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 589, Fri Mar 31 16:14:36 2000 UTC revision 959, Fri Oct 12 21:21:41 2001 UTC
# Line 13  Line 13 
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Allen Leung
17    Date: 2001/10/12 17:18:32 EDT 2001
18    Tag: leung-20011012-x86-printflowgraph
19    Description:
20    
21    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
22    
23    ----------------------------------------------------------------------
24    Name: Lal George
25    Date: 2001/10/11 23:51:34 EDT
26    Tag: george-20011011-too-many-instrs
27    Description:
28    
29    The representation of a program point never expected to see more
30    than 65536 instructions in a basic block!
31    
32    ----------------------------------------------------------------------
33    Name: Lal George
34    Date: 2001/10/09 09:41:37 EDT
35    Tag: george-20011008-mlrisc-labels
36    Description:
37    
38    Changed the machine description files to support printing of
39    local and global labels in assembly code, based on host assembler
40    conventions.
41    
42    ----------------------------------------------------------------------
43    Name: Matthias Blume
44    Date: 2001/09/25 15:25:00 EDT
45    Tag: blume-20010925-exninfo
46    Description:
47    
48    I provided a non-hook implementation of exnName (at the toplevel) and
49    made the "dummy" implementation of exnMessage (at the toplevel) more
50    useful: if nothing gets "hooked in", then at least you are going to
51    see the exception name and a message indicating why you don't see more.
52    
53    [For the time being, programs that need exnMessage and want to use
54    ml-build should either use General.exnMessage (strongly recommended) or
55    refer to structure General at some other point so that CM sees a
56    static dependency.]
57    
58    [Similar remarks go for "print" and "use":  If you want to use their
59    functionality in stand-alone programs generated by ml-build, then use
60    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
61    
62    ----------------------------------------------------------------------
63    Name: Matthias Blume
64    Date: 2001/09/20 17:28:00 EDT
65    Tag: blume-20010920-slowfp
66    Description:
67    
68    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
69    
70    ----------------------------------------------------------------------
71    Name: Matthias Blume
72    Date: 2001/09/20 17:20:00 EDT
73    Tag: blume-20010920-canonicalpaths
74    Description:
75    
76    0. Updated the BOOT file (something that I forgot to do earlier).
77    
78    1. Small internal change to CM so that it avoids "/../" in filenames
79       as much as possible (but only where it is safe).
80    
81    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
82       that arguments that contain delimiters are passed through correctly.
83       This change also means that all "special" arguments of the form
84       @SMLxxx... must come first.
85    
86    3. Changed install script to put relative anchor names for tool commands
87       into pathconfig.
88    
89    ----------------------------------------------------------------------
90    Name: Matthias Blume
91    >>>>>>> 1.169
92    Date: 2001/09/18 15:35:00 EDT
93    Tag: blume-20010918-readme11036
94    Description:
95    
96    Added README files.
97    
98    ----------------------------------------------------------------------
99    Name: Matthias Blume
100    Date: 2001/09/18 11:45:00 EDT
101    Tag: Release_110_36 (retag)
102    Description:
103    
104    Fixed mistake in config/preloads. Retagged as 110.36.
105    
106    ----------------------------------------------------------------------
107    Name: Matthias Blume
108    Date: 2001/09/18 09:40:00 EDT
109    Tag: Release_110_36_orig (tag changed)
110    Description:
111    
112    New version (110.36).  New bootfiles.
113    
114    ----------------------------------------------------------------------
115    Name: Matthias Blume
116    Date: 2001/09/14 16:15:00 EDT
117    Tag: blume-20010914-x86fastfp
118    Description:
119    
120    John committed some changes that Allen made, in particular a (hopefully)
121    correctly working version of the x86-fp module.
122    
123    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
124    flag to "true".  Everything seems to compile to a fixpoint ok, and
125    "mandelbrot" speeds up by about 15%.
126    
127    ----------------------------------------------------------------------
128    Name: Matthias Blume
129    Date: 2001/09/13 11:20:00 EDT
130    Tag: blume-20010913-minimal
131    Description:
132    
133    1. Stefan Monnier's patch to fix a miscompilation problem that
134       was brought to light by John Reppy's work on Moby.
135    
136    2. Implemented a minimal "structure Compiler" that contains just
137       "version" and "architecture".  The minimal version will be
138       available when the full version is not.  This is for backward-
139       compatibility with code that wants to test Compiler.version.
