312    
1313    2.  A new brand version of machine description tool (v3.0) that generates
1314        modules expecting the new interface.  The old version is removed.
1315    
1316    3.  The RA interface has been further abstracted into two new functors.
1317        RISC_RA and X86RA.  These functors have much simpler interfaces.
1318        [See also directory MLRISC/demo.]
1319    
1320    4.  Some other new source->source code generation tools are available:
1321    
1322        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
1323        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
1324           I use this tool to generate the peephole optimizers---with the new
1325           cell type changes, peephole rules are becoming difficult to write
1326           without conditional pattern matching.
1327    
1328    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
1329        cover the entire MLRISC source tree so many things broke.
1330    
1331    6.  CM files have been moved to the subdirectory MLRISC/cm.
1332        They are moved because there are a lot of them and they clutter up the
1333        root dir.
1334    
1335    7.  More detailed documentation to come...
1336    
1337        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
1338        a makeml -rebuild first.  This is because of other other
1339        changes that Matthias has made (see below).
1340    
1341    
1342    ----------------------------------------------------------------------
1343    Name: Matthias Blume
1344    Date: 2000/11/30 23:12:00 JST
1345    Tag: blume-20001130-filereorg
1346    Description:
1347    
1348    Some manual updates and some file reorganizations in CM.
1349    
1350    ----------------------------------------------------------------------
1351    Name: Matthias Blume
1352    Date: 2000/11/24 17:45:00 JST
1353    Tag: blume-20001124-link
1354    Description:
1355    
1356    Drastically improved link traversal code for the case that the dynamic
1357    value was already loaded at bootstrap time.  As a result, CM and CMB
1358    now both load blazingly fast -- even on a very slow machine.  Also,
1359    memory consumption has been further reduced by this.
1360    
1361    Warning: The format of the PIDMAP file has changed.  THerefore, to
1362    bootstrap you have to do this:
1363    
1364    1. Run CMB.make
1365    2. Make a symbolic link for the boot directory:
1366         ln -s sml.boot.ARCH-OS xxx
1367    3. "Rebuild" the boot directory:
1368         ./makeml -boot xxx -rebuild sml ; rm xxx
1369    4. Boot normally:
1370          ./makeml
1371    
1372    ----------------------------------------------------------------------
1373    Name: Matthias Blume
1374    Date: 2000/11/21 21:20:00 JST
1375    Tag: blume-20001121-tools
1376    Description:
1377    
1378    Continued hacking on autoloading problem -- with success this time.
1379    Also changed tool-plugin mechanism.  See new CM manual.
1380    
1381    ----------------------------------------------------------------------
1382    Name: Matthias Blume
1383    Date: 2000/11/19 14:30:00 JST
1384    Tag:  blume-20001119-autoload
1385    Description:
1386    
1387    Some hacking to make autoloading faster.  Success for CMB, no success
1388    so far for CM.  There is a reduced structure CM' that autoloads faster.
1389    (This is a temporary, non-documented hack to be eliminated again when
1390    the general problem is solved.)
1391    
1392    ----------------------------------------------------------------------
1393    Name: Matthias Blume
1394    Date: 2000/11/17 14:10:00 JST
1395    Tag: blume-20001117-pickle-lib
1396    Description:
1397    
1398    1. Eliminated comp-lib.cm
1399    2. Made pickle-lib.cm
1400    3. Eliminated all uses of intset.sml (from comp-lib.cm)
1401    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
1402       equivalent constructs from smlnj-lib.cm (INtHashTable).
1403    5. Point 4. also goes for those uses of intmap.* in MLRISC.
1404       Duplicated intmap modules thrown out.
1405    6. Hunted down all duplicated SCC code and replaced it with
1406       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
1407    7. Rewrote Feedback module.
1408    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
1409       should be thrown out and equivalent modules from smlnj-lib.cm
1410       should be used (IntRedBlackSet, IntListSet, ...).
1411    
1412    Confirmed that compiler compiles to fixpoint.
1413    
1414    ----------------------------------------------------------------------
1415    Name: Allen Leung
1416    Date: 2000/11/10 18:00:00
1417    Tag: leunga-20001110-new-x86-fp
1418    
1419    A new x86 floating point code generator has been added.
1420    By default this is turned off.  To turn this on, do:
1421    
1422        CM.autoload "$smlnj/compiler.cm";
1423        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
1424    
1425    Changes:
1426    
1427    1.  Changed FTAN to FPTAN so that the assembly output is correct.
1428    2.  Changed the extension callback for FTANGENT to generate:
1429    
1430              fptan
1431              fstp  %st(0)
1432        instead of
1433              fptan
1434              fstpl ftempmem
1435    
1436    3.  Numerous assembly fixes for x86.
