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 588, Fri Mar 31 09:00:02 2000 UTC revision 952, Tue Oct 9 14:00:40 2001 UTC
# Line 13  Line 13 
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Lal George
17    Date: 2001/10/09 09:41:37 EDT
18    Tag: george-20011008-mlrisc-labels
19    Description:
20    
21    Changed the machine description files to support printing of
22    local and global labels in assembly code, based on host assembler
23    conventions.
24    
25    ----------------------------------------------------------------------
26    Name: Matthias Blume
27    Date: 2001/09/25 15:25:00 EDT
28    Tag: blume-20010925-exninfo
29    Description:
30    
31    I provided a non-hook implementation of exnName (at the toplevel) and
32    made the "dummy" implementation of exnMessage (at the toplevel) more
33    useful: if nothing gets "hooked in", then at least you are going to
34    see the exception name and a message indicating why you don't see more.
35    
36    [For the time being, programs that need exnMessage and want to use
37    ml-build should either use General.exnMessage (strongly recommended) or
38    refer to structure General at some other point so that CM sees a
39    static dependency.]
40    
41    [Similar remarks go for "print" and "use":  If you want to use their
42    functionality in stand-alone programs generated by ml-build, then use
43    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
44    
45    ----------------------------------------------------------------------
46    Name: Matthias Blume
47    Date: 2001/09/20 17:28:00 EDT
48    Tag: blume-20010920-slowfp
49    Description:
50    
51    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
52    
53    ----------------------------------------------------------------------
54    Name: Matthias Blume
55    Date: 2001/09/20 17:20:00 EDT
56    Tag: blume-20010920-canonicalpaths
57    Description:
58    
59    0. Updated the BOOT file (something that I forgot to do earlier).
60    
61    1. Small internal change to CM so that it avoids "/../" in filenames
62       as much as possible (but only where it is safe).
63    
64    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
65       that arguments that contain delimiters are passed through correctly.
66       This change also means that all "special" arguments of the form
67       @SMLxxx... must come first.
68    
69    3. Changed install script to put relative anchor names for tool commands
70       into pathconfig.
71    
72    ----------------------------------------------------------------------
73    Name: Matthias Blume
74    >>>>>>> 1.169
75    Date: 2001/09/18 15:35:00 EDT
76    Tag: blume-20010918-readme11036
77    Description:
78    
79    Added README files.
80    
81    ----------------------------------------------------------------------
82    Name: Matthias Blume
83    Date: 2001/09/18 11:45:00 EDT
84    Tag: Release_110_36 (retag)
85    Description:
86    
87    Fixed mistake in config/preloads. Retagged as 110.36.
88    
89    ----------------------------------------------------------------------
90    Name: Matthias Blume
91    Date: 2001/09/18 09:40:00 EDT
92    Tag: Release_110_36_orig (tag changed)
93    Description:
94    
95    New version (110.36).  New bootfiles.
96    
97    ----------------------------------------------------------------------
98    Name: Matthias Blume
99    Date: 2001/09/14 16:15:00 EDT
100    Tag: blume-20010914-x86fastfp
101    Description:
102    
103    John committed some changes that Allen made, in particular a (hopefully)
104    correctly working version of the x86-fp module.
105    
106    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
107    flag to "true".  Everything seems to compile to a fixpoint ok, and
108    "mandelbrot" speeds up by about 15%.
109    
110    ----------------------------------------------------------------------
111    Name: Matthias Blume
112    Date: 2001/09/13 11:20:00 EDT
113    Tag: blume-20010913-minimal
114    Description:
115    
116    1. Stefan Monnier's patch to fix a miscompilation problem that
117       was brought to light by John Reppy's work on Moby.
118    
119    2. Implemented a minimal "structure Compiler" that contains just
120       "version" and "architecture".  The minimal version will be
121       available when the full version is not.  This is for backward-
122       compatibility with code that wants to test Compiler.version.
123    
124    ----------------------------------------------------------------------
125    Name: Matthias Blume
126    Date: 2001/08/28 14:03:00 EDT
127    Tag: blume-20010828-ml-lex
128    Description:
129    
130    Fix for bug 1581, received from Neophytos Michael.
131    
132    ----------------------------------------------------------------------
133    Name: Matthias Blume
134    Date: 2001/08/27 11:20:00 EDT
135    Tag: blume-20010827-readme11035
136    Description:
137    
138    Fleshed out the README file for 110.35.
139    
140    ----------------------------------------------------------------------
141    Name: Matthias Blume
142    Date: 2001/08/24 17:10:00 EDT
143    Tag: Release_110_35
144    Description:
145    
146    New version number (110.35).  New bootfiles.
147    
148    ----------------------------------------------------------------------
149    Name: Lal George
150    Date: 2001/08/24 13:47:18 EDT 2001
151    Tag: george-20010824-MLRISC-graphs
152    Description:
153    
154     removed clusters from MLRISC completely and replaced with graphs.
155    
156    ----------------------------------------------------------------------
157    Name: Matthias Blume
158    Date: 2001/08/23 17:50:00 EDT
159    Tag: blume-20010823-toplevel
160    Description:
161    
162    - some reorganization of the code that implements various kinds of
163      environments in the compiler (static, dynamic, symbolic, combined)
164    - re-implemented the EnvRef module so that evalStream works properly
165      (if the stream contains references to "use", "CM.make", etc.)
166    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
167    
168    ----------------------------------------------------------------------
169    Name: Matthias Blume
170    Date: 2001/08/20 15:50 EDT
171    Tag: blume20010820-slipup
172    Description:
173    
174    I forgot to commit a few files.  Here they are...
175    
176    ----------------------------------------------------------------------
177    Name: Matthias Blume
178    Date: 2001/08/20 15:35:00 EDT
179    Tag: blume-20010820-debugprof
180    Description:
181    
182    !!!! NEW BOOTFILES !!!!
183    
184    This is another round of reorganizing the compiler sources.  This
185    time the main goal was to factor out all the "instrumentation"
186    passes (for profiling and backtracing) into their own library.
187    The difficulty was to do it in such a way that it does not depend
188    on elaborate.cm but only on elabdata.cm.
189    
190    Therefore there have been further changes to both elaborate.cm and
191    elabdata.cm -- more "generic" things have been moved from the former
192    to the latter.  As a result, I was forced to split the assignment
193    of numbers indicating "primtyc"s into two portions: SML-generic and
194    SML/NJ-specific.  Since it would have been awkward to maintain,
195    I bit the bullet and actually _changed_ the mapping between these
196    numbers and primtycs.  The bottom line of this is that you need
197    a new set of bin- and bootfiles.
198    
199    I have built new bootfiles for all architectures, so doing a fresh
200    checkout and config/install.sh should be all you need.
201    
202    The newly created library's name is
203    
204        $smlnj/viscomp/debugprof.cm
205    
206    and its sources live under
207    
208        src/compiler/DebugProf
209    
210    ----------------------------------------------------------------------
211    Name: Matthias Blume
212    Date: 2001/08/15 17:15:00 EDT
213    Tag: blume-20010815-compreorg
214    Description:
215    
216    This is a first cut at reorganizing the CM libraries that make up the
217    core of the compiler.  The idea is to separate out pieces that could
218    be used independently by tools, e.g., the parser, the typechecker, etc.
219    
220    The current status is a step in this direction, but it is not quite
221    satisfactory yet.  Expect more changes in the future.
222    
223    Here is the current (new) organization...
224    
225        What used to be $smlnj/viscomp/core.cm is now divided into
226        six CM libraries:
227    
228             $smlnj/viscomp/basics.cm
229                           /parser.cm
230                           /elabdata.cm
231                           /elaborate.cm
232                           /execute.cm
233                           /core.cm
234    
235        The CM files for these libraries live under src/system/smlnj/viscomp.
236        All these libraries are proxy libraries that contain precisely
237        one CM library component.  Here are the locations of the components
238        (all within the src/compiler tree):
239    
240             Basics/basics.cm
241             Parse/parser.cm
242             ElabData/elabdata.cm
243             Elaborator/elaborate.cm
244             Execution/execute.cm
245             core.cm
246    
247         [This organization is the same that has been used already
248         for a while for the architecture-specific parts of the visible
249         compiler and for the old version of core.cm.]
250    
251         As you will notice, many source files have been moved from their
252         respective original locations to a new home in one of the above
253         subtrees.
254    
255         The division of labor between the new libraries is the following:
256    
257             basics.cm:
258                - Simple, basic definitions that pertain to many (or all) of
259                  the other libraries.
260             parser.cm:
261                - The SML parser, producing output of type Ast.dec.
262                - The type family for Ast is also defined and exported here.
263             elabdata.cm:
264                - The datatypes that describe input and output of the elaborator.
265                  This includes types, absyn, and static environments.
266             elaborator.cm:
267                - The SML/NJ type checker and elaborator.
268                  This maps an Ast.dec (with a given static environment) to
269                  an Absyn.dec (with a new static environment).
270                - This libraries implements certain modules that used to be
271                  structures as functors (to remove dependencies on FLINT).
272             execute.cm:
273                - Everything having to do with executing binary code objects.
274                - Dynamic environments.
275             core.cm:
276                - SML/NJ-specific instantiations of the elaborator and MLRISC.
277                - Top-level modules.
278                - FLINT (this should eventually become its own library)
279    
280    Notes:
281    
282    I am not 100% happy with the way I separated the elaborator (and its
283    data structures) from FLINT.  Two instances of the same problem:
284    
285        1. Data structures contain certain fields that carry FLINT-specific
286           information.  I hacked around this using exn and the property list
287           module from smlnj-lib.  But the fact that there are middle-end
288           specific fields around at all is a bit annoying.
289    
290        2. The elaborator calculates certain FLINT-related information.  I tried
291           to make this as abstract as I could using functorization, but, again,
292           the fact that the elaborator has to perform calculations on behalf
293           of the middle-end at all is not nice.
294    
295        3. Having to used exn and property lists is unfortunate because it
296           weakens type checking.  The other alternative (parameterizing
297           nearly *everything*) is not appealing, though.
298    
299    I removed the "rebinding =" warning hack because due to the new organization
300    it was awkward to maintain it.  As a result, the compiler now issues some of
301    these warnings when compiling init.cmi during bootstrap compilation. On
302    the plus side, you also get a warning when you do, for example:
303       val op = = Int32.+
304    which was not the case up to now.
305    
306    I placed "assign" and "deref" into the _Core structure so that the
307    code that deals with the "lazy" keyword can find them there.  This
308    removes the need for having access to the primitive environment
309    during elaboration.
310    
311    ----------------------------------------------------------------------
312    Name: Matthias Blume
313    Date: 2001/08/13
314    Tag: blume-20010813-closures
315    Description:
316    
317    This fix was sent to us by Zhong Shao.  It is supposed to improve the
318    performance of certain loops by avoiding needless closure allocation.
319    
320    ----------------------------------------------------------------------
321    Name: Lal George
322    Date: 2001/07/31 10:03:23 EDT 2001
323    Tag: george-20010731-x86-fmalloc
324    Description: Fixed bug in x86 calls
325    
326        There was a bug where call instructions would mysteriously
327        vanish. The call instruction had to be one that returned
328        a floating point value.
329    
330    ----------------------------------------------------------------------
331    Name: Lal George
332    Date: 2001/07/19 16:36:29 EDT 2001
333    Tag: george-20010719-simple-cells
334    Description:
335    
336    I have dramatically simplified the interface for CELLS in MLRISC.
337    
338    In summary, the cells interface is broken up into three parts:
339    
340      1. CellsBasis : CELLS_BASIS
341    
342            CellsBasis is a top level structure and common for all
343            architectures.  it contains the definitions of basic datatypes
344            and utility  functions over these types.
345    
346      2. functor Cells() : CELLS
347    
348            Cells generates an interface for CELLS that incorporates the
349            specific resources on the target architecture, such as the
350            presence of special register classes, their number and size,
351            and various useful substructures.
352    
353      3. <ARCH>CELLS
354    
355            e.g. SparcCells: SPARCCELLS
356    
357            <ARCH>CELLS usually contains additional bindings for special
358            registers  on the architecture, such as:
359    
360                    val r0 : cell           (* register zero *)
361                    val y : cell            (* Y register *)
362                    val psr : cell          (* processor status register *)
363                    ...
364    
365            The structure returned by applying the Cells functor is opened
366            in this interface.
367    
368    The main implication of all this is that the datatypes for cells is
369    split between CellsBasis and CELLS -- a fairly simple change for user
370    code.
371    
372    In the old scheme the CELLS interface had a definitional binding of
373    the form:
374    
375            signature CELLS = sig
376    
377               structure CellsBasis = CellsBasis
378    
379               ...
380    
381            end
382    
383    With all the sharing constraints that goes on in MLRISC, this old
384    design  quickly leads to errors such as:
385    
386            "structure definition spec inside of sharing ... "
387    
388    
389    and appears to require an unacceptable amount of sharing and where
390    constraint hackery.
391    
392    I think this error message (the interaction of definitional specs and
393    sharing) requires more explanation on our web page.