140    
141    ----------------------------------------------------------------------
142    Name: Matthias Blume
143    Date: 2001/08/28 14:03:00 EDT
144    Tag: blume-20010828-ml-lex
145    Description:
146    
147    Fix for bug 1581, received from Neophytos Michael.
148    
149    ----------------------------------------------------------------------
150    Name: Matthias Blume
151    Date: 2001/08/27 11:20:00 EDT
152    Tag: blume-20010827-readme11035
153    Description:
154    
155    Fleshed out the README file for 110.35.
156    
157    ----------------------------------------------------------------------
158    Name: Matthias Blume
159    Date: 2001/08/24 17:10:00 EDT
160    Tag: Release_110_35
161    Description:
162    
163    New version number (110.35).  New bootfiles.
164    
165    ----------------------------------------------------------------------
166    Name: Lal George
167    Date: 2001/08/24 13:47:18 EDT 2001
168    Tag: george-20010824-MLRISC-graphs
169    Description:
170    
171     removed clusters from MLRISC completely and replaced with graphs.
172    
173    ----------------------------------------------------------------------
174    Name: Matthias Blume
175    Date: 2001/08/23 17:50:00 EDT
176    Tag: blume-20010823-toplevel
177    Description:
178    
179    - some reorganization of the code that implements various kinds of
180      environments in the compiler (static, dynamic, symbolic, combined)
181    - re-implemented the EnvRef module so that evalStream works properly
182      (if the stream contains references to "use", "CM.make", etc.)
183    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
184    
185    ----------------------------------------------------------------------
186    Name: Matthias Blume
187    Date: 2001/08/20 15:50 EDT
188    Tag: blume20010820-slipup
189    Description:
190    
191    I forgot to commit a few files.  Here they are...
192    
193    ----------------------------------------------------------------------
194    Name: Matthias Blume
195    Date: 2001/08/20 15:35:00 EDT
196    Tag: blume-20010820-debugprof
197    Description:
198    
199    !!!! NEW BOOTFILES !!!!
200    
201    This is another round of reorganizing the compiler sources.  This
202    time the main goal was to factor out all the "instrumentation"
203    passes (for profiling and backtracing) into their own library.
204    The difficulty was to do it in such a way that it does not depend
205    on elaborate.cm but only on elabdata.cm.
206    
207    Therefore there have been further changes to both elaborate.cm and
208    elabdata.cm -- more "generic" things have been moved from the former
209    to the latter.  As a result, I was forced to split the assignment
210    of numbers indicating "primtyc"s into two portions: SML-generic and
211    SML/NJ-specific.  Since it would have been awkward to maintain,
212    I bit the bullet and actually _changed_ the mapping between these
213    numbers and primtycs.  The bottom line of this is that you need
214    a new set of bin- and bootfiles.
215    
216    I have built new bootfiles for all architectures, so doing a fresh
217    checkout and config/install.sh should be all you need.
218    
219    The newly created library's name is
220    
221        $smlnj/viscomp/debugprof.cm
222    
223    and its sources live under
224    
225        src/compiler/DebugProf
226    
227    ----------------------------------------------------------------------
228    Name: Matthias Blume
229    Date: 2001/08/15 17:15:00 EDT
230    Tag: blume-20010815-compreorg
231    Description:
232    
233    This is a first cut at reorganizing the CM libraries that make up the
234    core of the compiler.  The idea is to separate out pieces that could
235    be used independently by tools, e.g., the parser, the typechecker, etc.
236    
237    The current status is a step in this direction, but it is not quite
238    satisfactory yet.  Expect more changes in the future.
239    
240    Here is the current (new) organization...
241    
242        What used to be $smlnj/viscomp/core.cm is now divided into
243        six CM libraries:
244    
245             $smlnj/viscomp/basics.cm
246                           /parser.cm
247                           /elabdata.cm
248                           /elaborate.cm
249                           /execute.cm
250                           /core.cm
251    
252        The CM files for these libraries live under src/system/smlnj/viscomp.
253        All these libraries are proxy libraries that contain precisely
254        one CM library component.  Here are the locations of the components
255        (all within the src/compiler tree):
256    
257             Basics/basics.cm
258             Parse/parser.cm
259             ElabData/elabdata.cm
260             Elaborator/elaborate.cm
261             Execution/execute.cm
262             core.cm
263    
264         [This organization is the same that has been used already
265         for a while for the architecture-specific parts of the visible
266         compiler and for the old version of core.cm.]