1437    
1438    5.  Cleaned up the machine code output module x86/x86MC.sml and added
1439        support for a whole bunch of instructions and addressing modes:
1440    
1441          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
1442          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
1443          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
1444          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
1445          fxch %st(n)
1446          fld %st(n)
1447          fst %st(n)
1448          fst mem
1449          fstp %st(n)
1450          fucom %st(n)
1451          fucomp %st(n)
1452    
1453        All these are now generated when the fast fp mode is turned on.
1454    
1455    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
1456    
1457    ----------------------------------------------------------------------
1458    Name: Matthias Blume
1459    Date: 2000/11/09 11:20:00 JST
1460    Tag: blume-20001109-scc
1461    Description:
1462    
1463    Eliminated some code duplication:
1464    
1465    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
1466       (Otherwise the functor is useless.)
1467    2. Used GraphSCCFn where SCCUtilFun was used previously.
1468    3. Got rid of SCCUtilFun (in comp-lib.cm).
1469    
1470    ----------------------------------------------------------------------
1471    Name: Lal George
1472    Date: 2000/11/06 09:02:21 EST 2000
1473    Tag: Release_110_30
1474    Description:
1475    
1476    - Version 110.30
1477    ----------------------------------------------------------------------
1478    Name: Matthias Blume
1479    Date: 2000/11/04 14:45:00
1480    Tag: blume-20001104-mlbuild
1481    Description:
1482    
1483    - Made ml-build faster on startup.
1484    - Documentation fixes.
1485    
1486    ----------------------------------------------------------------------
1487    Name: Matthias Blume
1488    Date: 2000/11/02 17:00:00 JST
1489    Tag: blume-20001102-condcomp
1490    Description:
1491    
1492    - Small tweaks to pickler -- new BOOTFILES!
1493    - Version bumped to 110.29.2.
1494    - Added conditional compilation facility to init.cmi (see comment there).
1495    ----------------------------------------------------------------------
1496    Name: Allen Leung
1497    Date: 2000/10/23 19:31:00
1498    Tag: leunga-20001023-demo-ra
1499    
1500    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
1501    2. Test programs for the graph library updated
1502    3. Some new MLRISC demo programs added
1503    
1504    ----------------------------------------------------------------------
1505    Name: Matthias Blume
1506    Date: 2000/08/31 22:15:00 JST
1507    Tag: blume-20001017-errmsg
1508    Description:
1509    
1510    More error message grief: Where there used to be no messages, there
1511    now were some that had bogus error regions.  Fixed.
1512    
1513    ----------------------------------------------------------------------
1514    Name: Matthias Blume
1515    Date: 2000/08/31 17:30:00 JST
1516    Tag: blume-20001017-v110p29p1
1517    Description:
1518    
1519    I made a version 110.29.1 with new bootfiles.
1520    
1521    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
1522              CM documentation changes and a small bugfix in CM's error reporting.
1523    
1524    ----------------------------------------------------------------------
1525    Name: Lal George
1526    Date: 2000/09/27 14:42:35 EDT
1527    Tag: george-20000927-nodestatus
1528    Description:
1529    
1530    Changed the type of the nodestatus, so that:
1531    
1532            SPILLED(~1)             is now SPILLED
1533            SPILLED(m) where m>=0   is now MEMREG(m)
1534            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
1535    
1536    ----------------------------------------------------------------------
1537    Name: Matthias Blume
1538    Date: 2000/09/07 14:45:00 JST
1539    Tag: blume-20000907-cmerrmsg
1540    Description:
1541    
1542    Small tweak to CM to avoid getting ML syntax error messages twice.
1543    
1544    ----------------------------------------------------------------------
1545    Name: Matthias Blume
1546    Date: 2000/08/31 18:00:00 JST
1547    Tag: blume-20000831-cvsbootfiles
1548    Description:
1549    
1550    New URL for boot files (because the 110.29 files on the BL server do
1551    now work correctly with my updated install scripts for yacc and lex).
1552    
1553    ----------------------------------------------------------------------
1554    Name: Matthias Blume
1555    Date: 2000/08/08 12:33:00 JST
1556    Tag: blume-20000808-manual
1557    Description:
1558    
1559    Tiny update to CM manual.
1560    
1561    ----------------------------------------------------------------------
1562    Name: Allen Leung
1563    Date: 2000/08/7 19:31:00
1564    Tag: leunga-20000807-a-whole-bunch-of-stuff
1565    
1566      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
1567    and Mobdy.
1568    
1569    1.  x86
1570    
1571       a.  Fixes to peephole module by John and Dan.
1572       b.  Assembly fix to SETcc by Allen.
1573       c.  Fix to c-call by John.
1574       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
1575       e.  Instruction selection optimization to SETcc as suggested by John.
1576    
1577           For example,
1578    
1579            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
1580    
1581           should generate:
1582    
1583            MOVL a, x
1584            SUBL b, x
1585            SHRL 31, x
1586    
1587    2.  IR stuff
1588    
1589         A bunch of new DJ-graph related algorithms added.  These
1590         speed up SSA construction.
1591    
1592    3.  SSA + Scheduling
1593    
1594         Added code for SSA and scheduling to the repository
1595    
1596    ----------------------------------------------------------------------
1597    Name: Lal George
1598    Date: 2000/07/27 11:53:14 EDT
1599    
1600    Tag: lal-20000727-linux-ppc
1601    Description:
1602    
1603     Made changes to support Linux PPC.
1604     p.s. I have confirmation that the 110.29 boot files work fine.