394    
395    ----------------------------------------------------------------------
396    Name: Matthias Blume
397    Date: 2001/07/19 15:00:00 EDT
398    Tag: blume-20010719-libreorg
399    Description:
400    
401    This update puts together a fairly extensive but straightforward change
402    to the way the libraries that implement the interactive system are
403    organized:
404    
405       The biggest change is the elimination of structure Compiler.  As a
406       replacement for this structure, there is now a CM library
407       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
408       that exports all the substructures of the original structure Compiler
409       directly.  So instead of saying Compiler.Foo.bar one now simply
410       says Foo.bar.  (The CM libraries actually export a collection of
411       structures that is richer than the collection of substructures of
412       structure Compiler.)
413    
414       To make the transition smooth, there is a separate library called
415       $smlnj/compiler/compiler.cm which puts together and exports the
416       original structure Compiler (or at least something very close to it).
417    
418       There are five members of the original structure Compiler
419       that are not exported directly but which instead became members
420       of a new structure Backend (described by signature BACKEND).  These are:
421       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
422       Interact (: INTERACT), structure Machine (: MACHINE), and val
423       architecture (: string).
424    
425       Structure Compiler.Version has become structure CompilerVersion.
426    
427       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
428       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
429       or x86, respectively.
430       Each of these exports the same frontend structures that
431       $smlnj/compiler.cm exports.  But they do not have a structure Backend
432       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
433       Hppa, PPC, Sparc, or X86, respectively.
434    
435       Library $smlnj/compiler/all.cm exports the union of the exports of
436       $smlnj/compiler/<arch>.cm
437    
438       There are no structures <Arch>Compiler anymore, use
439       $smlnj/compiler/<arch>.cm instead.
440    
441       Library host-compiler-0.cm is gone.  Instead, the internal library
442       that instantiates CM is now called cm0.cm.  Selection of the host
443       compiler (backend) is no longer done here but. (Responsibility for it
444       now lies with $smlnj/compiler/current.cm.  This seems to be more
445       logical.)
446    
447       Many individual files have been moved or renamed.  Some files have
448       been split into multiple files, and some "dead" files have been deleted.
449    
450    Aside from these changes to library organization, there are also changes
451    to the way the code itself is organized:
452    
453       Structure Binfile has been re-implemented in such a way that it no
454       longer needs any knowledge of the compiler.  It exclusively deals
455       with the details of binfile layout.  It no longer invokes the
456       compiler (for the purpose of creating new prospective binfile
457       content), and it no longer has any knowledge of how to interpret
458       pickles.
459    
460       Structure Compile (: COMPILE) has been stripped down to the bare
461       essentials of compilation.  It no longer deals with linking/execution.
462       The interface has been cleaned up considerably.
463    
464       Utility routines for dealing with linking and execution have been
465       moved into their own substructures.
466    
467       (The ultimate goal of these changes is to provide a light-weight
468       binfile loader/linker (at least for, e.g., stable libraries) that
469       does not require CM or the compiler to be present.)
470    
471    CM documentation has been updated to reflect the changes to library
472    organization.
473    
474    ----------------------------------------------------------------------
475    Name: Matthias Blume
476    Date: 2001/07/10 17:30:00 EDT
477    Tag: Release_110_34
478    Description:
479    
480    Minor tweak to 110.34 (re-tagged):
481    
482      - README.html file added to CVS repository
483      - runtime compiles properly under FreeBSD 3.X and 4.X
484    
485    ----------------------------------------------------------------------
486    Name: Matthias Blume
487    Date: 2001/07/10 17:30:00 EDT
488    Tag: Release_110_34
489    Description:
490    
491    New version number (110.34). New bootfiles.
492    
493    ----------------------------------------------------------------------
494    Name: Matthias Blume
495    Date: 2001/07/09 16:00:00 EDT
496    Tag: blume-20010709-more-varargs
497    Description:
498    
499    I changed the handling of varargs in ml-nlffigen again:
500    The ellipsis ... will now simply be ignored (with an accompanying warning).
501    
502    The immediate effect is that you can actually call a varargs function
503    from ML -- but you can't actually supply any arguments beyond the ones
504    specified explicitly.  (For example, you can call printf with its format
505    string, but you cannot pass additional arguments.)
506    
507    This behavior is only marginally more useful than the one before, but
508    it has the advantage that a function or, more importantly, a function
509    type never gets dropped on the floor, thus avoiding follow-up problems with
510    other types that refer to the offending one.
511    
512    ----------------------------------------------------------------------
513    Name: Matthias Blume
514    Date: 2001/07/09 11:25:00 EDT
515    Tag: blume-20010709-varargs
516    Description:
517    
518    1. ckit-lib.cm now exports structure Error
519    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
520       with a warning accompanied by a source location.  Moreover, it
521       merely skips the offending function or type and proceeds with the
522       rest of its work.u  As a result, one can safely feed C code containing
523       "..." to ml-nlffigen.
524    3. There are some internal improvements to CM, providing slightly
525       more general string substitutions in the tools subsystem.
526    
527    ----------------------------------------------------------------------
528    Name: Matthias Blume
529    Date: 2001/06/27 15:10:00 EDT
530    Tag: blume-20010627-concur
531    Description:
532    
533    Fixed a small bug in CM's handling of parallel compilation.
534    (You could observe the bug by Control-C-interrupting an ordinary
535    CMB.make or CM.stabilize and then attaching some compile servers.
536    The result was that all of a sudden the previously interrupted
537    compilation would continue on its own.  This was because of
538    an over-optimization: CM did not bother to clean out certain queues
539    when no servers were attached "anyway", resulting in the contents
540    of these queues to grab control when new servers did get attached.)
541    
542    There is also another minor update to the CM manual.
543    
544    ----------------------------------------------------------------------
545    Name: Matthias Blume
546    Date: 2001/06/26 16:15:00 EDT
547    Tag: blume-20010626-cmdoc
548    Description:
549    
550    Minor typo fixed in CM manual (syntax diagram for libraries).
551    
552    ----------------------------------------------------------------------
553    Name: Matthias Blume
554    Date: 2001/06/25 22:55:00 EDT
555    Tag: blume-20010625-x86pc
556    Description:
557    
558    Fixed a nasty bug in the X86 assembly code that caused signal
559    handlers to fail (crash) randomly.
560    
561    ----------------------------------------------------------------------
562    Name: Matthias Blume
563    Date: 2001/06/25 12:05:00 EDT
564    Tag: blume-20010625-nlffigen
565    Description:
566    
567    This update fixes a number of minor bugs in ml-nlffigen as reported by
568    Nick Carter <nbc@andrew.cmu.edu>.
569    
570      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
571      2. Default names for generated files are now derived from the name of
572         the C file *without its directory*.  In particular, this causes generated
573         files to be placed locally even if the C file is in some system directory.
574      3. Default names for generated signatures and structures are also derived
575         from the C file name without its directory.  This avoids silly things
576         like "structure GL/GL".
577         (Other silly names are still possible because ml-nlffigen does not do
578          a thorough check of whether generated names are legal ML identifiers.
579          When in doubt, use command line arguments to force particular names.)
580    
581    ----------------------------------------------------------------------
582    Name: Matthias Blume
583    Date: 2001/06/21 12:25:00 EDT
584    Tag: blume-20010621-eXene
585    Description:
586    
587    eXene now compiles and (sort of) works again.
588    
589    The library name (for version > 110.33) is $/eXene.cm.
590    
591    I also added an new example in src/eXene/examples/nbody.  See the
592    README file there for details.
593    
594    ----------------------------------------------------------------------
595    Name: Matthias Blume
596    Date: 2001/06/20 16:40:00 EDT
597    Tag: blume-20010620-cml
598    Description:
599    
600    CML now compiles and works again.
601    
602    Libraries (for version > 110.33):
603    
604      $cml/cml.cm            Main CML library.
605      $cml/basis.cm          CML's version of $/basis.cm.
606      $cml/cml-internal.cm   Internal helper library.
607      $cml/core-cml.cm       Internal helper library.
608      $cml-lib/trace-cml.cm  Tracing facility.
609      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
610    
611    The installer (config/install.sh) has been taught how to properly
612    install this stuff.
613    
614    ----------------------------------------------------------------------
615    Name: Matthias Blume
616    Date: 2001/06/19 17:55:00 EDT
617    Tag: blume-20010619-instantiate
618    Description:
619    
620    This un-breaks the fix for bug 1432.
621    (The bug was originally fixed in 110.9 but I broke it again some
622    time after that.)
623    
624    ----------------------------------------------------------------------
625    Name: Matthias Blume
626    Date: 2001/06/19 17:25:00 EDT
627    Tag: blume-20010619-signals
628    Description:
629    
630    This should (hopefully) fix the long-standing signal handling bug.
631    (The runtime system was constructing a continuation record with an
632    incorrect descriptor which would cause the GC to drop data on the floor...)
633    
634    ----------------------------------------------------------------------
635    Name: Matthias Blume
636    Date: 2001/06/15 15:05:00 EDT
637    Tag: blume-20010615-moresparc
638    Description:
639    
640    Here is a short late-hour update related to Sparc c-calls:
641    
642     -- made handling of double-word arguments a bit smarter
643    
644     -- instruction selection phase tries to collapse certain clumsily
645        constructed ML-Trees; typical example:
646    
647            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
648    
649        This currently has no further impact on SML/NJ since mlriscGen does
650        not seem to generate such patterns in the first place, and c-calls
651        (which did generate them in the beginning) has meanwhile been fixed
652        so as to avoid them as well.
653    
654    ----------------------------------------------------------------------
655    Name: Matthias Blume
656    Date: 2001/06/15 15:05:00 EDT
657    Tag: blume-20010615-sparc
658    Description:
659    
660    The purpose of this update is to provide an implementation of NLFFI
661    on Sparc machines.
662    
663    Here are the changes in detail:
664    
665       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
666       the Sparc implementation of the c-calls API.
667       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
668       for accessing the ML frame.  Thus, we have a real frame pointer and
669       can freely modify %sp without need for an omit-frame-ptr phase.
670       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
671       this case.
672       * ml-nlffigen has been taught to produce code for different architectures
673       and calling conventions.
674       * In a way similar to what was done in the x86 case, the Sparc
675       backend uses its own specific extension to mltree.  (For example,
676       it needs to be able to generate UNIMP instructions which are part
677       of the calling convention.)
678       * ml-nlffi-lib was reorganized to make it more modular (in particular,
679       to make it easier to plug in new machine- and os-dependent parts).
680    
681    There are some other fairly unrelated bug fixes and cleanups as well:
682    
683       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
684       that they properly share their libraries with existing SML/NJ libraries.
685       * I fixed a minor cosmetic bug in CM, supressing certain spurious
686       follow-up error messages.
687       * Updates to CM/CMB documentation.
688    
689    TODO items:
690    
691       * MLRISC should use a different register as its asmTemp on the Sparc.
692         (The current %o2 is a really bad choice because it is part of the
693         calling conventions, so things might interfere in unexpected ways.)
694    
695    ----------------------------------------------------------------------
696    Name: Matthias Blume
697    Date: 2001/06/07
698    Tag: blume-20010607-calls
699    Description:
700    
701    A number of internal changes related to C calls and calling conventions:
702    
703    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
704       number of bytes popped implicitly (by the callee).  In most cases
705       this field is 0 but on x86/win32 it is some non-zero value.  This
706       is information provided for the benefit of the "omit-frameptr" pass.
707    2. The CALL instruction on the x86 carries a similar "pops" field.
708       The instruction selection phase copies its value from the ML-Tree
709       CALL statement.
710    3. On all other architectures, the instruction selection phase checks
711       whether "pops=0" and complains if not.
712    4. The c-calls implementation for x86 now accepts two calling conventions:
713       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
714       up after the call and pops is set to 0.  For "stdcall", the caller
715       does nothing, leaving the cleanup to the callee; pops is set to
716       the number of bytes that were pushed onto the stack.
717    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
718       distinguish between "ccall" and "stdcall".
719    6. The UNIMP instruction has been added to the supported Sparc instruction
720       set. (This is needed for implementing the official C calling convention
721       on this architecture.)
722    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
723       work with the latest CM.
724    
725    ----------------------------------------------------------------------
726    Name: Matthias Blume
727    Date: 2001/06/05 15:10:00 EDT
728    Tag: blume-20010605-cm-index
729    Description:
730    
731    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
732    
733    1. CM can now generate "index files".  These are human-readable files
734       that list on a per-.cm-file basis each toplevel symbol defined or
735       imported.  The location of the index file for
736       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
737       To enable index-file generation, set CM.Control.generate_index to true
738       or export an environment-symbol: export CM_GENERATE_INDEX=true.
739    
740       The CM manual has been updated accordingly.
741    
742    2. I made some slight modifications to the c-calls API in MLRISC.
743    
744         a) There is now a callback to support saving/restoring of
745            dedicated but caller-save registers around the actual call
746            instruction.
747         b) One can optionally specify a comment-annotation for the
748            call instruction.
749    
750    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
751       (For example, the comment annotation shows the C prototype of
752        the function being called.)
753    
754    ----------------------------------------------------------------------
755    Name: Matthias Blume
756    Date: 2001/06/01 13:30:00 EDT
757    Tag: blume-20010601-nlffi-cleanup
758    Description:
759    
760    This is mostly a cleanup of MLFFI stuff:
761    
762       - some signature files have been put into a more exposed place
763       - the ugly 'f type parameter is gone (simplifies types tremendously!)