267    
268         As you will notice, many source files have been moved from their
269         respective original locations to a new home in one of the above
270         subtrees.
271    
272         The division of labor between the new libraries is the following:
273    
274             basics.cm:
275                - Simple, basic definitions that pertain to many (or all) of
276                  the other libraries.
277             parser.cm:
278                - The SML parser, producing output of type Ast.dec.
279                - The type family for Ast is also defined and exported here.
280             elabdata.cm:
281                - The datatypes that describe input and output of the elaborator.
282                  This includes types, absyn, and static environments.
283             elaborator.cm:
284                - The SML/NJ type checker and elaborator.
285                  This maps an Ast.dec (with a given static environment) to
286                  an Absyn.dec (with a new static environment).
287                - This libraries implements certain modules that used to be
288                  structures as functors (to remove dependencies on FLINT).
289             execute.cm:
290                - Everything having to do with executing binary code objects.
291                - Dynamic environments.
292             core.cm:
293                - SML/NJ-specific instantiations of the elaborator and MLRISC.
294                - Top-level modules.
295                - FLINT (this should eventually become its own library)
296    
297    Notes:
298    
299    I am not 100% happy with the way I separated the elaborator (and its
300    data structures) from FLINT.  Two instances of the same problem:
301    
302        1. Data structures contain certain fields that carry FLINT-specific
303           information.  I hacked around this using exn and the property list
304           module from smlnj-lib.  But the fact that there are middle-end
305           specific fields around at all is a bit annoying.
306    
307        2. The elaborator calculates certain FLINT-related information.  I tried
308           to make this as abstract as I could using functorization, but, again,
309           the fact that the elaborator has to perform calculations on behalf
310           of the middle-end at all is not nice.
311    
312        3. Having to used exn and property lists is unfortunate because it
313           weakens type checking.  The other alternative (parameterizing
314           nearly *everything*) is not appealing, though.
315    
316    I removed the "rebinding =" warning hack because due to the new organization
317    it was awkward to maintain it.  As a result, the compiler now issues some of
318    these warnings when compiling init.cmi during bootstrap compilation. On
319    the plus side, you also get a warning when you do, for example:
320       val op = = Int32.+
321    which was not the case up to now.
322    
323    I placed "assign" and "deref" into the _Core structure so that the
324    code that deals with the "lazy" keyword can find them there.  This
325    removes the need for having access to the primitive environment
326    during elaboration.
327    
328    ----------------------------------------------------------------------
329    Name: Matthias Blume
330    Date: 2001/08/13
331    Tag: blume-20010813-closures
332    Description:
333    
334    This fix was sent to us by Zhong Shao.  It is supposed to improve the
335    performance of certain loops by avoiding needless closure allocation.
336    
337    ----------------------------------------------------------------------
338    Name: Lal George
339    Date: 2001/07/31 10:03:23 EDT 2001
340    Tag: george-20010731-x86-fmalloc
341    Description: Fixed bug in x86 calls
342    
343        There was a bug where call instructions would mysteriously
344        vanish. The call instruction had to be one that returned
345        a floating point value.
346    
347    ----------------------------------------------------------------------
348    Name: Lal George
349    Date: 2001/07/19 16:36:29 EDT 2001
350    Tag: george-20010719-simple-cells
351    Description:
352    
353    I have dramatically simplified the interface for CELLS in MLRISC.
354    
355    In summary, the cells interface is broken up into three parts:
356    
357      1. CellsBasis : CELLS_BASIS
358    
359            CellsBasis is a top level structure and common for all
360            architectures.  it contains the definitions of basic datatypes
361            and utility  functions over these types.
362    
363      2. functor Cells() : CELLS
364    
365            Cells generates an interface for CELLS that incorporates the
366            specific resources on the target architecture, such as the
367            presence of special register classes, their number and size,
368            and various useful substructures.
369    
370      3. <ARCH>CELLS
371    
372            e.g. SparcCells: SPARCCELLS
373    
374            <ARCH>CELLS usually contains additional bindings for special
375            registers  on the architecture, such as:
376    
377                    val r0 : cell           (* register zero *)
378                    val y : cell            (* Y register *)
379                    val psr : cell          (* processor status register *)
380                    ...