1605    
1606    ----------------------------------------------------------------------
1607    Name: Matthias Blume
1608    Date: 2000/07/27 17:40:00 JST
1609    Tag: blume-20000727-scripts
1610    Description:
1611    
1612    !!!! WARNING !!!!
1613    You must recompile the runtime system!
1614    !!!! WARNING !!!!
1615    
1616    This is basically another round of script-enhancements:
1617    
1618    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
1619       and undefine CM preprocessor symbols.
1620    
1621    2. ml-build avoids generating a new heap image if it finds that the
1622       existing one is still ok.  (The condition is that no ML file had to
1623       be recompiled and all ML files are found to be older that the heap
1624       file.)
1625    
1626       To make this work smoothly, I also hacked the runtime system as
1627       well as SMLofNJ.SysInfo to get access to the heap image suffix
1628       (.sparc-solaris, ...) that is currently being used.
1629    
1630       Moreover, the signature of CM.mk_standalone has changed.  See the
1631       CM manual.
1632    
1633    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
1634       CM manual for details.)
1635    
1636    4. More CM manual updates:
1637        - all of the above has been documented.
1638        - there is now a section describing the (CM-related) command line
1639          arguments that are accepted by the "sml" command
1640    
1641    ----------------------------------------------------------------------
1642    Name: Matthias Blume
1643    Date: 2000/07/25 16:20:00 JST
1644    Tag: blume-20000725-makedepend
1645    Description:
1646    
1647    Added a script called ml-makedepend.  This can be used in makefiles
1648    for Unix' make in a way very similar to the "makedepend" command for
1649    C.
1650    
1651    The script internally uses function CM.sources.
1652    
1653    Synopsis:
1654    
1655        ml-makedepend [-f makefile] cmfile targetname
1656    
1657    The default for the makefile is "makefile" (or "Makefile" should
1658    "makefile" not exist).
1659    
1660    ml-makedepend adds a cmfile/targetname-specific section to this
1661    makefile (after removing the previous version of this section).  The
1662    section contains a single dependency specification with targetname on
1663    the LHS (targetname is an arbitrary name), and a list of files derived
1664    from the cmfile on the RHS.  Some of the files on the RHS are
1665    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
1666    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
1667    path names.  The makefile writer is responsible for making sure that
1668    these variables have correct at the time "make" is invoked.
1669    
1670    ----------------------------------------------------------------------
1671    Name: Matthias Blume
1672    Date: 2000/07/22 23:30:00 JST
1673    Tag: blume-20000722-urlupdate
1674    Description:
1675    
1676    Changed BOOT and config/srcarchiveurl to point to BL server:
1677    
1678        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
1679    
1680    ----------------------------------------------------------------------
1681    Name: Matthias Blume
1682    Date: 2000/07/18 18:00:00 JST
1683    Tag: blume-20000718-Version_110_29
1684    Description:
1685    
1686    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
1687    
1688    2. Updated config/version to 110.29
1689    
1690    3. Updated config/srcarchiveurl
1691    
1692    3. New boot files!
1693       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
1694    
1695    ----------------------------------------------------------------------
1696    Name: Matthias Blume
1697    Date: 2000/07/11 13:58:00 JST
1698    Tag: blume-20000711-doctypo
1699    Description:
1700    
1701    Fixed a few typos in CM manual.
1702    
1703    ----------------------------------------------------------------------
1704    Name: Allen Leung
1705    Date: 2000/06/15 00:38:00
1706    Tag: leunga-20000704-sparc-x86
1707    
1708    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
1709    2. fix to x86 RET bug [found by Dan Grossman]
1710    3. sparc assembly bug fix for ticc instructions [found by Fermin]
1711    
1712       Affects c-- and moby only
1713    
1714    ----------------------------------------------------------------------
1715    Name: Matthias Blume
1716    Date: 2000/07/04 15:26:00
1717    Tag: blume-20000704-trigger
1718    Description:
1719    
1720    1. Improvements to CM manual.
1721    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
1722       of getting a back-trace.  The function, when called, raises an
1723       internal exception which explicitly carries the full back-trace history,
1724       so it is unaffected by any intervening handle-raise pairs ("trivial"
1725       or not).  The interactive loop will print that history once it arrives
1726       at top level.
1727       Short of having all exceptions implicitly carry the full history, the
1728       recommended way of using this facility is:
1729         - compile your program with instrumentation "on"
1730         - run it, when it raises an exception, look at the history
1731         - if the history is "cut off" because of some handler, go and modify
1732           your program so that it explicitly calls BTrace.trigger
1733         - recompile (still instrumented), and rerun; look at the full history
1734    
1735    ----------------------------------------------------------------------
1736    Name: Matthias Blume
1737    Date: 2000/07/03 15:36:00 JST
1738    Tag: blume-20000702-manual
1739    Description:
1740    
1741    Small corrections and updates to CM manual.
1742    
1743    ----------------------------------------------------------------------
1744    Name: Matthias Blume
1745    Date: 2000/06/29 16:04:00 JST
1746    Tag: blume-20000629-yacctool
1747    Description:
1748    
1749    Changes:
1750    
1751    1. Class "mlyacc" now takes separate arguments to pass options to
1752       generated .sml- and .sig-files independently.