764       - ml-nlffigen changed accordingly
765       - tutorial updated
766    
767    Other changes:
768    
769       - author's affiliation in CM manual(s) updated
770       - some more recognized keywords added to Allen's sml.sty
771    
772    ----------------------------------------------------------------------
773    Name: Matthias Blume
774    Date: 2001/05/25 15:30:00 EDT
775    Tag: blume-20010525-iptr
776    Description:
777    
778      - put the official 110.33-README (as it appears on the ftp server) under
779        CVS
780      - fixed a small bug related to incomplete pointer types in
781        ml-nlffigen
782      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
783        (it does not need the 'f type parameter)
784    
785    ----------------------------------------------------------------------
786    Name: Matthias Blume
787    Date: 2001/05/23 14:30:00 EDT
788    Tag: Release_110_33
789    Description:
790    
791    New version number (110.33).  New bootfiles.
792    
793    ----------------------------------------------------------------------
794    Name: Matthias Blume
795    Date: 2001/05/22 18:06:00 EDT
796    Tag: blume-20010522-targets
797    Description:
798    
799    Made install.sh use file config/targets.customized if it exists, falling
800    back to config/targets if it doesn't.  This way one can have a customized
801    version of the targets file without touching the "real thing", thus
802    eliminating the constant fear of accidentally checking something bogus
803    back into the CVS repository...  (File config/targets.customized must
804    not be added to the repository!)
805    
806    ----------------------------------------------------------------------
807    Name: Matthias Blume
808    Date: 2001/05/22 16:30:00 EDT
809    Tag: blume-20010522-minitut
810    Description:
811    
812    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
813       struct returns.
814    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
815       incomplete, preliminary documentation for NLFFI.
816    
817    ----------------------------------------------------------------------
818    Name: Matthias Blume
819    Date: 2001/05/14 11:30:00 EDT
820    Tag: blume-20010514-script
821    Description:
822    
823    Some bugs in install script fixed.
824    
825    In addition to that I also made a slight change to the NLFFI API:
826    Functors generated by ml-nlffigen now take the dynamic library as a
827    straight functor argument, not as a suspended one.  (The original
828    functor code used to force the suspension right away anyway, so there
829    was nothing gained by this complication of the interface.)
830    
831    ----------------------------------------------------------------------
832    Name: Matthias Blume
833    Date: 2001/05/11 14:35:00 EDT
834    Tag: blume-20010511-ml-nlffi
835    Description:
836    
837    I finally took the plunge and added my new FFI code to the main
838    repository.  For x86-linux it is now ready for prime-time.
839    
840    There are two new subdirectories of "src":
841    
842      - ml-nlffi-lib:
843           The utility library for programs using the FFI interface.
844           Here is the implementation of $/c.cm and its associated low-level
845           partners $/c-int.cm and $/memory.cm.
846      - ml-nlffigen:
847           A stand-alone program for generating ML glue code from C source
848           code.
849    
850    Building ml-nlffigen requires $/ckit-lib.cm.
851    
852    The config/install.sh script has been updates to do the Right Thing
853    (hopefully).
854    
855    Notice that the source tree for the C-Kit will not be put under "src"
856    but directly under the installation root directory.  (This is the
857    structure that currently exists on the CVS server when you check out
858    module "sml".)  Fortunately, config/install.sh knows about this oddity.
859    
860    Bugs: No documentation yet.
861    
862    ----------------------------------------------------------------------
863    Name: Matthias Blume
864    Date: 2001/05/09 16:35:00 EDT
865    Tag: blume-20010509-cpscontract
866    Description:
867    
868    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
869    wrapper/unwrapper elimination did not decrement usage counts and some
870    dead variables got overlooked by the dead-up logic.)
871    
872    ----------------------------------------------------------------------
873    Name: Lal George
874    Date: 2001/05/08  17:26:09 EDT
875    Tag: george-20010508-omit-frameptr
876    Description:
877    
878    Changes to implement the omit-frame-pointer optimization to support
879    raw C calls. For now, there is only support on the Intel x86, but
880    other architectures will follow as more experience is gained with this.
881    
882    
883    ----------------------------------------------------------------------
884    Name: Matthias Blume
885    Date: 2001/05/07 14:40:00 EDT
886    Tag: blume-20010507-proxies
887    Description:
888    
889    I made into "proxy libraries" all libraries that qualify for such a
890    change.  (A qualifying library is a library that has another library or
891    groups as its sole member and repeats that member's export list
892    verbatim.  A proxy library avoids this repetition by omitting its export
893    list, effectively inheriting the list that its (only) member exports.
894    See the CM manual for more explanation.)
895    The main effect is that explicit export lists for these libraries
896    do not have to be kepts in sync, making maintenance a bit easier.
897    
898    I also added copyright notices to many .cm-files.
899    
900    Last but not least, I made a new set of bootfiles.
901    
902    ----------------------------------------------------------------------
903    Name: Matthias Blume
904    Date: 2001/05/04 17:00:00 EDT
905    Tag: blume-20010504-cm-lsplit
906    Description:
907    
908    0. John merged pending changes to $/smlnj-lib.cm
909    
910    1. Allen's previous change accidentally backed out of one of Lal's
911       earlier changes.  I undid this mistake (re-introducing Lal's change).
912    
913    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
914       within the compiler where applicable.  There is some code simplification
915       because of that.
916    
917    3. The "split" phase (in FLINT) is now part of the default list of phases.
918       Compiler.Control.LambdaSplitting.* can be used to globally control the
919       lambda-splitting (cross-module-inlining) engine.  In addition to that,
920       it can now also be controlled on a per-source basis: CM has been taught
921       a new tool parameter applicable to ML source files.
922    
923       - To turn lambda-splitting off completely:
924            local open Compiler.Control.LambdaSplitting in
925                val _ = set Off
926            end
927       - To make "no lambda-splitting" the global default (but allow per-source
928         overriding); this is the initial setting:
929            local open Compiler.Control.LambdaSplitting in
930                val _ = set (Default NONE)
931            end
932       - To make "lambda-splitting with aggressiveness a" the global default
933         (and allow per-source overriding):
934            local open Compiler.Control.LambdaSplitting in
935                val _ = set (Default (SOME a))
936            end
937    
938       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
939         write (in the respective .cm-file):
940            a.sml (lambdasplitting:off)
941       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
942            a.sml (lambdasplitting:on)
943       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
944         <a> is a decimal non-negative integer):
945            a.sml (lambdasplitting:<a>)
946       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
947            a.sml (lambdasplitting:infinity)
948       - To use the global default for a.sml:
949            a.sml (lambdasplitting:default)
950         or simply
951            a.sml
952    
953    ----------------------------------------------------------------------
954    Name: Allen Leung
955    Date: 2001/05/04 01:57:00 EDT
956    Tag: leunga-20010504-sync
957    Description:
958    
959      MLRISC features.
960    
961      1. Fix to CMPXCHG instructions.
962      2. Changed RA interface to allow annotations in callbacks.
963      3. Added a new method to the stream interface to allow annotations updates.
964    
965    ----------------------------------------------------------------------
966    Name: Matthias Blume
967    Date: 2001/05/01 11:45:00 EDT
968    Tag: blume-20010501-pcedittmp
969    Description:
970    
971    Changed install.sh to use the current working directory instead of
972    /usr/tmp for a temporary file (pcedittmp).  The previous choice
973    of /usr/tmp caused trouble with MacOS X because of file premission
974    problems.
975    
976    ----------------------------------------------------------------------
977    Name: Matthias Blume
978    Date: 2001/04/20 11:10:00 EDT
979    Tag: blume-20010420-inMLflag
980    Description:
981    
982     - added vp_limitPtrMask to vproc-state.h
983       (for use by the raw-C-calls mechanism to implement proper interrupt
984        handling)
985     - made the ML compiler aware of various data-structure offsets so it
986       can generate code for accessing the vp_inML flag and vp_limitPtrMask
987     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
988       raw C-calls
989    
990    ----------------------------------------------------------------------
991    Name: Lal George
992    Date: 2001/04/20 09:15:28 EDT
993    Tag: george-20010420-macosX
994    Description:
995    
996     - Changes to port to Mac OS X; Darwin.
997    
998     - In the process I found that sqrt was broken on the PPC, because the
999       fsqrt instruction is not implemented.
1000    
1001    ----------------------------------------------------------------------
1002    Name: Matthias Blume
1003    Date: 2001/04/18 12:45:00 EDT
1004    Tag: blume-20010418-ccalls
1005    Description:
1006    
1007     - fixed two off-by-4 errors in the x86-specific c-calls implementation
1008       (this bug prevented structure arguments containing pointers from being
1009        passed correctly)
1010     - changed the raw-C-call code in mlriscGen.sml in such a way that
1011       structure arguments are represented as a pointer to the beginning
1012       of the structure (instead of having a series of synthesized arguments,
1013       one for each structure member)
1014    
1015     - made makeml script's verbosity level configurable via environment
1016       variable (MAKEML_VERBOSITY)
1017    
1018     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
1019       in rawmem-x86.sml; we are now using the real thing
1020    
1021    ----------------------------------------------------------------------
1022    Name: Matthias Blume
1023    Date: 2001/03/22 16:25:00 EST
1024    Tag: blume-20010322-bootfiles
1025    Description:
1026    
1027    Created a new set of bootfiles (for your automatic installation convenience).
1028    
1029    ----------------------------------------------------------------------
1030    Name: Matthias Blume
1031    Date: 2001/03/22 15:10:00 EST
1032    Tag: blume-20010322-rawmem-parcm
1033    Description:
1034    
1035    1. All "raw memory access" primitives for the new FFI are implemented now
1036       (at least on the x86).
1037    2. Some further cleanup of CM's parallel make mechanism.
1038    
1039    ----------------------------------------------------------------------
1040    Name: Matthias Blume
1041    Date: 2001/03/19 17:53:00 EST
1042    Tag: blume-20010319-parallel
1043    Description:
1044    
1045    Parallel make (using compile servers) now works again.
1046    
1047    To this end, CM.stabilize and CMB.make have been modified to work in
1048    two passes when compile servers are attached:
1049       1. Compile everything, do not perform stabilization; this pass
1050          uses compile servers
1051       2. Stabilize everything; this pass does not use compile servers
1052    If there are no compile servers, the two passes are combined into one
1053    (as before).  Splitting the passes increases the inherent parallelism
1054    in the dependency graph because the entire graph including all
1055    libraries is available at the same time.  This, in turn, improves
1056    server utilization.  The downside is that the master process will
1057    have to do some extra work after compilation is done (because for
1058    technical reasons it must re-read all the binfiles during stabilization).
1059    
1060    ----------------------------------------------------------------------
1061    Name: Matthias Blume
1062    Date: 2001/03/16 12:22:00 EST
1063    Tag: blume-20010316-bootfiles
1064    Description:
1065    
1066    Created a new set of bootfiles (for your automatic installation convenience).
1067    
1068    ----------------------------------------------------------------------
1069    Name: Matthias Blume
1070    Date: 2001/03/16 11:00:00 EST
1071    Tag: blume-20010316-MLTREE-fixup
1072    Description:
1073    
1074    This is a minor fixup for an (untagged) earlier commit by Allen.
1075    (A file was missing).
1076    
1077    ----------------------------------------------------------------------
1078    Name: Allen Leung
1079    Date: Mon Mar  5 18:54:57 EST 2001
1080    Tag: leunga-20010305-cut-support
1081    
1082    1. New support for alternative control-flow in MLTREE.
1083       Currently we support
1084    
1085          FLOW_TO(CALL ...., [k1,...,kn])
1086    
1087       This is needed for 'cuts to' in C-- and try/handle-like constructs
1088       in Moby
1089    
1090       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
1091    
1092    2. Register Allocator
1093    
1094       Changes in interface [from Fermin, John]
1095    
1096    3. Alpha 8-bit SLL support [Fermin]
1097    
1098    4. All architectures
1099    
1100       A new module (ClusterExpandCopies) for expanding parallel copies.
1101    
1102    ----------------------------------------------------------------------
1103    Name: Allen Leung
1104    Date: 2001/02/27 23:07:00 EST
1105    Tag: leunga-20010227-minor-stuff
1106    
1107    1. Alpha bug fix for CMOVNE
1108    2. Handle mltree COND(..,FCMP ...,...)
1109    3. Bug fix in simplifier
1110    
1111    ----------------------------------------------------------------------
1112    Name: Matthias Blume
1113    Date: 2001/01/30 17:50:00 EST
1114    Tag: blume-20010130-sync
1115    Description:
1116    
1117    This is just a minor update to sync my devel branch with the main brach.
1118    The only visible change is the addition of some README files.
1119    
1120    ----------------------------------------------------------------------
1121    Name: Matthias Blume
1122    Date: 2001/01/12 23:30:00 JST
1123    Tag: blume-20010112-bootfiles
1124    Description:
1125    
1126    Made a new set of bootfiles that goes with the current state of the
1127    repository.
1128    
1129    ----------------------------------------------------------------------
1130    Name: Matthias Blume
1131    Date: 2001/01/12 21:20:00 JST
1132    Tag: blume-20010112-sync
1133    Description:
1134    
1135    I am just flushing out some minor changes that had accumulated in
1136    my private branch in order to sync with the main tree.  (This is
1137    mainly because I had CVS trouble when trying to merge _into_ my
1138    private branch.)