381    
382            The structure returned by applying the Cells functor is opened
383            in this interface.
384    
385    The main implication of all this is that the datatypes for cells is
386    split between CellsBasis and CELLS -- a fairly simple change for user
387    code.
388    
389    In the old scheme the CELLS interface had a definitional binding of
390    the form:
391    
392            signature CELLS = sig
393    
394               structure CellsBasis = CellsBasis
395    
396               ...
397    
398            end
399    
400    With all the sharing constraints that goes on in MLRISC, this old
401    design  quickly leads to errors such as:
402    
403            "structure definition spec inside of sharing ... "
404    
405    
406    and appears to require an unacceptable amount of sharing and where
407    constraint hackery.
408    
409    I think this error message (the interaction of definitional specs and
410    sharing) requires more explanation on our web page.
411    
412    ----------------------------------------------------------------------
413    Name: Matthias Blume
414    Date: 2001/07/19 15:00:00 EDT
415    Tag: blume-20010719-libreorg
416    Description:
417    
418    This update puts together a fairly extensive but straightforward change
419    to the way the libraries that implement the interactive system are
420    organized:
421    
422       The biggest change is the elimination of structure Compiler.  As a
423       replacement for this structure, there is now a CM library
424       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
425       that exports all the substructures of the original structure Compiler
426       directly.  So instead of saying Compiler.Foo.bar one now simply
427       says Foo.bar.  (The CM libraries actually export a collection of
428       structures that is richer than the collection of substructures of
429       structure Compiler.)
430    
431       To make the transition smooth, there is a separate library called
432       $smlnj/compiler/compiler.cm which puts together and exports the
433       original structure Compiler (or at least something very close to it).
434    
435       There are five members of the original structure Compiler
436       that are not exported directly but which instead became members
437       of a new structure Backend (described by signature BACKEND).  These are:
438       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
439       Interact (: INTERACT), structure Machine (: MACHINE), and val
440       architecture (: string).
441    
442       Structure Compiler.Version has become structure CompilerVersion.
443    
444       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
445       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
446       or x86, respectively.
447       Each of these exports the same frontend structures that
448       $smlnj/compiler.cm exports.  But they do not have a structure Backend
449       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
450       Hppa, PPC, Sparc, or X86, respectively.
451    
452       Library $smlnj/compiler/all.cm exports the union of the exports of
453       $smlnj/compiler/<arch>.cm
454    
455       There are no structures <Arch>Compiler anymore, use
456       $smlnj/compiler/<arch>.cm instead.
457    
458       Library host-compiler-0.cm is gone.  Instead, the internal library
459       that instantiates CM is now called cm0.cm.  Selection of the host
460       compiler (backend) is no longer done here but. (Responsibility for it
461       now lies with $smlnj/compiler/current.cm.  This seems to be more
462       logical.)
463    
464       Many individual files have been moved or renamed.  Some files have
465       been split into multiple files, and some "dead" files have been deleted.
466    
467    Aside from these changes to library organization, there are also changes
468    to the way the code itself is organized:
469    
470       Structure Binfile has been re-implemented in such a way that it no
471       longer needs any knowledge of the compiler.  It exclusively deals
472       with the details of binfile layout.  It no longer invokes the
473       compiler (for the purpose of creating new prospective binfile
474       content), and it no longer has any knowledge of how to interpret
475       pickles.
476    
477       Structure Compile (: COMPILE) has been stripped down to the bare
478       essentials of compilation.  It no longer deals with linking/execution.
479       The interface has been cleaned up considerably.
480    
481       Utility routines for dealing with linking and execution have been
482       moved into their own substructures.
483    
484       (The ultimate goal of these changes is to provide a light-weight
485       binfile loader/linker (at least for, e.g., stable libraries) that
486       does not require CM or the compiler to be present.)
487    
488    CM documentation has been updated to reflect the changes to library
489    organization.