1753    2. Corresponding CM manual updates.
1754    3. BTrace module now also reports call sites.  (However, for loop clusters
1755       it only shows from where the cluster was entered.)  There are associated
1756       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1757    
1758    ----------------------------------------------------------------------
1759    Name: Matthias Blume
1760    Date: 2000/06/27 16:51:00 JST
1761    Tag: blume-20000627-noweb
1762    Description:
1763    
1764    Changes:
1765    
1766     1. Implemented "subdir" and "witness" options for noweb tool.
1767        This caused some slight internal changes in CM's tool implementation.
1768     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
1769        some remaining issues from earlier path anchor changes.
1770     3. Updated CM manual accordingly.
1771    
1772     4. Changed implementation of back-tracing so that I now consider it
1773        ready for prime-time.
1774    
1775        In particular, you don't have to explicitly trigger the back-trace
1776        anymore.  Instead, if you are running BTrace-instrumented code and
1777        there is an uncaught exception (regardless of whether or not it was
1778        raised in instrumented code), the top-level evalloop will print
1779        the back-trace.
1780    
1781        Features:
1782    
1783          - Instrumented and uninstrumented code work together seemlessly.
1784            (Of course, uninstrumented code is never mentioned in actual
1785             back-traces.)
1786    
1787          - Asymptotic time- and space-complexity of instrumented code is
1788            equal to that of uninstrumented code.  (This means that
1789            tail-recursion is preserved by the instrumentation phase.)
1790    
1791          - Modules whose code has been instrumented in different sessions
1792            work together without problem.
1793    
1794          - There is no penalty whatsoever on uninstrumented code.
1795    
1796          - There is no penalty on "raise" expressions, even in
1797            instrumented code.
1798    
1799        A potential bug (or perhaps it is a feature, too):
1800    
1801          A back-trace reaches no further than the outermost instrumented
1802          non-trivial "raise".  Here, a "trivial" raise is one that is the
1803          sole RHS of a "handle" rule.  Thus, back-traces reach trough
1804    
1805               <exp> handle e => raise e
1806    
1807          and even
1808    
1809               <exp> handle Foo => raise Bar
1810    
1811          and, of course, through
1812    
1813               <exp> handle Foo => ...
1814    
1815         if the exception was not Foo.
1816    
1817         Back-traces always reach right through any un-instrumented code
1818         including any of its "handle" expressions, trivial or not.
1819    
1820       To try this out, do the following:
1821    
1822         - Erase all existing binfiles for your program.
1823           (You may keep binfiles for those modules where you think you
1824            definitely don't need back-tracing.)
1825         - Turn on back-trace instrumentation:
1826              SMLofNJ.Internals.BTrace.mode (SOME true);
1827         - Recompile your program.  (I.e., run "CM.make" or "use".)
1828         - You may now turn instrumentation off again (if you want):
1829              SMLofNJ.Internals.BTrace.mode (SOME false);
1830         - Run your program as usual.  If it raises an exception that
1831           reaches the interactive toplevel, then a back-trace will
1832           automatically be printed.  After that, the toplevel loop
1833           will print the exception history as usual.
1834    
1835    ----------------------------------------------------------------------
1836    Name: Matthias Blume
1837    Date: 2000/06/26 09:56:46 JST
1838    Tag: blume-20000626-setup
1839    Description:
1840    
1841    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1842          ML code before and after compiling a file (e.g., to set compiler
1843          flags)
1844    
1845    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1846              - associated changes to btrace.sml (BTrace instrumentation pass)
1847              - cleaner implementation of btimp.sml (BTrace tracing and report
1848                module)
1849    
1850    CM manual: * new path encoding documented
1851               * description of setup-parameter to "sml" added
1852    
1853    The biggest user-visible change to back-tracing is that it is no
1854    longer necessary to compile all traced modules within the same
1855    session.  (This was a real limitation.)
1856    
1857    ----------------------------------------------------------------------
1858    Name: Matthias Blume
1859    Date: 2000/06/24 12:40:00 JST
1860    Tag: blume-20000624-startup
1861    Description:
1862    
1863    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1864    bit too often -- to put it mildly. :)
1865    
1866    ----------------------------------------------------------------------
1867    Name: Matthias Blume
1868    Date: 2000/06/23 18:20:00 JST
1869    Tag: blume-20000623-btrace
1870    Description:
1871    
1872    This updates adds a backtrace facility to aid programmers in debugging
1873    their programs.  This involves the following changes:
1874    
1875    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1876       keeping track of the current call stack.  When programs are compiled
1877       in a special mode, the compiler will insert calls to these hooks
1878       into the user program.
1879       "Hook" means that it is possible for different implementations of
1880       back-tracing to register themselves (at different times).