1139    
1140    Most people should be completely unaffected by this.
1141    
1142    ----------------------------------------------------------------------
1143    Name: Allen Leung
1144    Date: Thu Jan 11 21:03:00 EST 2001
1145    Tag: leunga-20010111-labexp=mltree
1146    Description:
1147    
1148    1.  Removed the type LabelExp and replace it by MLTree.
1149    2.  Rewritten mltree-simplify with the pattern matcher tool.
1150    3.  There were some bugs in alpha code generator which would break
1151        64-bit code generation.
1152    4.  Redo the tools to generate code with the
1153    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
1154        are now generated by perl scripts.
1155    
1156    ----------------------------------------------------------------------
1157    Name: Matthias Blume
1158    Date: 2001/01/10 21:55:00 JST
1159    Tag: blume-20010110-rcc
1160    Description:
1161    
1162    The RCC stuff now seems to work (but only on the x86).
1163    This required hacking of the c-calls interface (and -implementation) in
1164    MLRISC.
1165    
1166    Normal compiler users should be unaffected.
1167    
1168    ----------------------------------------------------------------------
1169    Name: Matthias Blume
1170    Date: 2001/01/09 01:20:00 JST
1171    Tag: blume-20010109-rcc
1172    Description:
1173    
1174    This is a fairly big patch, flushing out a large number of pending
1175    changes that I made to my development copy over the last couple of days.
1176    
1177    Of practical relevance at this moment is a workaround for a pickling
1178    bug that Allen ran into the other day.  The cause of the bug itself is
1179    still unknown and it might be hard to fix it properly, but the
1180    workaround has some merits of its own (namely somewhat reducing pickling
1181    overhead for certain libraries).  Therefore, I think this solution should
1182    be satisfactory at this time.
1183    
1184    The rest of the changes (i.e., the vast majority) has to do with my
1185    ongoing efforts of providing direct support for C function calls from
1186    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
1187    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
1188    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
1189    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
1190    handle RCC.
1191    
1192    The last part (the changes to mlriscGen.sml) are still known to be
1193    wrong on the x86 and not implemented on all other architectures.  But
1194    the infrastructure is in place. I had to change a few functor
1195    signatures in the backend to be able to route the CCalls interface
1196    from MLRISC there, and I had to specialize the mltree type (on the
1197    x86) to include the necessary extensions. (The extensions themselves
1198    were already there and redy to go in MLRISC/x86).
1199    
1200    Everything should be very happy as soon as someone helps me with
1201    mlriscGen.sml...
1202    
1203    In any case, nothing of this should matter to anyone as long as the
1204    new primop is not being used (which is going to be the case unless you
1205    find it where I hid it :). The rest of the compiler is completely
1206    unaffected.
1207    
1208    ----------------------------------------------------------------------
1209    Name: Matthias Blume
1210    Date: 2001/01/05 00:30:00 JST
1211    Tag: blume-20010105-primops
1212    Description:
1213    
1214    Added some experimental support for work that I am doing right now.
1215    These changes mostly concern added primops, but there is also a new
1216    experimental C library in the runtime system (but currently not enabled
1217    anywhere except on Linux/X86).
1218    
1219    In the course of adding primops (and playing with them), I discovered that
1220    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
1221    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
1222    To recover, I made type information in INL_PRIM mandatory and changed
1223    prim.sml as well as built-in.sml accordingly.  The InLine structure now
1224    has complete, correct type information (i.e., no bottom types).
1225    
1226    Since all these changes mean that we need new binfiles, I also bumped the
1227    version number to 110.32.1.
1228    
1229    ----------------------------------------------------------------------
1230    Name: Matthias Blume
1231    Date: 2000/12/30 22:10:00 JST
1232    Tag: blume-20001230-various
1233    Description:
1234    
1235    Added proxy libraries for MLRISC and let MLRISC libraries refer
1236    to each other using path anchors.  (See CM manual for explanation.)
1237    
1238    Updated CM documentation.
1239    
1240    Fixed some bugs in CM.
1241    
1242    Implemented "proxy" libraries (= syntactic sugar for CM).
1243    
1244    Added "-quiet" option to makeml and changed runtime system accordingly.
1245    
1246    Added cleanup handler for exportML to reset timers and compiler stats.
1247    
1248    ----------------------------------------------------------------------
1249    Name: Lal George
1250    Date: 2000/12/22 22:22:58 EST 2000
1251    Tag: Release_110_32
1252    Description:
1253    
1254            Infinite precision used throughout MLRISC.
1255            see MLRISC/mltree/machine-int.sig
1256    
1257    ----------------------------------------------------------------------
1258    Name: Matthias Blume
1259    Date: 2000/12/22 23:16:00 JST
1260    Tag: blume-20001222-warn
1261    Description:
1262    
1263    Corrected wording and formatting of some CM warning message which I
1264    broke in my previous patch.
1265    
1266    ----------------------------------------------------------------------
1267    Name: Matthias Blume
1268    Date: 2000/12/22 21:20:00 JST
1269    Tag: blume-20001222-anchorenv
1270    Description:
1271    
1272    Fixed CM's handling of anchor environments in connection with CMB.make.
1273    
1274    ----------------------------------------------------------------------
1275    Name: Matthias Blume
1276    Date: 2000/12/22 13:15:00 JST
1277    Tag: blume-20001222-cleanup
1278    Description:
1279    
1280    Removed src/cm/ffi which does not (and did not) belong here.
1281    
1282    ----------------------------------------------------------------------
1283    Name: Matthias Blume
1284    Date: 2000/12/21 23:55:00 JST
1285    Tag: blume-20001221-exn
1286    Description:
1287    
1288    Probably most important: CM no longer silently swallows all exceptions
1289    in the compiler.
1290    Plus: some other minor CM changes.  For example, CM now reports some
1291    sizes for generated binfiles (code, data, envpickle, lambdapickle).
1292    
1293    ----------------------------------------------------------------------
1294    Name: Matthias Blume
1295    Date: 2000/12/15 00:01:05 JST
1296    Tag: blume-20001215-dirtool
1297    Description:
1298    
1299    - "dir" tool added.
1300    - improvements and cleanup to Tools structure
1301    - documentation updates
1302    
1303    ----------------------------------------------------------------------
1304    Name: Allen Leung
1305    Date: Thu Dec 14 03:45:24 EST 2000
1306    Description:
1307    Tag:  leunga-20001214-int-inf
1308    Description:
1309    
1310       In IntInf, added these standard functions, which are missing from our
1311    implementation:
1312    
1313        andb : int * int -> int
1314        xorb : int * int -> int
1315        orb  : int * int -> int
1316        notb : int -> int
1317         <<   : int * word -> int
1318        ~>>  : int * word -> int
1319    
1320       Not tested, I hope they are correct.
1321    
1322    ----------------------------------------------------------------------
1323    Name: Allen Leung
1324    Date: Fri Dec  8 19:23:26 EST 2000
1325    Description:
1326    Tag:  leunga-20001208-nowhere
1327    Description:
1328    
1329      Slight improvements to the 'nowhere' tool to handle OR-patterns,
1330    to generate better error messages etc.  Plus a brief manual.
1331    
1332    ----------------------------------------------------------------------
1333    Name: Lal George
1334    Date: 2000/12/08 09:54:02 EST 2000
1335    Tag: Release_110_31
1336    Description:
1337    
1338    - Version 110.31
1339    ----------------------------------------------------------------------
1340    Name: Allen Leung
1341    Date: Thu Dec  7 22:01:04 EST 2000
1342    Tag:  leunga-20001207-cell-monster-hack
1343    Description:
1344    
1345    Major MLRISC internal changes.  Affect all clients.
1346    Summary:
1347    
1348    1.  Type CELLS.cell = int is now replaced by a datatype.
1349        As a result, the old regmap is now gone.  Almost all interfaces
1350        in MLRISC change as a consequence.
1351    
1352    2.  A new brand version of machine description tool (v3.0) that generates
1353        modules expecting the new interface.  The old version is removed.
1354    
1355    3.  The RA interface has been further abstracted into two new functors.
1356        RISC_RA and X86RA.  These functors have much simpler interfaces.
1357        [See also directory MLRISC/demo.]
1358    
1359    4.  Some other new source->source code generation tools are available:
1360    
1361        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
1362        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
1363           I use this tool to generate the peephole optimizers---with the new
1364           cell type changes, peephole rules are becoming difficult to write
1365           without conditional pattern matching.
1366    
1367    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
1368        cover the entire MLRISC source tree so many things broke.
1369    
1370    6.  CM files have been moved to the subdirectory MLRISC/cm.
1371        They are moved because there are a lot of them and they clutter up the
1372        root dir.
1373    
1374    7.  More detailed documentation to come...
1375    
1376        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
1377        a makeml -rebuild first.  This is because of other other
1378        changes that Matthias has made (see below).
1379    
1380    
1381    ----------------------------------------------------------------------
1382    Name: Matthias Blume
1383    Date: 2000/11/30 23:12:00 JST
1384    Tag: blume-20001130-filereorg
1385    Description:
1386    
1387    Some manual updates and some file reorganizations in CM.
1388    
1389    ----------------------------------------------------------------------
1390    Name: Matthias Blume
1391    Date: 2000/11/24 17:45:00 JST
1392    Tag: blume-20001124-link
1393    Description:
1394    
1395    Drastically improved link traversal code for the case that the dynamic
1396    value was already loaded at bootstrap time.  As a result, CM and CMB
1397    now both load blazingly fast -- even on a very slow machine.  Also,
1398    memory consumption has been further reduced by this.
1399    
1400    Warning: The format of the PIDMAP file has changed.  THerefore, to
1401    bootstrap you have to do this:
1402    
1403    1. Run CMB.make
1404    2. Make a symbolic link for the boot directory:
1405         ln -s sml.boot.ARCH-OS xxx
1406    3. "Rebuild" the boot directory:
1407         ./makeml -boot xxx -rebuild sml ; rm xxx
1408    4. Boot normally:
1409          ./makeml
1410    
1411    ----------------------------------------------------------------------
1412    Name: Matthias Blume
1413    Date: 2000/11/21 21:20:00 JST
1414    Tag: blume-20001121-tools
1415    Description:
1416    
1417    Continued hacking on autoloading problem -- with success this time.
1418    Also changed tool-plugin mechanism.  See new CM manual.
1419    
1420    ----------------------------------------------------------------------
1421    Name: Matthias Blume
1422    Date: 2000/11/19 14:30:00 JST
1423    Tag:  blume-20001119-autoload
1424    Description:
1425    
1426    Some hacking to make autoloading faster.  Success for CMB, no success
1427    so far for CM.  There is a reduced structure CM' that autoloads faster.
1428    (This is a temporary, non-documented hack to be eliminated again when
1429    the general problem is solved.)
1430    
1431    ----------------------------------------------------------------------
1432    Name: Matthias Blume
1433    Date: 2000/11/17 14:10:00 JST
1434    Tag: blume-20001117-pickle-lib
1435    Description:
1436    
1437    1. Eliminated comp-lib.cm
1438    2. Made pickle-lib.cm
1439    3. Eliminated all uses of intset.sml (from comp-lib.cm)
1440    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
1441       equivalent constructs from smlnj-lib.cm (INtHashTable).
1442    5. Point 4. also goes for those uses of intmap.* in MLRISC.
1443       Duplicated intmap modules thrown out.
1444    6. Hunted down all duplicated SCC code and replaced it with
1445       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
1446    7. Rewrote Feedback module.
1447    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
1448       should be thrown out and equivalent modules from smlnj-lib.cm
1449       should be used (IntRedBlackSet, IntListSet, ...).
1450    
1451    Confirmed that compiler compiles to fixpoint.
1452    
1453    ----------------------------------------------------------------------
1454    Name: Allen Leung
1455    Date: 2000/11/10 18:00:00
1456    Tag: leunga-20001110-new-x86-fp
1457    
1458    A new x86 floating point code generator has been added.
1459    By default this is turned off.  To turn this on, do:
1460    
1461        CM.autoload "$smlnj/compiler.cm";
1462        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
1463    
1464    Changes:
1465    
1466    1.  Changed FTAN to FPTAN so that the assembly output is correct.
1467    2.  Changed the extension callback for FTANGENT to generate:
1468    
1469              fptan
1470              fstp  %st(0)
1471        instead of
1472              fptan
1473              fstpl ftempmem
1474    
1475    3.  Numerous assembly fixes for x86.
1476    
1477    5.  Cleaned up the machine code output module x86/x86MC.sml and added
1478        support for a whole bunch of instructions and addressing modes:
1479    
1480          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
1481          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
1482          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
1483          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
1484          fxch %st(n)
1485          fld %st(n)
1486          fst %st(n)
1487          fst mem
1488          fstp %st(n)
1489          fucom %st(n)
1490          fucomp %st(n)
1491    
1492        All these are now generated when the fast fp mode is turned on.
1493    
1494    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
1495    
1496    ----------------------------------------------------------------------
1497    Name: Matthias Blume
1498    Date: 2000/11/09 11:20:00 JST
1499    Tag: blume-20001109-scc
1500    Description:
1501    
1502    Eliminated some code duplication:
1503    
1504    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
1505       (Otherwise the functor is useless.)
1506    2. Used GraphSCCFn where SCCUtilFun was used previously.
1507    3. Got rid of SCCUtilFun (in comp-lib.cm).