490    
491    ----------------------------------------------------------------------
492    Name: Matthias Blume
493    Date: 2001/07/10 17:30:00 EDT
494    Tag: Release_110_34
495    Description:
496    
497    Minor tweak to 110.34 (re-tagged):
498    
499      - README.html file added to CVS repository
500      - runtime compiles properly under FreeBSD 3.X and 4.X
501    
502    ----------------------------------------------------------------------
503    Name: Matthias Blume
504    Date: 2001/07/10 17:30:00 EDT
505    Tag: Release_110_34
506    Description:
507    
508    New version number (110.34). New bootfiles.
509    
510    ----------------------------------------------------------------------
511    Name: Matthias Blume
512    Date: 2001/07/09 16:00:00 EDT
513    Tag: blume-20010709-more-varargs
514    Description:
515    
516    I changed the handling of varargs in ml-nlffigen again:
517    The ellipsis ... will now simply be ignored (with an accompanying warning).
518    
519    The immediate effect is that you can actually call a varargs function
520    from ML -- but you can't actually supply any arguments beyond the ones
521    specified explicitly.  (For example, you can call printf with its format
522    string, but you cannot pass additional arguments.)
523    
524    This behavior is only marginally more useful than the one before, but
525    it has the advantage that a function or, more importantly, a function
526    type never gets dropped on the floor, thus avoiding follow-up problems with
527    other types that refer to the offending one.
528    
529    ----------------------------------------------------------------------
530    Name: Matthias Blume
531    Date: 2001/07/09 11:25:00 EDT
532    Tag: blume-20010709-varargs
533    Description:
534    
535    1. ckit-lib.cm now exports structure Error
536    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
537       with a warning accompanied by a source location.  Moreover, it
538       merely skips the offending function or type and proceeds with the
539       rest of its work.u  As a result, one can safely feed C code containing
540       "..." to ml-nlffigen.
541    3. There are some internal improvements to CM, providing slightly
542       more general string substitutions in the tools subsystem.
543    
544    ----------------------------------------------------------------------
545    Name: Matthias Blume
546    Date: 2001/06/27 15:10:00 EDT
547    Tag: blume-20010627-concur
548    Description:
549    
550    Fixed a small bug in CM's handling of parallel compilation.
551    (You could observe the bug by Control-C-interrupting an ordinary
552    CMB.make or CM.stabilize and then attaching some compile servers.
553    The result was that all of a sudden the previously interrupted
554    compilation would continue on its own.  This was because of
555    an over-optimization: CM did not bother to clean out certain queues
556    when no servers were attached "anyway", resulting in the contents
557    of these queues to grab control when new servers did get attached.)
558    
559    There is also another minor update to the CM manual.
560    
561    ----------------------------------------------------------------------
562    Name: Matthias Blume
563    Date: 2001/06/26 16:15:00 EDT
564    Tag: blume-20010626-cmdoc
565    Description:
566    
567    Minor typo fixed in CM manual (syntax diagram for libraries).
568    
569    ----------------------------------------------------------------------
570    Name: Matthias Blume
571    Date: 2001/06/25 22:55:00 EDT
572    Tag: blume-20010625-x86pc
573    Description:
574    
575    Fixed a nasty bug in the X86 assembly code that caused signal
576    handlers to fail (crash) randomly.
577    
578    ----------------------------------------------------------------------
579    Name: Matthias Blume
580    Date: 2001/06/25 12:05:00 EDT
581    Tag: blume-20010625-nlffigen
582    Description:
583    
584    This update fixes a number of minor bugs in ml-nlffigen as reported by
585    Nick Carter <nbc@andrew.cmu.edu>.
586    
587      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
588      2. Default names for generated files are now derived from the name of
589         the C file *without its directory*.  In particular, this causes generated
590         files to be placed locally even if the C file is in some system directory.
591      3. Default names for generated signatures and structures are also derived
592         from the C file name without its directory.  This avoids silly things
593         like "structure GL/GL".
594         (Other silly names are still possible because ml-nlffigen does not do
595          a thorough check of whether generated names are legal ML identifiers.
596          When in doubt, use command line arguments to force particular names.)
597    
598    ----------------------------------------------------------------------
599    Name: Matthias Blume
600    Date: 2001/06/21 12:25:00 EDT
601    Tag: blume-20010621-eXene
602    Description:
603    
604    eXene now compiles and (sort of) works again.
605    
606    The library name (for version > 110.33) is $/eXene.cm.
607    
608    I also added an new example in src/eXene/examples/nbody.  See the
609    README file there for details.
610    
611    ----------------------------------------------------------------------
612    Name: Matthias Blume
613    Date: 2001/06/20 16:40:00 EDT
614    Tag: blume-20010620-cml
615    Description:
616    
617    CML now compiles and works again.
618    
619    Libraries (for version > 110.33):
620    
621      $cml/cml.cm            Main CML library.