1881    
1882    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1883       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1884       It can be turned on using this call:
1885         SMLofNJ.Internals.BTrace.mode (SOME true);
1886       Turning it off again:
1887         SMLofNJ.Internals.BTrace.mode (SOME false);
1888       Querying the current status:
1889         SMLofNJ.Internals.BTrace.mode NONE;
1890       Annotated programs are about twice as big as normal ones, and they
1891       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1892       where all hooks do nothing).  The slowdown with a plugin that is
1893       actually useful (such as the one supplied by default) is even greater,
1894       but in the case of the default plugin it is still only an constant
1895       factor (amortized).
1896    
1897    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1898       with a sub-structure BTrace for controlling back-tracing.  In particular,
1899       the above-mentioned function "mode" controls whether the annotation
1900       phase is invoked by the compiler.  Another important function is
1901       "trigger": when called it aborts the current execution and causes
1902       the top-level loop to print a full back-trace.
1903    
1904    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1905       for back-tracing.  It keeps track of the dynamic call stack and in
1906       addition to that it keeps a partial history at each "level" of that
1907       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1908       d and b (at separate times, dynamically), then the report will show:
1909    
1910       GOTO   d
1911             /c
1912       GOTO  \b
1913       CALL   a
1914    
1915       This shows that there was an initial non-tail call of a, then a
1916       tail-call to b or c, looping behavior in a cluster of functions that
1917       consist of b and c, and then a goto from that cluster (i.e., either from
1918       b or from c) to d.
1919    
1920       Note that (depending on the user program) the amount of information
1921       that the back-trace module has to keep track of at each level is bounded
1922       by a constant.  Thus, the whole implementation has the same asymptotical
1923       complexity as the original program (both in space and in time).
1924    
1925    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1926       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1927       which is raised by the "trigger" function mentioned above.
1928    
1929    Notes on usage:
1930    
1931    - Annotated code works well together with unannotated code:
1932    Unannotated calls simply do not show up at all in the backtrace.
1933    
1934    - It is not a good idea to let modules that were annotated during
1935    different sessions run at the same time.  This is because the compiler
1936    chooses small integers to identify individual functions, and there
1937    will be clashes if different modules were compiled in separate sessions.
1938    (Nothing will crash, and you will even be told about the clashes, but
1939    back-trace information will in general not be useful.)
1940    
1941    - Back-tracing can be confused by callcc and capture.
1942    
1943    - The only way of getting a back-trace right now is to explicitly
1944    invoke the "trigger" function from your user program.  Eventually, we
1945    should make every exception carry back-trace information (if
1946    available).  But since this creates more overhead at "raise"-time
1947    (similar to the current exnHistory overhead), I have not yet
1948    implemented this.  (The implementation will be rather easy.)  With
1949    exceptions carrying back-trace information, this facility will be even
1950    more useful because users don't need to modify their programs...
1951    
1952    - While it is possible to compile the compiler with back-trace
1953    annotations turned on (I did it to get some confidence in
1954    correctness), you must make absolutely sure that core.sml and
1955    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1956    be compiled with annotation because there is no core access yet, but
1957    if you compile btimp.sml with annotation, then the system will go into
1958    an infinite recursion and crash.)
1959    Since CM currently does not know about BTrace, the only way to turn
1960    annotations on and off for different modules of the compiler is to
1961    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1962    this is awkward and clumsy.
1963    
1964    Sample sessions:
1965    
1966    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1967    - SMLofNJ.Internals.BTrace.mode (SOME true);
1968    [autoloading]
1969    [autoloading done]
1970    val it = false : bool
1971    - structure X = struct
1972    -     fun main n = let
1973    -         fun a (x, 0) = d x
1974    -           | a (x, n) = b (x, n - 1)
1975    -         and b (x, n) = c (x, n)
1976    -         and c (x, n) = a (x, n)
1977    -         and d x = e (x, 3)
1978    -         and e (x, 0) = f x
1979    -           | e (x, n) = e (x, n - 1)
1980    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1981    -           | f n = n * g (n - 1)
1982    -         and g n = a (n, 3)
1983    -     in
1984    -         f n
1985    -     end
1986    - end;
1987    structure X : sig val main : int -> int end
1988    - X.main 3;
1989    *** BACK-TRACE ***
1990    GOTO   stdIn:4.2-13.20: X.main[2].f
1991    GOTO-( stdIn:4.2-13.20: X.main[2].e
1992    GOTO   stdIn:4.2-13.20: X.main[2].d
1993         / stdIn:4.2-13.20: X.main[2].a
1994         | stdIn:4.2-13.20: X.main[2].b
1995    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1996    CALL   stdIn:4.2-13.20: X.main[2].g
1997    GOTO   stdIn:4.2-13.20: X.main[2].f
1998    GOTO-( stdIn:4.2-13.20: X.main[2].e
1999    GOTO   stdIn:4.2-13.20: X.main[2].d
2000         / stdIn:4.2-13.20: X.main[2].a
2001         | stdIn:4.2-13.20: X.main[2].b
2002    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2003    CALL   stdIn:4.2-13.20: X.main[2].g
2004    GOTO   stdIn:4.2-13.20: X.main[2].f
2005    GOTO-( stdIn:4.2-13.20: X.main[2].e
2006    GOTO   stdIn:4.2-13.20: X.main[2].d
2007         / stdIn:4.2-13.20: X.main[2].a
2008         | stdIn:4.2-13.20: X.main[2].b
2009    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2010    CALL   stdIn:4.2-13.20: X.main[2].g
2011    GOTO   stdIn:4.2-13.20: X.main[2].f
2012    CALL   stdIn:2.15-17.4: X.main[2]
2013    -
2014    
2015    (Note that because of a FLINt bug the above code currently does not
2016    compile without BTrace turned on.)