1508    
1509    ----------------------------------------------------------------------
1510    Name: Lal George
1511    Date: 2000/11/06 09:02:21 EST 2000
1512    Tag: Release_110_30
1513    Description:
1514    
1515    - Version 110.30
1516    ----------------------------------------------------------------------
1517    Name: Matthias Blume
1518    Date: 2000/11/04 14:45:00
1519    Tag: blume-20001104-mlbuild
1520    Description:
1521    
1522    - Made ml-build faster on startup.
1523    - Documentation fixes.
1524    
1525    ----------------------------------------------------------------------
1526    Name: Matthias Blume
1527    Date: 2000/11/02 17:00:00 JST
1528    Tag: blume-20001102-condcomp
1529    Description:
1530    
1531    - Small tweaks to pickler -- new BOOTFILES!
1532    - Version bumped to 110.29.2.
1533    - Added conditional compilation facility to init.cmi (see comment there).
1534    ----------------------------------------------------------------------
1535    Name: Allen Leung
1536    Date: 2000/10/23 19:31:00
1537    Tag: leunga-20001023-demo-ra
1538    
1539    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
1540    2. Test programs for the graph library updated
1541    3. Some new MLRISC demo programs added
1542    
1543    ----------------------------------------------------------------------
1544    Name: Matthias Blume
1545    Date: 2000/08/31 22:15:00 JST
1546    Tag: blume-20001017-errmsg
1547    Description:
1548    
1549    More error message grief: Where there used to be no messages, there
1550    now were some that had bogus error regions.  Fixed.
1551    
1552    ----------------------------------------------------------------------
1553    Name: Matthias Blume
1554    Date: 2000/08/31 17:30:00 JST
1555    Tag: blume-20001017-v110p29p1
1556    Description:
1557    
1558    I made a version 110.29.1 with new bootfiles.
1559    
1560    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
1561              CM documentation changes and a small bugfix in CM's error reporting.
1562    
1563    ----------------------------------------------------------------------
1564    Name: Lal George
1565    Date: 2000/09/27 14:42:35 EDT
1566    Tag: george-20000927-nodestatus
1567    Description:
1568    
1569    Changed the type of the nodestatus, so that:
1570    
1571            SPILLED(~1)             is now SPILLED
1572            SPILLED(m) where m>=0   is now MEMREG(m)
1573            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
1574    
1575    ----------------------------------------------------------------------
1576    Name: Matthias Blume
1577    Date: 2000/09/07 14:45:00 JST
1578    Tag: blume-20000907-cmerrmsg
1579    Description:
1580    
1581    Small tweak to CM to avoid getting ML syntax error messages twice.
1582    
1583    ----------------------------------------------------------------------
1584    Name: Matthias Blume
1585    Date: 2000/08/31 18:00:00 JST
1586    Tag: blume-20000831-cvsbootfiles
1587    Description:
1588    
1589    New URL for boot files (because the 110.29 files on the BL server do
1590    now work correctly with my updated install scripts for yacc and lex).
1591    
1592    ----------------------------------------------------------------------
1593    Name: Matthias Blume
1594    Date: 2000/08/08 12:33:00 JST
1595    Tag: blume-20000808-manual
1596    Description:
1597    
1598    Tiny update to CM manual.
1599    
1600    ----------------------------------------------------------------------
1601    Name: Allen Leung
1602    Date: 2000/08/7 19:31:00
1603    Tag: leunga-20000807-a-whole-bunch-of-stuff
1604    
1605      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
1606    and Mobdy.
1607    
1608    1.  x86
1609    
1610       a.  Fixes to peephole module by John and Dan.
1611       b.  Assembly fix to SETcc by Allen.
1612       c.  Fix to c-call by John.
1613       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
1614       e.  Instruction selection optimization to SETcc as suggested by John.
1615    
1616           For example,
1617    
1618            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
1619    
1620           should generate:
1621    
1622            MOVL a, x
1623            SUBL b, x
1624            SHRL 31, x
1625    
1626    2.  IR stuff
1627    
1628         A bunch of new DJ-graph related algorithms added.  These
1629         speed up SSA construction.
1630    
1631    3.  SSA + Scheduling
1632    
1633         Added code for SSA and scheduling to the repository
1634    
1635    ----------------------------------------------------------------------
1636    Name: Lal George
1637    Date: 2000/07/27 11:53:14 EDT
1638    
1639    Tag: lal-20000727-linux-ppc
1640    Description:
1641    
1642     Made changes to support Linux PPC.
1643     p.s. I have confirmation that the 110.29 boot files work fine.
1644    
1645    ----------------------------------------------------------------------
1646    Name: Matthias Blume
1647    Date: 2000/07/27 17:40:00 JST
1648    Tag: blume-20000727-scripts
1649    Description:
1650    
1651    !!!! WARNING !!!!
1652    You must recompile the runtime system!
1653    !!!! WARNING !!!!
1654    
1655    This is basically another round of script-enhancements:
1656    
1657    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
1658       and undefine CM preprocessor symbols.
1659    
1660    2. ml-build avoids generating a new heap image if it finds that the
1661       existing one is still ok.  (The condition is that no ML file had to
1662       be recompiled and all ML files are found to be older that the heap
1663       file.)
1664    
1665       To make this work smoothly, I also hacked the runtime system as
1666       well as SMLofNJ.SysInfo to get access to the heap image suffix
1667       (.sparc-solaris, ...) that is currently being used.
1668    
1669       Moreover, the signature of CM.mk_standalone has changed.  See the
1670       CM manual.
1671    
1672    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
1673       CM manual for details.)
1674    
1675    4. More CM manual updates:
1676        - all of the above has been documented.
1677        - there is now a section describing the (CM-related) command line
1678          arguments that are accepted by the "sml" command
1679    
1680    ----------------------------------------------------------------------
1681    Name: Matthias Blume
1682    Date: 2000/07/25 16:20:00 JST
1683    Tag: blume-20000725-makedepend
1684    Description:
1685    
1686    Added a script called ml-makedepend.  This can be used in makefiles
1687    for Unix' make in a way very similar to the "makedepend" command for
1688    C.
1689    
1690    The script internally uses function CM.sources.
1691    
1692    Synopsis:
1693    
1694        ml-makedepend [-f makefile] cmfile targetname
1695    
1696    The default for the makefile is "makefile" (or "Makefile" should
1697    "makefile" not exist).
1698    
1699    ml-makedepend adds a cmfile/targetname-specific section to this
1700    makefile (after removing the previous version of this section).  The
1701    section contains a single dependency specification with targetname on
1702    the LHS (targetname is an arbitrary name), and a list of files derived
1703    from the cmfile on the RHS.  Some of the files on the RHS are
1704    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
1705    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
1706    path names.  The makefile writer is responsible for making sure that
1707    these variables have correct at the time "make" is invoked.
1708    
1709    ----------------------------------------------------------------------
1710    Name: Matthias Blume
1711    Date: 2000/07/22 23:30:00 JST
1712    Tag: blume-20000722-urlupdate
1713    Description:
1714    
1715    Changed BOOT and config/srcarchiveurl to point to BL server:
1716    
1717        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
1718    
1719    ----------------------------------------------------------------------
1720    Name: Matthias Blume
1721    Date: 2000/07/18 18:00:00 JST
1722    Tag: blume-20000718-Version_110_29
1723    Description:
1724    
1725    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
1726    
1727    2. Updated config/version to 110.29
1728    
1729    3. Updated config/srcarchiveurl
1730    
1731    3. New boot files!
1732       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
1733    
1734    ----------------------------------------------------------------------
1735    Name: Matthias Blume
1736    Date: 2000/07/11 13:58:00 JST
1737    Tag: blume-20000711-doctypo
1738    Description:
1739    
1740    Fixed a few typos in CM manual.
1741    
1742    ----------------------------------------------------------------------
1743    Name: Allen Leung
1744    Date: 2000/06/15 00:38:00
1745    Tag: leunga-20000704-sparc-x86
1746    
1747    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
1748    2. fix to x86 RET bug [found by Dan Grossman]
1749    3. sparc assembly bug fix for ticc instructions [found by Fermin]
1750    
1751       Affects c-- and moby only
1752    
1753    ----------------------------------------------------------------------
1754    Name: Matthias Blume
1755    Date: 2000/07/04 15:26:00
1756    Tag: blume-20000704-trigger
1757    Description:
1758    
1759    1. Improvements to CM manual.
1760    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
1761       of getting a back-trace.  The function, when called, raises an
1762       internal exception which explicitly carries the full back-trace history,
1763       so it is unaffected by any intervening handle-raise pairs ("trivial"
1764       or not).  The interactive loop will print that history once it arrives
1765       at top level.
1766       Short of having all exceptions implicitly carry the full history, the
1767       recommended way of using this facility is:
1768         - compile your program with instrumentation "on"
1769         - run it, when it raises an exception, look at the history
1770         - if the history is "cut off" because of some handler, go and modify
1771           your program so that it explicitly calls BTrace.trigger
1772         - recompile (still instrumented), and rerun; look at the full history
1773    
1774    ----------------------------------------------------------------------
1775    Name: Matthias Blume
1776    Date: 2000/07/03 15:36:00 JST
1777    Tag: blume-20000702-manual
1778    Description:
1779    
1780    Small corrections and updates to CM manual.
1781    
1782    ----------------------------------------------------------------------
1783    Name: Matthias Blume
1784    Date: 2000/06/29 16:04:00 JST
1785    Tag: blume-20000629-yacctool
1786    Description:
1787    
1788    Changes:
1789    
1790    1. Class "mlyacc" now takes separate arguments to pass options to
1791       generated .sml- and .sig-files independently.
1792    2. Corresponding CM manual updates.
1793    3. BTrace module now also reports call sites.  (However, for loop clusters
1794       it only shows from where the cluster was entered.)  There are associated
1795       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1796    
1797    ----------------------------------------------------------------------
1798    Name: Matthias Blume
1799    Date: 2000/06/27 16:51:00 JST
1800    Tag: blume-20000627-noweb
1801    Description:
1802    
1803    Changes:
1804    
1805     1. Implemented "subdir" and "witness" options for noweb tool.
1806        This caused some slight internal changes in CM's tool implementation.
1807     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
1808        some remaining issues from earlier path anchor changes.
1809     3. Updated CM manual accordingly.
1810    
1811     4. Changed implementation of back-tracing so that I now consider it
1812        ready for prime-time.
1813    
1814        In particular, you don't have to explicitly trigger the back-trace
1815        anymore.  Instead, if you are running BTrace-instrumented code and
1816        there is an uncaught exception (regardless of whether or not it was
1817        raised in instrumented code), the top-level evalloop will print
1818        the back-trace.
1819    
1820        Features:
1821    
1822          - Instrumented and uninstrumented code work together seemlessly.
1823            (Of course, uninstrumented code is never mentioned in actual
1824             back-traces.)
1825    
1826          - Asymptotic time- and space-complexity of instrumented code is
1827            equal to that of uninstrumented code.  (This means that
1828            tail-recursion is preserved by the instrumentation phase.)
1829    
1830          - Modules whose code has been instrumented in different sessions
1831            work together without problem.
1832    
1833          - There is no penalty whatsoever on uninstrumented code.
1834    
1835          - There is no penalty on "raise" expressions, even in
1836            instrumented code.
1837    
1838        A potential bug (or perhaps it is a feature, too):
1839    
1840          A back-trace reaches no further than the outermost instrumented
1841          non-trivial "raise".  Here, a "trivial" raise is one that is the
1842          sole RHS of a "handle" rule.  Thus, back-traces reach trough
1843    
1844               <exp> handle e => raise e
1845    
1846          and even
1847    
1848               <exp> handle Foo => raise Bar
1849    
1850          and, of course, through
1851    
1852               <exp> handle Foo => ...
1853    
1854         if the exception was not Foo.
1855    
1856         Back-traces always reach right through any un-instrumented code
1857         including any of its "handle" expressions, trivial or not.
1858    
1859       To try this out, do the following:
1860    
1861         - Erase all existing binfiles for your program.
1862           (You may keep binfiles for those modules where you think you
1863            definitely don't need back-tracing.)
1864         - Turn on back-trace instrumentation:
1865              SMLofNJ.Internals.BTrace.mode (SOME true);
1866         - Recompile your program.  (I.e., run "CM.make" or "use".)
1867         - You may now turn instrumentation off again (if you want):
1868              SMLofNJ.Internals.BTrace.mode (SOME false);
1869         - Run your program as usual.  If it raises an exception that
1870           reaches the interactive toplevel, then a back-trace will
1871           automatically be printed.  After that, the toplevel loop
1872           will print the exception history as usual.
1873    
1874    ----------------------------------------------------------------------
1875    Name: Matthias Blume
1876    Date: 2000/06/26 09:56:46 JST
1877    Tag: blume-20000626-setup
1878    Description:
1879    
1880    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1881          ML code before and after compiling a file (e.g., to set compiler
1882          flags)
1883    
1884    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1885              - associated changes to btrace.sml (BTrace instrumentation pass)
1886              - cleaner implementation of btimp.sml (BTrace tracing and report
1887                module)
1888    
1889    CM manual: * new path encoding documented
1890               * description of setup-parameter to "sml" added
1891    
1892    The biggest user-visible change to back-tracing is that it is no
1893    longer necessary to compile all traced modules within the same
1894    session.  (This was a real limitation.)