622      $cml/basis.cm          CML's version of $/basis.cm.
623      $cml/cml-internal.cm   Internal helper library.
624      $cml/core-cml.cm       Internal helper library.
625      $cml-lib/trace-cml.cm  Tracing facility.
626      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
627    
628    The installer (config/install.sh) has been taught how to properly
629    install this stuff.
630    
631    ----------------------------------------------------------------------
632    Name: Matthias Blume
633    Date: 2001/06/19 17:55:00 EDT
634    Tag: blume-20010619-instantiate
635    Description:
636    
637    This un-breaks the fix for bug 1432.
638    (The bug was originally fixed in 110.9 but I broke it again some
639    time after that.)
640    
641    ----------------------------------------------------------------------
642    Name: Matthias Blume
643    Date: 2001/06/19 17:25:00 EDT
644    Tag: blume-20010619-signals
645    Description:
646    
647    This should (hopefully) fix the long-standing signal handling bug.
648    (The runtime system was constructing a continuation record with an
649    incorrect descriptor which would cause the GC to drop data on the floor...)
650    
651    ----------------------------------------------------------------------
652    Name: Matthias Blume
653    Date: 2001/06/15 15:05:00 EDT
654    Tag: blume-20010615-moresparc
655    Description:
656    
657    Here is a short late-hour update related to Sparc c-calls:
658    
659     -- made handling of double-word arguments a bit smarter
660    
661     -- instruction selection phase tries to collapse certain clumsily
662        constructed ML-Trees; typical example:
663    
664            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
665    
666        This currently has no further impact on SML/NJ since mlriscGen does
667        not seem to generate such patterns in the first place, and c-calls
668        (which did generate them in the beginning) has meanwhile been fixed
669        so as to avoid them as well.
670    
671    ----------------------------------------------------------------------
672    Name: Matthias Blume
673    Date: 2001/06/15 15:05:00 EDT
674    Tag: blume-20010615-sparc
675    Description:
676    
677    The purpose of this update is to provide an implementation of NLFFI
678    on Sparc machines.
679    
680    Here are the changes in detail:
681    
682       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
683       the Sparc implementation of the c-calls API.
684       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
685       for accessing the ML frame.  Thus, we have a real frame pointer and
686       can freely modify %sp without need for an omit-frame-ptr phase.
687       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
688       this case.
689       * ml-nlffigen has been taught to produce code for different architectures
690       and calling conventions.
691       * In a way similar to what was done in the x86 case, the Sparc
692       backend uses its own specific extension to mltree.  (For example,
693       it needs to be able to generate UNIMP instructions which are part
694       of the calling convention.)
695       * ml-nlffi-lib was reorganized to make it more modular (in particular,
696       to make it easier to plug in new machine- and os-dependent parts).
697    
698    There are some other fairly unrelated bug fixes and cleanups as well:
699    
700       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
701       that they properly share their libraries with existing SML/NJ libraries.
702       * I fixed a minor cosmetic bug in CM, supressing certain spurious
703       follow-up error messages.
704       * Updates to CM/CMB documentation.
705    
706    TODO items:
707    
708       * MLRISC should use a different register as its asmTemp on the Sparc.
709         (The current %o2 is a really bad choice because it is part of the
710         calling conventions, so things might interfere in unexpected ways.)
711    
712    ----------------------------------------------------------------------
713    Name: Matthias Blume
714    Date: 2001/06/07
715    Tag: blume-20010607-calls
716    Description:
717    
718    A number of internal changes related to C calls and calling conventions:
719    
720    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
721       number of bytes popped implicitly (by the callee).  In most cases
722       this field is 0 but on x86/win32 it is some non-zero value.  This
723       is information provided for the benefit of the "omit-frameptr" pass.
724    2. The CALL instruction on the x86 carries a similar "pops" field.
725       The instruction selection phase copies its value from the ML-Tree
726       CALL statement.
727    3. On all other architectures, the instruction selection phase checks
728       whether "pops=0" and complains if not.
729    4. The c-calls implementation for x86 now accepts two calling conventions:
730       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
731       up after the call and pops is set to 0.  For "stdcall", the caller
732       does nothing, leaving the cleanup to the callee; pops is set to
733       the number of bytes that were pushed onto the stack.