2017    
2018    Here is another example, using my modified Tiger compiler:
2019    
2020    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
2021    - SMLofNJ.Internals.BTrace.mode (SOME true);
2022    [autoloading]
2023    [autoloading done]
2024    val it = false : bool
2025    - CM.make "sources.cm";
2026    [autoloading]
2027    ...
2028    [autoloading done]
2029    [scanning sources.cm]
2030    [parsing (sources.cm):parse.sml]
2031    [creating directory CM/SKEL ...]
2032    [parsing (sources.cm):tiger.lex.sml]
2033    ...
2034    [wrote CM/sparc-unix/semant.sml]
2035    [compiling (sources.cm):main.sml]
2036    [wrote CM/sparc-unix/main.sml]
2037    [New bindings added.]
2038    val it = true : bool
2039    - Main.compile ("../testcases/merge.tig", "foo.out");
2040    *** BACK-TRACE ***
2041    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
2042    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2043    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2044    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2045    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2046    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2047    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
2048         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
2049    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
2050    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2051    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
2052    CALL   main.sml:1.18-118.4: Main.compile[2]
2053    -
2054    
2055    ----------------------------------------------------------------------
2056    Name: Matthias Blumen
2057    Date: 2000/06/21 18:00:00 JST
2058    Tag: blume-20000621-manual
2059    Description:
2060    
2061    CM manual update: Path environments documented.
2062    
2063    ----------------------------------------------------------------------
2064    Name: Matthias Blume
2065    Date: 2000/06/19 13:40:00
2066    Tag: blume-20000619-manual
2067    Description:
2068    
2069    CM manual and system/README update.  This only covers the fact that
2070    there are no more implicit anchors.  (Path environments and the "bind"
2071    option to "cm" have yet to be documented.)
2072    
2073    ----------------------------------------------------------------------
2074    Name: Matthias Blume
2075    Date: 2000/06/19 11:05:00 JST
2076    Tag: blume-20000619-chdir-bugfix
2077    Description:
2078    
2079    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
2080    
2081    ----------------------------------------------------------------------
2082    Name: Matthias Blume
2083    Date: 2000/06/18 22:00:10 JST
2084    Tag: blume-20000618-implicit-anchors-really-gone
2085    Description:
2086    
2087    I updates the previous HISTORY entry where I forgot to mention that
2088    implicit anchors are no longer with us.
2089    
2090    The current update also gets rid of the (now useless) controller
2091    CM.Control.implicit_anchors.
2092    
2093    ----------------------------------------------------------------------
2094    Name: Matthias Blume
2095    Date: 2000/06/16 17:30:00 JST
2096    Tag: blume-20000616-anchorenv
2097    Description:
2098    
2099    This patch implements the long anticipated (just kidding :) "anchor
2100    environment" mechanism.  In the course of doing this, I also
2101    re-implemented CM's internal "SrcPath" module from scratch.  The new
2102    one should be more robust in certain boundary cases.  In any case, it
2103    is a lot cleaner than its predecessor (IMHO).
2104    
2105    This time, although there is yet another boot file format change, I
2106    kept the unpickler backward-compatible.  As a result, no new bootfiles
2107    are necessary and bootstrapping is straightforward.  (You cannot read
2108    new bootfiles into an old system, but the other way around is no
2109    problem.)
2110    
2111    Visible changes:
2112    
2113    ** 0. Implicit path anchors (without the leading $-symbol) are no
2114    longer recognized at all. This means that such path names are not
2115    illegal either.  For example, the name basis.cm simply refers to a
2116    local file called "basis.cm" (i.e, the name is an ordinary path
2117    relative to .cm-files directory).  Or, to put it differently, only
2118    names that start with $ are anchored paths.
2119    
2120    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
2121    vanished.
2122    
2123    John (Reppy) had critizised this as soon as I originally proposed and
2124    implemented it, but at that time I did not really deeply believe
2125    him. :) Now I came full-circle because I need the $<singlearc> syntax
2126    in another place where it cannot be seen as an abbreviation for
2127    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
2128    seems to mean (i.e., it "expands" into the corresponding anchor
2129    value).
2130    
2131    However, when paths are used as members in CM description files, it
2132    continues to be true that there must be at least another arc after the
2133    anchor.  This is now enforced separately during semantic analysis
2134    (i.e., from a lexical/syntactical point of view, the notation is ok.)