1895    
1896    ----------------------------------------------------------------------
1897    Name: Matthias Blume
1898    Date: 2000/06/24 12:40:00 JST
1899    Tag: blume-20000624-startup
1900    Description:
1901    
1902    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1903    bit too often -- to put it mildly. :)
1904    
1905    ----------------------------------------------------------------------
1906    Name: Matthias Blume
1907    Date: 2000/06/23 18:20:00 JST
1908    Tag: blume-20000623-btrace
1909    Description:
1910    
1911    This updates adds a backtrace facility to aid programmers in debugging
1912    their programs.  This involves the following changes:
1913    
1914    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1915       keeping track of the current call stack.  When programs are compiled
1916       in a special mode, the compiler will insert calls to these hooks
1917       into the user program.
1918       "Hook" means that it is possible for different implementations of
1919       back-tracing to register themselves (at different times).
1920    
1921    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1922       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1923       It can be turned on using this call:
1924         SMLofNJ.Internals.BTrace.mode (SOME true);
1925       Turning it off again:
1926         SMLofNJ.Internals.BTrace.mode (SOME false);
1927       Querying the current status:
1928         SMLofNJ.Internals.BTrace.mode NONE;
1929       Annotated programs are about twice as big as normal ones, and they
1930       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1931       where all hooks do nothing).  The slowdown with a plugin that is
1932       actually useful (such as the one supplied by default) is even greater,
1933       but in the case of the default plugin it is still only an constant
1934       factor (amortized).
1935    
1936    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1937       with a sub-structure BTrace for controlling back-tracing.  In particular,
1938       the above-mentioned function "mode" controls whether the annotation
1939       phase is invoked by the compiler.  Another important function is
1940       "trigger": when called it aborts the current execution and causes
1941       the top-level loop to print a full back-trace.
1942    
1943    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1944       for back-tracing.  It keeps track of the dynamic call stack and in
1945       addition to that it keeps a partial history at each "level" of that
1946       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1947       d and b (at separate times, dynamically), then the report will show:
1948    
1949       GOTO   d
1950             /c
1951       GOTO  \b
1952       CALL   a
1953    
1954       This shows that there was an initial non-tail call of a, then a
1955       tail-call to b or c, looping behavior in a cluster of functions that
1956       consist of b and c, and then a goto from that cluster (i.e., either from
1957       b or from c) to d.
1958    
1959       Note that (depending on the user program) the amount of information
1960       that the back-trace module has to keep track of at each level is bounded
1961       by a constant.  Thus, the whole implementation has the same asymptotical
1962       complexity as the original program (both in space and in time).
1963    
1964    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1965       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1966       which is raised by the "trigger" function mentioned above.
1967    
1968    Notes on usage:
1969    
1970    - Annotated code works well together with unannotated code:
1971    Unannotated calls simply do not show up at all in the backtrace.
1972    
1973    - It is not a good idea to let modules that were annotated during
1974    different sessions run at the same time.  This is because the compiler
1975    chooses small integers to identify individual functions, and there
1976    will be clashes if different modules were compiled in separate sessions.
1977    (Nothing will crash, and you will even be told about the clashes, but
1978    back-trace information will in general not be useful.)
1979    
1980    - Back-tracing can be confused by callcc and capture.
1981    
1982    - The only way of getting a back-trace right now is to explicitly
1983    invoke the "trigger" function from your user program.  Eventually, we
1984    should make every exception carry back-trace information (if
1985    available).  But since this creates more overhead at "raise"-time
1986    (similar to the current exnHistory overhead), I have not yet
1987    implemented this.  (The implementation will be rather easy.)  With
1988    exceptions carrying back-trace information, this facility will be even
1989    more useful because users don't need to modify their programs...
1990    
1991    - While it is possible to compile the compiler with back-trace
1992    annotations turned on (I did it to get some confidence in
1993    correctness), you must make absolutely sure that core.sml and
1994    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1995    be compiled with annotation because there is no core access yet, but
1996    if you compile btimp.sml with annotation, then the system will go into
1997    an infinite recursion and crash.)
1998    Since CM currently does not know about BTrace, the only way to turn
1999    annotations on and off for different modules of the compiler is to
2000    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
2001    this is awkward and clumsy.
2002    
2003    Sample sessions:
2004    
2005    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
2006    - SMLofNJ.Internals.BTrace.mode (SOME true);
2007    [autoloading]
2008    [autoloading done]
2009    val it = false : bool
2010    - structure X = struct
2011    -     fun main n = let
2012    -         fun a (x, 0) = d x
2013    -           | a (x, n) = b (x, n - 1)
2014    -         and b (x, n) = c (x, n)
2015    -         and c (x, n) = a (x, n)
2016    -         and d x = e (x, 3)
2017    -         and e (x, 0) = f x
2018    -           | e (x, n) = e (x, n - 1)
2019    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
2020    -           | f n = n * g (n - 1)
2021    -         and g n = a (n, 3)
2022    -     in
2023    -         f n
2024    -     end
2025    - end;
2026    structure X : sig val main : int -> int end
2027    - X.main 3;
2028    *** BACK-TRACE ***
2029    GOTO   stdIn:4.2-13.20: X.main[2].f
2030    GOTO-( stdIn:4.2-13.20: X.main[2].e
2031    GOTO   stdIn:4.2-13.20: X.main[2].d
2032         / stdIn:4.2-13.20: X.main[2].a
2033         | stdIn:4.2-13.20: X.main[2].b
2034    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2035    CALL   stdIn:4.2-13.20: X.main[2].g
2036    GOTO   stdIn:4.2-13.20: X.main[2].f
2037    GOTO-( stdIn:4.2-13.20: X.main[2].e
2038    GOTO   stdIn:4.2-13.20: X.main[2].d
2039         / stdIn:4.2-13.20: X.main[2].a
2040         | stdIn:4.2-13.20: X.main[2].b
2041    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2042    CALL   stdIn:4.2-13.20: X.main[2].g
2043    GOTO   stdIn:4.2-13.20: X.main[2].f
2044    GOTO-( stdIn:4.2-13.20: X.main[2].e
2045    GOTO   stdIn:4.2-13.20: X.main[2].d
2046         / stdIn:4.2-13.20: X.main[2].a
2047         | stdIn:4.2-13.20: X.main[2].b
2048    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2049    CALL   stdIn:4.2-13.20: X.main[2].g
2050    GOTO   stdIn:4.2-13.20: X.main[2].f
2051    CALL   stdIn:2.15-17.4: X.main[2]
2052    -
2053    
2054    (Note that because of a FLINt bug the above code currently does not
2055    compile without BTrace turned on.)
2056    
2057    Here is another example, using my modified Tiger compiler:
2058    
2059    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
2060    - SMLofNJ.Internals.BTrace.mode (SOME true);
2061    [autoloading]
2062    [autoloading done]
2063    val it = false : bool
2064    - CM.make "sources.cm";
2065    [autoloading]
2066    ...
2067    [autoloading done]
2068    [scanning sources.cm]
2069    [parsing (sources.cm):parse.sml]
2070    [creating directory CM/SKEL ...]
2071    [parsing (sources.cm):tiger.lex.sml]
2072    ...
2073    [wrote CM/sparc-unix/semant.sml]
2074    [compiling (sources.cm):main.sml]
2075    [wrote CM/sparc-unix/main.sml]
2076    [New bindings added.]
2077    val it = true : bool
2078    - Main.compile ("../testcases/merge.tig", "foo.out");
2079    *** BACK-TRACE ***
2080    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
2081    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2082    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2083    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2084    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2085    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2086    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
2087         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
2088    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
2089    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2090    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
2091    CALL   main.sml:1.18-118.4: Main.compile[2]
2092    -
2093    
2094    ----------------------------------------------------------------------
2095    Name: Matthias Blumen
2096    Date: 2000/06/21 18:00:00 JST
2097    Tag: blume-20000621-manual
2098    Description:
2099    
2100    CM manual update: Path environments documented.
2101    
2102    ----------------------------------------------------------------------
2103    Name: Matthias Blume
2104    Date: 2000/06/19 13:40:00
2105    Tag: blume-20000619-manual
2106    Description:
2107    
2108    CM manual and system/README update.  This only covers the fact that
2109    there are no more implicit anchors.  (Path environments and the "bind"
2110    option to "cm" have yet to be documented.)
2111    
2112    ----------------------------------------------------------------------
2113    Name: Matthias Blume
2114    Date: 2000/06/19 11:05:00 JST
2115    Tag: blume-20000619-chdir-bugfix
2116    Description:
2117    
2118    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
2119    
2120    ----------------------------------------------------------------------
2121    Name: Matthias Blume
2122    Date: 2000/06/18 22:00:10 JST
2123    Tag: blume-20000618-implicit-anchors-really-gone
2124    Description:
2125    
2126    I updates the previous HISTORY entry where I forgot to mention that
2127    implicit anchors are no longer with us.
2128    
2129    The current update also gets rid of the (now useless) controller
2130    CM.Control.implicit_anchors.
2131    
2132    ----------------------------------------------------------------------
2133    Name: Matthias Blume
2134    Date: 2000/06/16 17:30:00 JST
2135    Tag: blume-20000616-anchorenv
2136    Description:
2137    
2138    This patch implements the long anticipated (just kidding :) "anchor
2139    environment" mechanism.  In the course of doing this, I also
2140    re-implemented CM's internal "SrcPath" module from scratch.  The new
2141    one should be more robust in certain boundary cases.  In any case, it
2142    is a lot cleaner than its predecessor (IMHO).
2143    
2144    This time, although there is yet another boot file format change, I
2145    kept the unpickler backward-compatible.  As a result, no new bootfiles
2146    are necessary and bootstrapping is straightforward.  (You cannot read
2147    new bootfiles into an old system, but the other way around is no
2148    problem.)
2149    
2150    Visible changes:
2151    
2152    ** 0. Implicit path anchors (without the leading $-symbol) are no
2153    longer recognized at all. This means that such path names are not
2154    illegal either.  For example, the name basis.cm simply refers to a
2155    local file called "basis.cm" (i.e, the name is an ordinary path
2156    relative to .cm-files directory).  Or, to put it differently, only
2157    names that start with $ are anchored paths.
2158    
2159    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
2160    vanished.
2161    
2162    John (Reppy) had critizised this as soon as I originally proposed and
2163    implemented it, but at that time I did not really deeply believe
2164    him. :) Now I came full-circle because I need the $<singlearc> syntax
2165    in another place where it cannot be seen as an abbreviation for
2166    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
2167    seems to mean (i.e., it "expands" into the corresponding anchor
2168    value).
2169    
2170    However, when paths are used as members in CM description files, it
2171    continues to be true that there must be at least another arc after the
2172    anchor.  This is now enforced separately during semantic analysis
2173    (i.e., from a lexical/syntactical point of view, the notation is ok.)
2174    
2175    ** 2. The "cm" class now accepts an option "bind".  The option's value
2176    is a sub-option list of precisely two items -- one labeled "anchor"
2177    and the other one labeled "value".  As you might expect, "anchor" is
2178    used to specify an anchor name to be bound, and "value" specifies what
2179    the anchor is being bound to.
2180    
2181    The value must be a directory name and can be given in either standard
2182    syntax (including the possibility that it is itself an anchored path)
2183    or native syntax.
2184    
2185    Examples:
2186    
2187       foo.cm (bind:(anchor:bar value:$mystuff/bar))
2188       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
2189    
2190    and so on.
2191    
2192    The meaning of this is that the .cm-file will be processed with an
2193    augmented anchor environment where the given anchor(s) is/are bound to
2194    the given values(s).
2195    
2196    The rationale for having this feature is this: Suppose you are trying
2197    to use two different (already stable) libraries a.cm and b.cm (that
2198    you perhaps didn't write yourself).  Further, suppose each of these
2199    two libraries internally uses its own auxiliary library $aux/lib.cm.
2200    Normally you would now have a problem because the anchor "lib" can not
2201    be bound to more than one value globally.  Therefore, the project that
2202    uses both a.cm and b.cm must locally redirect the anchor to some other
2203    place:
2204    
2205       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
2206       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
2207    
2208    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
2209    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
2210    
2211    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
2212    you when you do so at the time of CM.stabilize).  Therefore, you can
2213    also use an anchored path as the value:
2214    
2215      a.cm (bind:(anchor:lib value:$a-lib))
2216      b.cm (bind:(anchor:lib value:$b-lib))
2217    
2218    Now you can globally configure (using the usual CM.Anchor.anchor or
2219    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
2220    itself is always locally bound, setting it globally is no longer
2221    meaningful or necessary (but it does not hurt either).  In fact, "lib"
2222    can still be used as a global anchor for separate purposes.  As a
2223    matter of fact, one can locally define "lib" in terms of a global
2224    "lib":
2225    
2226      a.cm (bind:(anchor:lib value:$lib/a))
2227      b.cm (bind:(anchor:lib value:$lib/b))
2228    
2229    ** 3: The encoding of path names has changed.  This affects the way
2230    path names are shown in CM's progress report and also the internal
2231    protocol encoding used for parallel make.
2232    
2233    The encoding now uses one or more ':'-separated segments.  Each
2234    segments corresponds to a file that has been specified relative to the
2235    file given by its preceding segment.  The first segment is either
2236    relative to the CWD, absolute, or anchored.  Each segment itself is
2237    basically a Unix pathname; all segments but the first are relative.