734    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
735       distinguish between "ccall" and "stdcall".
736    6. The UNIMP instruction has been added to the supported Sparc instruction
737       set. (This is needed for implementing the official C calling convention
738       on this architecture.)
739    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
740       work with the latest CM.
741    
742    ----------------------------------------------------------------------
743    Name: Matthias Blume
744    Date: 2001/06/05 15:10:00 EDT
745    Tag: blume-20010605-cm-index
746    Description:
747    
748    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
749    
750    1. CM can now generate "index files".  These are human-readable files
751       that list on a per-.cm-file basis each toplevel symbol defined or
752       imported.  The location of the index file for
753       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
754       To enable index-file generation, set CM.Control.generate_index to true
755       or export an environment-symbol: export CM_GENERATE_INDEX=true.
756    
757       The CM manual has been updated accordingly.
758    
759    2. I made some slight modifications to the c-calls API in MLRISC.
760    
761         a) There is now a callback to support saving/restoring of
762            dedicated but caller-save registers around the actual call
763            instruction.
764         b) One can optionally specify a comment-annotation for the
765            call instruction.
766    
767    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
768       (For example, the comment annotation shows the C prototype of
769        the function being called.)
770    
771    ----------------------------------------------------------------------
772    Name: Matthias Blume
773    Date: 2001/06/01 13:30:00 EDT
774    Tag: blume-20010601-nlffi-cleanup
775    Description:
776    
777    This is mostly a cleanup of MLFFI stuff:
778    
779       - some signature files have been put into a more exposed place
780       - the ugly 'f type parameter is gone (simplifies types tremendously!)
781       - ml-nlffigen changed accordingly
782       - tutorial updated
783    
784    Other changes:
785    
786       - author's affiliation in CM manual(s) updated
787       - some more recognized keywords added to Allen's sml.sty
788    
789    ----------------------------------------------------------------------
790    Name: Matthias Blume
791    Date: 2001/05/25 15:30:00 EDT
792    Tag: blume-20010525-iptr
793    Description:
794    
795      - put the official 110.33-README (as it appears on the ftp server) under
796        CVS
797      - fixed a small bug related to incomplete pointer types in
798        ml-nlffigen
799      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
800        (it does not need the 'f type parameter)
801    
802    ----------------------------------------------------------------------
803    Name: Matthias Blume
804    Date: 2001/05/23 14:30:00 EDT
805    Tag: Release_110_33
806    Description:
807    
808    New version number (110.33).  New bootfiles.
809    
810    ----------------------------------------------------------------------
811    Name: Matthias Blume
812    Date: 2001/05/22 18:06:00 EDT
813    Tag: blume-20010522-targets
814    Description:
815    
816    Made install.sh use file config/targets.customized if it exists, falling
817    back to config/targets if it doesn't.  This way one can have a customized
818    version of the targets file without touching the "real thing", thus
819    eliminating the constant fear of accidentally checking something bogus
820    back into the CVS repository...  (File config/targets.customized must
821    not be added to the repository!)
822    
823    ----------------------------------------------------------------------
824    Name: Matthias Blume
825    Date: 2001/05/22 16:30:00 EDT
826    Tag: blume-20010522-minitut
827    Description:
828    
829    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
830       struct returns.
831    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
832       incomplete, preliminary documentation for NLFFI.
833    
834    ----------------------------------------------------------------------
835    Name: Matthias Blume
836    Date: 2001/05/14 11:30:00 EDT
837    Tag: blume-20010514-script
838    Description:
839    
840    Some bugs in install script fixed.
841    
842    In addition to that I also made a slight change to the NLFFI API:
843    Functors generated by ml-nlffigen now take the dynamic library as a
844    straight functor argument, not as a suspended one.  (The original
845    functor code used to force the suspension right away anyway, so there
846    was nothing gained by this complication of the interface.)
847    
848    ----------------------------------------------------------------------
849    Name: Matthias Blume
850    Date: 2001/05/11 14:35:00 EDT
851    Tag: blume-20010511-ml-nlffi
852    Description:
853    
854    I finally took the plunge and added my new FFI code to the main
855    repository.  For x86-linux it is now ready for prime-time.
856    
857    There are two new subdirectories of "src":
858    
859      - ml-nlffi-lib:
860           The utility library for programs using the FFI interface.
861           Here is the implementation of $/c.cm and its associated low-level
862           partners $/c-int.cm and $/memory.cm.