2135    
2136    ** 2. The "cm" class now accepts an option "bind".  The option's value
2137    is a sub-option list of precisely two items -- one labeled "anchor"
2138    and the other one labeled "value".  As you might expect, "anchor" is
2139    used to specify an anchor name to be bound, and "value" specifies what
2140    the anchor is being bound to.
2141    
2142    The value must be a directory name and can be given in either standard
2143    syntax (including the possibility that it is itself an anchored path)
2144    or native syntax.
2145    
2146    Examples:
2147    
2148       foo.cm (bind:(anchor:bar value:$mystuff/bar))
2149       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
2150    
2151    and so on.
2152    
2153    The meaning of this is that the .cm-file will be processed with an
2154    augmented anchor environment where the given anchor(s) is/are bound to
2155    the given values(s).
2156    
2157    The rationale for having this feature is this: Suppose you are trying
2158    to use two different (already stable) libraries a.cm and b.cm (that
2159    you perhaps didn't write yourself).  Further, suppose each of these
2160    two libraries internally uses its own auxiliary library $aux/lib.cm.
2161    Normally you would now have a problem because the anchor "lib" can not
2162    be bound to more than one value globally.  Therefore, the project that
2163    uses both a.cm and b.cm must locally redirect the anchor to some other
2164    place:
2165    
2166       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
2167       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
2168    
2169    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
2170    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
2171    
2172    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
2173    you when you do so at the time of CM.stabilize).  Therefore, you can
2174    also use an anchored path as the value:
2175    
2176      a.cm (bind:(anchor:lib value:$a-lib))
2177      b.cm (bind:(anchor:lib value:$b-lib))
2178    
2179    Now you can globally configure (using the usual CM.Anchor.anchor or
2180    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
2181    itself is always locally bound, setting it globally is no longer
2182    meaningful or necessary (but it does not hurt either).  In fact, "lib"
2183    can still be used as a global anchor for separate purposes.  As a
2184    matter of fact, one can locally define "lib" in terms of a global
2185    "lib":
2186    
2187      a.cm (bind:(anchor:lib value:$lib/a))
2188      b.cm (bind:(anchor:lib value:$lib/b))
2189    
2190    ** 3: The encoding of path names has changed.  This affects the way
2191    path names are shown in CM's progress report and also the internal
2192    protocol encoding used for parallel make.
2193    
2194    The encoding now uses one or more ':'-separated segments.  Each
2195    segments corresponds to a file that has been specified relative to the
2196    file given by its preceding segment.  The first segment is either
2197    relative to the CWD, absolute, or anchored.  Each segment itself is
2198    basically a Unix pathname; all segments but the first are relative.
2199    
2200    Example:
2201    
2202       $foo/bar/baz.cm:a/b/c.sml
2203    
2204    This path denotes the file bar/a/b/c.sml relative to the directory
2205    denoted by anchor "foo".  Notice that the encoding also includes
2206    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
2207    paths are resolved relative to the .cm-files directory, so baz.cm must
2208    be ignored to get the "real" pathname.
2209    
2210    To make this fact more obvious, CM puts the names of such "virtual
2211    arcs" into parentheses when they appear in progress reports. (No
2212    parentheses will appear in the internal protocol encoding.)  Thus,
2213    what you really see is:
2214    
2215      $foo/bar/(baz.cm):a/b/c.sml
2216    
2217    I find this notation to be much more informative than before.
2218    
2219    Another new feature of the encoding is that special characters
2220    including parentheses, colons, (back)slashes, and white space are
2221    written as \ddd (where ddd is the decimal encoding of the character).
2222    
2223    *** The CM manual still needs to be updated.
2224    
2225    ----------------------------------------------------------------------
2226    Name: Allen Leung
2227    Date: 2000/06/15 00:38:00
2228    Tag: leunga-20000615-x86-peephole
2229    
2230    x86 Peephole fix by Fermin.  Affects c-- and moby only.
2231    
2232    ----------------------------------------------------------------------
2233    Name: Matthias Blume
2234    Date: 2000/06/12 11:40:00
2235    Tag: blume-20000612-parmakefix
2236    Description:
2237    
2238    More cleanup after changing the file naming scheme: This time I
2239    repaired the parallel make mechanism for CMB.make which I broke earlier.
2240    
2241    ----------------------------------------------------------------------
2242    Name: Allen Leung
2243    Date: 2000/06/09 01:25:00
2244    Tag: leunga-20000609-various
2245    
2246    None of these things should affect normal SML/NJ operations
2247    
2248    1. Peephole improvements provided by Fermin (c--)
2249    2. New annotation DEFUSE for adding extra dependence (moby)
2250    3. New X86 LOCK instructions (moby)
2251    4. New machine description language for reservation tables (scheduling)
2252    5. Fixes to various optimization/analysis modules (branch chaining, dominator
2253       trees etc.)