2238    
2239    Example:
2240    
2241       $foo/bar/baz.cm:a/b/c.sml
2242    
2243    This path denotes the file bar/a/b/c.sml relative to the directory
2244    denoted by anchor "foo".  Notice that the encoding also includes
2245    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
2246    paths are resolved relative to the .cm-files directory, so baz.cm must
2247    be ignored to get the "real" pathname.
2248    
2249    To make this fact more obvious, CM puts the names of such "virtual
2250    arcs" into parentheses when they appear in progress reports. (No
2251    parentheses will appear in the internal protocol encoding.)  Thus,
2252    what you really see is:
2253    
2254      $foo/bar/(baz.cm):a/b/c.sml
2255    
2256    I find this notation to be much more informative than before.
2257    
2258    Another new feature of the encoding is that special characters
2259    including parentheses, colons, (back)slashes, and white space are
2260    written as \ddd (where ddd is the decimal encoding of the character).
2261    
2262    *** The CM manual still needs to be updated.
2263    
2264    ----------------------------------------------------------------------
2265    Name: Allen Leung
2266    Date: 2000/06/15 00:38:00
2267    Tag: leunga-20000615-x86-peephole
2268    
2269    x86 Peephole fix by Fermin.  Affects c-- and moby only.
2270    
2271    ----------------------------------------------------------------------
2272    Name: Matthias Blume
2273    Date: 2000/06/12 11:40:00
2274    Tag: blume-20000612-parmakefix
2275    Description:
2276    
2277    More cleanup after changing the file naming scheme: This time I
2278    repaired the parallel make mechanism for CMB.make which I broke earlier.
2279    
2280    ----------------------------------------------------------------------
2281    Name: Allen Leung
2282    Date: 2000/06/09 01:25:00
2283    Tag: leunga-20000609-various
2284    
2285    None of these things should affect normal SML/NJ operations
2286    
2287    1. Peephole improvements provided by Fermin (c--)
2288    2. New annotation DEFUSE for adding extra dependence (moby)
2289    3. New X86 LOCK instructions (moby)
2290    4. New machine description language for reservation tables (scheduling)
2291    5. Fixes to various optimization/analysis modules (branch chaining, dominator
2292       trees etc.)
2293    6. I've changed the CM files so that they can work with versions
2294       110.0.6, 110.25 and 110.28
2295    
2296    ----------------------------------------------------------------------
2297    Name: Matthias Blume
2298    Date: 2000/06/09 12:40:00
2299    Tag: blume-20000609-log
2300    Description:
2301    
2302    - Removed all(?) remaining RCS Log entries from sources.
2303    
2304    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
2305      anchored paths).
2306    
2307    ----------------------------------------------------------------------
2308    Name: Matthias Blume
2309    Date: 2000/06/07 17:00:00 JST
2310    Tag: blume-20000607-no-implicit-anchors
2311    Description:
2312    
2313    1. This update changes the default setting for
2314    CM.Control.implicit_anchors from true to false.  This means that
2315    implicit anchors are no longer permitted by default.  I also tried to
2316    make sure that nothing else still relies on implicit anchors.
2317    (This is the next step on the schedule towards a CM that does not even
2318    have the notion of implicit anchors anymore.)
2319    
2320    2. More CM manual updates.
2321    
2322    3. I managed to track down and fix the pickling bug I mentioned last
2323    time.  Because of the previously existing workaround, this entails no
2324    immediate practical changes.
2325    
2326    ----------------------------------------------------------------------
2327    Name: Matthias Blume
2328    Date: 2000/06/06 11:15:00 JST
2329    Tag: blume-20000606-lazierpickle
2330    Description:
2331    
2332    !!!! NEW BOOT FILES !!!!
2333    
2334    * The main purpose of this update is to make library pickles lazier in
2335    order to reduce the initial space penalty for autoloading a library.
2336    As a result, it is now possible to have $smlnj/compiler.cm
2337    pre-registered.  This should take care of the many complaints or
2338    inquiries about missing structure Compiler.  This required changes to
2339    CM's internal data structures and small tweaks to some algorithms.
2340    
2341    As a neat additional effect, it is no longer necessary (for the sake
2342    of lean heap image files) to distinguish between a "minimal" CM and a
2343    "full" CM.  Now, there is only one CM (i.e., the "full" version:
2344    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
2345    interactive top level. ($smlnj/cm/minimal.cm is gone.)
2346    
2347    To make the life of compiler-hackers easier, "makeml" now also
2348    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
2349    words, after you bootstrap a new sml for the first time, you will not
2350    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
2351    around you will still have to do it, though.)
2352    
2353    * A second change consists of major updates to the CM manual.  There
2354    are now several appendices with summary information and also a full
2355    specification of the CM description file syntax.
2356    
2357    * In directory src/system I added the script "allcross".  This script
2358    invokes sml and cross-compiles the compiler for all supported
2359    architectures.  (Useful when providing a new set of boot files.)
2360    
2361    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
2362    added a small tweak to pickle-util.sml to work around this problem,
2363    but it is not a proper fix yet.  I will investigate further.  (The
2364    effect of the bug was an inflation of library pickle size.)
2365    
2366    * Version number increased to 110.28.1 (to avoid compatibility problems).
2367    
2368    ----------------------------------------------------------------------
2369    Name: Allen Leung
2370    Date: 2000/05/25 17:28 EDT
2371    Tag: leunga-20000525-ra
2372    Description:
2373    
2374      Fixed a bug in freezing phase of the register allocator.
2375    
2376    ----------------------------------------------------------------------
2377    Name: Allen Leung
2378    Date: 2000/05/15 22:53 EDT
2379    Tag: leunga-20000515-alpha-x86-ra
2380    Description:
2381    
2382      1. Alpha
2383    
2384          Slight cleanup.  Removed the instruction SGNXL
2385    
2386      2. X86
2387    
2388          Added the following instructions to the instruction set:
2389    
2390            ROLx, RORx,
2391            BTx, BTSx, BTLx, BTRx,
2392            XCHGx, and variants with the LOCK prefix
2393    
2394      3. Register Allocation
2395    
2396          The module ra-rewrite-with-renaming has been improved.
2397    
2398      These have no effect on SML/NJ.
2399    
2400    ----------------------------------------------------------------------
2401    Name: Matthias Blume
2402    Date: 2000/05/15 16:20:00 JST
2403    Tag: blume-20000515-lightrebuild
2404    Description:
2405    
2406    1. I added an alternative to "-rebuild" to "makeml".  The difference is
2407       that prior to calling CMB.make' the CM-variable "LIGHT" will be
2408       defined.  In effect, the command will not build any cross-compiler
2409       backends and therefore finish more quickly.
2410    
2411       The "fixpt" script also takes a "-light" switch to be able to use
2412       this new facility while compiling for a fixpoint.
2413    
2414    2. I replaced all mentions of anchored paths in group owner specifications
2415       with simple relative paths (usually starting with "..").
2416       The rationale is that a library's internal workings should not be
2417       compromised by the lack of some anchor.  (An anchor is necessary
2418       for someone who wants to refer to the library by an anchored path,
2419       but it should not be necessary to build the same library in the first
2420       place.)
2421    
2422    3. I changed the way CM's tool mechanism determines the shell command
2423       string used for things like ml-yacc etc. so that it does not break
2424       when CM.Control.implicit_anchors is turned off.
2425    
2426    ----------------------------------------------------------------------
2427    Name: Matthias Blume
2428    Date: 2000/05/12 18:20:00 JST
2429    Tag: blume-20000512-ml-build
2430    Description:
2431    
2432    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
2433    getting installed properly (by config/install.sh).
2434    
2435    ----------------------------------------------------------------------
2436    Name: Matthias Blume
2437    Date: 2000/05/12 17:30:00 JST
2438    Tag: blume-20000512-anchors
2439    Description:
2440    
2441    !!! NEW BOOT FILES !!!
2442    
2443    This change is in preparation of fading out support for "implicitly
2444    anchored path names".  I went through all sources and used the
2445    explicit (and relatively new) $-notation.  See system/README and the
2446    CM manual for more info on this.
2447    
2448    I also modified the anchoring scheme for some things such as "smlnj",
2449    "MLRISC", "cm", etc. to take advantage of the fact that explicit
2450    anchors are more expressive: anchor name and first arc do not have to
2451    coincide.  This entails the following user-visible change:
2452    
2453    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
2454    particular, when you fire up sml with a command-line argument, say,
2455    e.g.:
2456    
2457       sml '$smlnj/cmb.cm'
2458    
2459    At the ML toplevel prompt:
2460    
2461       CM.autoload "$smlnj/cmb.cm";
2462    
2463    There is also a new controller in CM.Control that can be used to turn
2464    off all remaining support for implicit anchors by saying:
2465    
2466        CM.autoload "$smlnj/
2467        #set CM.Control.implicit_anchors false;
2468    
2469    This causes CM to reject implicitly anchored paths.  This is (for the
2470    time being) less permissive than the "final" version where there will
2471    be no more such implicit anchors and relative paths will be just that:
2472    relative.
2473    
2474    The next step (version after next version?) will be to make the
2475    default for CM.Control.implicit_anchors false.  After the dust has
2476    settled, I can then produce the "final" version of this...
2477    
2478    Note: Since bootstrapping is a bit tricky, I provided new boot files.
2479    
2480    ----------------------------------------------------------------------
2481    Name: Matthias Blume
2482    Date: 2000/05/11 16:30:00 JST
2483    Tag: blume-20000511-sources
2484    Description:
2485    
2486    The main change is that I added function CM.sources as a generalized
2487    version of the earlier CM.makedepend.  This entails the following
2488    additional changes:
2489    
2490      - CM.makedepend has been dropped.
2491    
2492      - CM manual has been updated.
2493    
2494      - TOOLS signature and API have been changed.
2495    
2496    ----------------------------------------------------------------------
2497    Name: Allen Leung
2498    Date: 2000/05/10 21:17 EDT
2499    Tag: leunga-20000510-moby-c--ssa
2500    Description:
2501    
2502      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
2503    None of these affect SML/NJ.
2504    
2505    1. Register Allocation
2506    
2507        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
2508           This module tries to remove local (i.e. basic block level) redundancies
2509           during spilling.
2510    
2511        b. A new framework for performing region based register allocation.
2512           Not yet entirely functional.
2513    
2514    2. X86
2515    
2516       a. DefUse for POP was missing the stack pointer [found by Lal]
2517       b. Reload for CALL was incorrect in X86Spill [found by John]
2518       c. Various fixes in X86Spill so that it can be used correctly for
2519          the new spilling module.
2520    
2521    3. SSA/IR
2522    
2523       a. New module ir/dj-dataflow.sml implements elimination based
2524          data flow analysis.
2525    
2526    4. MLRiscGen
2527    
2528       a. Fix for gc type annotation
2529    
2530    5. MDGen
2531    
2532       Various fixes for machine description -> ml code translation.  For ssa
2533       only.
2534    
2535    ----------------------------------------------------------------------
2536    Name: Allen Leung
2537    Date: 2000/05/08 22:17 EDT
2538    Tag: leunga-20000508-labexp
2539    Description:
2540    
2541      Fermin has found a few assembly problems with constant expressions
2542      generated in LabelExp.  Mostly, the problems involve extra parentheses,
2543      which choke on dumb assemblers.  This is his fix.
2544    
2545    ----------------------------------------------------------------------
2546    Name: Dave MacQueen
2547    Date: 2000/04/09 14:00 EDT
2548    Tag: dbm-20000502-Version_110_28
2549    Description:
2550    
2551    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
2552    
2553    2. Updated config/version to 110.28
2554    
2555    3. Updated config/srcarchiveurl
2556    
2557    3. New boot files!
2558       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
2559    
2560    ----------------------------------------------------------------------
2561    Name: Matthias Blume
2562    Date: 2000/05/01 19:05:00 JST
2563    Tag: blume-20000501-noweb
2564    Description:
2565    
2566    A new noweb tool has been added.  The existing system is entirely
2567    unaffected by this, but some CM users have asked for renewed noweb
2568    support.  Everything is documented in the CM manual.
2569    
2570    New (plugin) libraries:
2571    
2572       noweb-tool.cm
2573       nw-ext.cm
2574    
2575    ----------------------------------------------------------------------
2576    Name: Dave MacQueen
2577    Date: 2000/04/30 12:40PM EDT
2578    Tag: dbm-20000430-bug_fixes
2579    Description:
2580    
2581    1. Fix for bug 1498
2582       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
2583       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
2584         added toRealArray function
2585       smlnj/src/compiler/MiscUtil/print/ppobj.sml
2586         added check for tag Obj.RealArray to array printing case in ppObj
2587    
2588    2. Fix for bug 1510
2589       smlnj/src/compiler/Semant/types/typesutil.sml
2590         fixed definition of dummyargs (used by equalTycon) so that
2591         dummy args are distinct types
2592    
2593    ----------------------------------------------------------------------
2594    Name: Matthias Blume
2595    Date: 2000/04/30 01:00:00 JST
2596    Tag: blume-20000430-versions
2597    Description:
2598    
2599    1. CM version numbering added.  This is an implementation of Lal's
2600       proposal for adding version numbers and version checking to .cm
2601       files.  Lal said that his proposal was just that -- a proposal.
2602       For the time being I went ahead and implemented it so that people
2603       can comment on it.  Everything is completely backward-compatible
2604       (except for the stable library format, i.e., new bootfiles!).