2254    6. I've changed the CM files so that they can work with versions
2255       110.0.6, 110.25 and 110.28
2256    
2257    ----------------------------------------------------------------------
2258    Name: Matthias Blume
2259    Date: 2000/06/09 12:40:00
2260    Tag: blume-20000609-log
2261    Description:
2262    
2263    - Removed all(?) remaining RCS Log entries from sources.
2264    
2265    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
2266      anchored paths).
2267    
2268    ----------------------------------------------------------------------
2269    Name: Matthias Blume
2270    Date: 2000/06/07 17:00:00 JST
2271    Tag: blume-20000607-no-implicit-anchors
2272    Description:
2273    
2274    1. This update changes the default setting for
2275    CM.Control.implicit_anchors from true to false.  This means that
2276    implicit anchors are no longer permitted by default.  I also tried to
2277    make sure that nothing else still relies on implicit anchors.
2278    (This is the next step on the schedule towards a CM that does not even
2279    have the notion of implicit anchors anymore.)
2280    
2281    2. More CM manual updates.
2282    
2283    3. I managed to track down and fix the pickling bug I mentioned last
2284    time.  Because of the previously existing workaround, this entails no
2285    immediate practical changes.
2286    
2287    ----------------------------------------------------------------------
2288    Name: Matthias Blume
2289    Date: 2000/06/06 11:15:00 JST
2290    Tag: blume-20000606-lazierpickle
2291    Description:
2292    
2293    !!!! NEW BOOT FILES !!!!
2294    
2295    * The main purpose of this update is to make library pickles lazier in
2296    order to reduce the initial space penalty for autoloading a library.
2297    As a result, it is now possible to have $smlnj/compiler.cm
2298    pre-registered.  This should take care of the many complaints or
2299    inquiries about missing structure Compiler.  This required changes to
2300    CM's internal data structures and small tweaks to some algorithms.
2301    
2302    As a neat additional effect, it is no longer necessary (for the sake
2303    of lean heap image files) to distinguish between a "minimal" CM and a
2304    "full" CM.  Now, there is only one CM (i.e., the "full" version:
2305    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
2306    interactive top level. ($smlnj/cm/minimal.cm is gone.)
2307    
2308    To make the life of compiler-hackers easier, "makeml" now also
2309    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
2310    words, after you bootstrap a new sml for the first time, you will not
2311    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
2312    around you will still have to do it, though.)
2313    
2314    * A second change consists of major updates to the CM manual.  There
2315    are now several appendices with summary information and also a full
2316    specification of the CM description file syntax.
2317    
2318    * In directory src/system I added the script "allcross".  This script
2319    invokes sml and cross-compiles the compiler for all supported
2320    architectures.  (Useful when providing a new set of boot files.)
2321    
2322    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
2323    added a small tweak to pickle-util.sml to work around this problem,
2324    but it is not a proper fix yet.  I will investigate further.  (The
2325    effect of the bug was an inflation of library pickle size.)
2326    
2327    * Version number increased to 110.28.1 (to avoid compatibility problems).
2328    
2329  ----------------------------------------------------------------------  ----------------------------------------------------------------------
2330  Name: Allen Leung  Name: Allen Leung
2331  Date: 2000/05/25 17:28 EDT  Date: 2000/05/25 17:28 EDT
# Line 502  Line 2817 
2817    
2818  3.  Assembly  3.  Assembly
2819    
2820      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
2821      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
2822      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
2823    
# Line 525  Line 2840 
2840    
2841  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
2842  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
2843  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
2844  legal surface-syntax symbol).  legal surface-syntax symbol).
2845    
2846  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 535  Line 2850 
2850  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
2851    
2852  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2853  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
2854  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
2855  more details.  more details.
2856    
# Line 650  Line 2965 
2965     (specified in the .cm file at each instance where the tool's class is     (specified in the .cm file at each instance where the tool's class is
2966     used).     used).
2967    
2968     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
2969     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
2970     Makefiles or Shell scripts.     Makefiles or Shell scripts.
2971    
2972     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 663  Line 2978 
2978    
2979     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
2980    
2981     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
2982     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
2983    
2984  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 694  Line 3009 
3009       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3010       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3011       Please, don't be too harsh on me for messing with this code a bit more       Please, don't be too harsh on me for messing with this code a bit more
3012       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3013       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3014       Dave's hands. :)       Dave's hands. :)
3015    
# Line 797  Line 3112 
3112    
3113       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
3114       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
3115       testl is perferable to andl.       testl is preferable to andl.
3116    
3117  7. RA (x86 only)  7. RA (x86 only)
3118    
# Line 979  Line 3294 
3294    
3295  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
3296  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
3297  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
3298  explanation.  explanation.
3299    
3300  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
3301  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
3302  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
3303  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 1026  Line 3341 
3341  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
3342  <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,
3343  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
3344  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
3345  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
3346  and, thus, saves a lot of time.  and, thus, saves a lot of time.
3347    
# Line 1076  Line 3391 
3391    it from that remote directory.    it from that remote directory.
3392    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
3393    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
3394    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
3395    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
3396    
3397    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.648  
changed lines
  Added in v.937

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