2605    
2606       As usual, see the CM manual for details.
2607    
2608    2. An alternative syntax for anchored paths has been implemented.
2609       Dave has recently voiced the same concerns that I had when I did
2610       this, so there should be some support.  My take is that eventually
2611       I will let support for the current syntax (where anchors are
2612       "implicit") fade out in favor of the new, explicit syntax.
2613       In order to be backward-compatible, both old and new syntax are
2614       currently supported.
2615    
2616       Again, see the CM manual for details.
2617    
2618    3. Parallel make is trying to be slightly smarter:  When the master
2619       process finds a "bottleneck", i.e., when there is only one
2620       compilation unit that can be compiled and everybody else is
2621       waiting on it, then it will simply compile it directly instead
2622       of clumsily telling one of the slaves to do it.
2623    
2624    4. Support for "unsharing" added.  This is necessary in order to be
2625       able to have two different versions of the same library running
2626       at the same time (e.g., for trying out a new MLRISC while still
2627       having the old MLRISC linked into the current compiler, etc.)
2628       See the CM manual.
2629    
2630    5. Simple "makedepend" functionality added for generating Makefile
2631       dependency information.  (This is rather crude at the moment.
2632       Expect some changes here in the future.)
2633    
2634    6. ".fun" added as a recognized suffix for ML files. Also documented
2635       explicitly in the manual that the fallback behavior (unknown suffix
2636       -> ML file) is not an official feature!
2637    
2638    7. Small changes to the pickler for stable libraries.
2639    
2640    8. Several internal changes to CM (for cleanup/improvement).
2641    
2642    
2643    !!!! NEW BINFILES !!!!
2644    
2645    ----------------------------------------------------------------------
2646    Name: Matthias Blume
2647    Date: 2000/04/28 17:30:00 JST
2648    Tag: blume-20000428-pathconfig
2649    Description:
2650    
2651    1. I changed config/install.sh to remove duplicate entries from the
2652       lib/pathconfig file at the end.  Moreover, the final version of
2653       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
2654       in src/system/installml.
2655    
2656    2. The config/install.sh script now consistently uses relative
2657       pathnames in lib/pathconfig whenever the anchor is in the lib
2658       directory.  (So far this was true for the libraries that come
2659       pre-compiled and bundled as part of the bootfiles but not for
2660       libraries that are compiled by the script itself.)
2661    
2662    ----------------------------------------------------------------------
2663    Name: Matthias Blume
2664    Date: 2000/04/26 13:10:00 JST
2665    Tag: blume-20000426-fun_suffix
2666    Description:
2667    
2668    Added ".fun" as a recognized file name suffix (for ML code).
2669    
2670    ----------------------------------------------------------------------
2671    Name: Allen Leung
2672    Date: 2000/04/25 17:00:00 EST
2673    Tag: leunga-20000425-alpha-ra
2674    Description:
2675    
2676    1. Alpha
2677    
2678        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
2679    in C--.
2680    
2681    2. RA
2682    
2683       Added a flag "ra-dump-size" to print out the size of the flowgraph
2684       and the interference graph.
2685    
2686    ----------------------------------------------------------------------
2687    Name: Dave MacQueen
2688    Date: 2000/04/25/
2689    Tag: dbm-20000425-mlyacc_doc_examples
2690    Description:
2691      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
2692      Updated all three examples in src/ml-yacc/examples to run
2693      under 110.* using CM.make.
2694    
2695    ----------------------------------------------------------------------
2696    Name: Allen Leung
2697    Date: 2000/04/20 23:04:00 EST
2698    Tag: leunga-20000420-ssa-c---stuff
2699    Description:
2700    
2701      This update synchronizes my repository with Yale's.  Most of these
2702    changes, however, do not affect SML/NJ at all (the RA is an exception).
2703    
2704    1. Register Allocator
2705    
2706       a. An improvement in the interference graph construction:
2707          Given a copy
2708    
2709                s <- t
2710    
2711          no interference edge between s and t is added for this definition of s.
2712    
2713       b. I've added two new spill heuristic modules that Fermin and I developed
2714          (in the new library RA.cm). These are unused in SML/NJ but maybe
2715          useful for others (Moby?)
2716    
2717    2. X86
2718    
2719       a. Various fixes in the backend provided by Fermin [C--] and Lal.
2720    
2721    3. Alpha
2722    
2723       a. Added the BSR instruction and code generation that goes with it [C--]
2724       b. Other fixes too numerous to recount provided by Fermin [C--]
2725    
2726    4. Regmaps
2727    
2728       a. The regmaps are not initialized with the identity physical bindings
2729          at creation time.  This is unneeded.
2730    
2731    5. MLRISC Optimizations
2732    
2733       a. The DJ-Graph module can now compute the iterated dominance frontiers
2734          intersects with liveness incrementally in linear time! Woohoo!
2735          This is now used in my new SSA construction algorithm.
2736    
2737       b. THe branch reorganization module is now smarter about linear chains of
2738          basic blocks.
2739    
2740    
2741    ----------------------------------------------------------------------
2742    Name: Matthias Blume
2743    Date: 2000/04/12 13:52:00 JST
2744    Tag: blume_main_v110p27_1
2745    Description:
2746    
2747    Changed install.sh script to handle archive files without version number
2748    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
2749    name of the boot file archive.
2750    
2751    ----------------------------------------------------------------------
2752    Name: Dave MacQueen
2753    Date: 2000/04/09 14:00 EDT
2754    Tag: dbm-20000410-Version_110_27
2755    Description:
2756    
2757    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
2758    
2759    2. Updated src/config/version to 110.27
2760    
2761    3. New boot files!
2762    
2763    ----------------------------------------------------------------------
2764    Name: Allen Leung
2765    Date: 2000/04/09 19:09:00 EST
2766    Tag: leunga-20000409-misc
2767    Description:
2768    
2769    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
2770    
2771    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
2772    
2773    ----------------------------------------------------------------------
2774    Name: Stefan
2775    Date: 2000/04/07 10:00:00 EDT
2776    Tag: monnier-20000406-branch-handling
2777    Description:
2778    
2779    Improved handling of branches (mostly those generated from
2780    polymorphic equality), removed switchoff and changed the
2781    default optimization settings (more cpsopt and less flintopt).
2782    
2783    ----------------------------------------------------------------------
2784    Name: Allen Leung
2785    Date: 2000/04/06 01:30:00 EST
2786    Tag: leunga-20000406-peephole-x86-SSA-2
2787    Description:
2788    
2789       Forgot a few files.
2790    
2791    ----------------------------------------------------------------------
2792    Name: Allen Leung
2793    Date: 2000/04/06 00:36:00 EST
2794    Tag: leunga-20000406-peephole-x86-SSA
2795    Description:
2796    
2797    1.  New Peephole code
2798    
2799    2.  Minor improvement to X86 instruction selection
2800    
2801    3.  Various fixes to SSA and machine description -> code translator
2802    
2803    ----------------------------------------------------------------------
2804    Name: Matthias Blume
2805    Date: 2000/04/05 12:30:00 JST
2806    Tag: blume_main_v110p26p2_3
2807    Description:
2808    
2809    This update just merges three minor cosmetic updates to CM's sources
2810    to get ready for the 110.27 code freeze on Friday.  No functionality
2811    has changed.
2812    
2813    ----------------------------------------------------------------------
2814    Name: Allen Leung
2815    Date: 2000/04/04 19:39:00 EST
2816    Tag: leunga-20000404-x86-asm
2817    Description:
2818    
2819    1.  Fixed a problem in X86 assembly.
2820    
2821        Things like
2822    
2823           jmp %eax
2824           jmp (%eax)
2825    
2826        should be output as
2827    
2828           jmp *%eax
2829           jmp *(%eax)
2830    
2831    2.  Assembly output
2832    
2833          Added a new flag
2834    
2835              "asm-indent-copies" (default to false)
2836    
2837          When this flag is on, parallel copies will be indented an extra level.
2838    
2839    ----------------------------------------------------------------------
2840    Name: Allen Leung
2841    Date: 2000/04/04 03:18:00 EST
2842    Tag: leunga-20000404-C--Moby
2843    Description:
2844    
2845        All of these fixes are related to C--, Moby, and my own optimization
2846        stuff; so they shouldn't affect SML/NJ.
2847    
2848    1.  X86
2849    
2850        Various fixes related floating point, and extensions.
2851    
2852    2.  Alpha
2853    
2854        Some extra patterns related to loads with signed/zero extension
2855        provided by Fermin.
2856    
2857    3.  Assembly
2858    
2859        When generating assembly, resolve the value of client defined constants,
2860        instead of generating symbolic values.  This is controlled by the
2861        new flag "asm-resolve-constants", which is default to true.
2862    
2863    4.  Machine Descriptions
2864    
2865        a. The precedence parser was slightly broken when parsing infixr symbols.
2866        b. The type generalizing code had the bound variables reversed, resulting
2867           in a problem during arity raising.
2868        c. Various fixes in machine descriptions.
2869    
2870    ----------------------------------------------------------------------
2871    Name: Matthias Blume
2872    Date: 2000/04/03 16:05:00 JST
2873    Tag: blume_main_v110p26p2_2
2874    Description:
2875    
2876    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
2877    now done via the ordinary static environment that is context to each
2878    compilation unit.
2879    
2880    To this end, I arranged that instead of "structure Core" as "structure
2881    _Core" is bound in the pervasive environment.  Core access is done via
2882    _Core (which can never be accidentally rebound because _Core is not a
2883    legal surface-syntax symbol).
2884    
2885    The current solution is much cleaner because the core environment is
2886    now simply part of the pervasive environment which is part of every
2887    compilation unit's context anyway.  In particular, this eliminates all
2888    special-case handling that was necessary until now in order to deal
2889    with dynamic and symbolic parts of the core environment.
2890    
2891    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2892    compilation manager's bootstrap compiler (actually: in the "init group"
2893    handling).  See the comments in src/system/smlnj/init/init.cmi for
2894    more details.
2895    
2896    I also tried to track down all mentions of "Core" (as string argument
2897    to Symbol.strSymbol) in the compiler and replaced them with a
2898    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
2899    name appears in one place only.
2900    
2901    Binfile and bootfile format have not changed, but the switchover from
2902    the old "init.cmi" to the new one is a bit tricky, so I supplied new
2903    bootfiles anyway.
2904    
2905    ----------------------------------------------------------------------
2906    Name: Allen Leung
2907    Date: 2000/04/02 21:17:00 EST
2908    Tag: leunga-20000402-mltree
2909    Description:
2910    
2911       1. Renamed the constructor CALL in MLTREE by popular demand.
2912       2. Added a bunch of files from my repository.  These are currently
2913          used by other non-SMLNJ backends.
2914    
2915    ----------------------------------------------------------------------
2916    Name: Allen Leung
2917    Date: 2000/03/31 21:15:00 EST
2918    Tag: leunga-20000331-aliasing
2919    Description:
2920    
2921    This update contains a rewritten (and hopefully more correct) module
2922    for extracting aliasing information from CPS.
2923    
2924       To turn on this feature:
2925    
2926            Compiler.Control.CG.memDisambiguate := true
2927    
2928       To pretty print the region information with assembly
2929    
2930           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
2931    
2932       To control how many levels of aliasing information are printed, use:
2933    
2934           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
2935    
2936       The default of n is 3.
2937    
2938    ----------------------------------------------------------------------
2939    Name: David MacQueen
2940    Date: 2000/03/31 11:15:00 EST
2941    Tag: dbm-20000331-runtime_fix
2942    Description:
2943    
2944    This update contains:
2945    
2946    1. runtime/c-lib/c-libraries.c
2947       includes added in revision 1.2 caused compilation errors on hppa-hpux
2948    
2949    2. fix for bug 1556
2950       system/Basis/Implementation/NJ/internal-signals.sml
2951    
2952    ----------------------------------------------------------------------
2953  Name: Matthias Blume  Name: Matthias Blume
2954  Date: 2000/03/31 18:00:00 JST  Date: 2000/03/31 18:00:00 JST
2955  Tag: blume_main_v110p26p2_1  Tag: blume_main_v110p26p2_1
# Line 67  Line 3004 
3004     (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
3005     used).     used).
3006    
3007     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
3008     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
3009     Makefiles or Shell scripts.     Makefiles or Shell scripts.
3010    
3011     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 80  Line 3017 
3017    
3018     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
3019    
3020     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
3021     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
3022    
3023  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 111  Line 3048 
3048       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3049       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3050       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
3051       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3052       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3053       Dave's hands. :)       Dave's hands. :)
3054    
# Line 214  Line 3151 
3151    
3152       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
3153       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
3154       testl is perferable to andl.       testl is preferable to andl.
3155    
3156  7. RA (x86 only)  7. RA (x86 only)
3157    
# Line 396  Line 3333 
3333    
3334  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
3335  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
3336  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
3337  explanation.  explanation.
3338    
3339  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
3340  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
3341  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
3342  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 443  Line 3380 
3380  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
3381  <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,
3382  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
3383  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
3384  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
3385  and, thus, saves a lot of time.  and, thus, saves a lot of time.
3386    
# Line 493  Line 3430 
3430    it from that remote directory.    it from that remote directory.
3431    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
3432    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
3433    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
3434    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
3435    
3436    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.588  
changed lines
  Added in v.952

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