863      - ml-nlffigen:
864           A stand-alone program for generating ML glue code from C source
865           code.
866    
867    Building ml-nlffigen requires $/ckit-lib.cm.
868    
869    The config/install.sh script has been updates to do the Right Thing
870    (hopefully).
871    
872    Notice that the source tree for the C-Kit will not be put under "src"
873    but directly under the installation root directory.  (This is the
874    structure that currently exists on the CVS server when you check out
875    module "sml".)  Fortunately, config/install.sh knows about this oddity.
876    
877    Bugs: No documentation yet.
878    
879    ----------------------------------------------------------------------
880    Name: Matthias Blume
881    Date: 2001/05/09 16:35:00 EDT
882    Tag: blume-20010509-cpscontract
883    Description:
884    
885    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
886    wrapper/unwrapper elimination did not decrement usage counts and some
887    dead variables got overlooked by the dead-up logic.)
888    
889    ----------------------------------------------------------------------
890    Name: Lal George
891    Date: 2001/05/08  17:26:09 EDT
892    Tag: george-20010508-omit-frameptr
893    Description:
894    
895    Changes to implement the omit-frame-pointer optimization to support
896    raw C calls. For now, there is only support on the Intel x86, but
897    other architectures will follow as more experience is gained with this.
898    
899    
900    ----------------------------------------------------------------------
901    Name: Matthias Blume
902    Date: 2001/05/07 14:40:00 EDT
903    Tag: blume-20010507-proxies
904    Description:
905    
906    I made into "proxy libraries" all libraries that qualify for such a
907    change.  (A qualifying library is a library that has another library or
908    groups as its sole member and repeats that member's export list
909    verbatim.  A proxy library avoids this repetition by omitting its export
910    list, effectively inheriting the list that its (only) member exports.
911    See the CM manual for more explanation.)
912    The main effect is that explicit export lists for these libraries
913    do not have to be kepts in sync, making maintenance a bit easier.
914    
915    I also added copyright notices to many .cm-files.
916    
917    Last but not least, I made a new set of bootfiles.
918    
919    ----------------------------------------------------------------------
920    Name: Matthias Blume
921    Date: 2001/05/04 17:00:00 EDT
922    Tag: blume-20010504-cm-lsplit
923    Description:
924    
925    0. John merged pending changes to $/smlnj-lib.cm
926    
927    1. Allen's previous change accidentally backed out of one of Lal's
928       earlier changes.  I undid this mistake (re-introducing Lal's change).
929    
930    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
931       within the compiler where applicable.  There is some code simplification
932       because of that.
933    
934    3. The "split" phase (in FLINT) is now part of the default list of phases.
935       Compiler.Control.LambdaSplitting.* can be used to globally control the
936       lambda-splitting (cross-module-inlining) engine.  In addition to that,
937       it can now also be controlled on a per-source basis: CM has been taught
938       a new tool parameter applicable to ML source files.
939    
940       - To turn lambda-splitting off completely:
941            local open Compiler.Control.LambdaSplitting in
942                val _ = set Off
943            end
944       - To make "no lambda-splitting" the global default (but allow per-source
945         overriding); this is the initial setting:
946            local open Compiler.Control.LambdaSplitting in
947                val _ = set (Default NONE)
948            end
949       - To make "lambda-splitting with aggressiveness a" the global default
950         (and allow per-source overriding):
951            local open Compiler.Control.LambdaSplitting in
952                val _ = set (Default (SOME a))
953            end
954    
955       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
956         write (in the respective .cm-file):
957            a.sml (lambdasplitting:off)
958       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
959            a.sml (lambdasplitting:on)
960       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
961         <a> is a decimal non-negative integer):
962            a.sml (lambdasplitting:<a>)
963       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
964            a.sml (lambdasplitting:infinity)
965       - To use the global default for a.sml:
966            a.sml (lambdasplitting:default)
967         or simply
968            a.sml
969    
970    ----------------------------------------------------------------------
971    Name: Allen Leung
972    Date: 2001/05/04 01:57:00 EDT
973    Tag: leunga-20010504-sync
974    Description:
975    
976      MLRISC features.
977    
978      1. Fix to CMPXCHG instructions.
979      2. Changed RA interface to allow annotations in callbacks.
980      3. Added a new method to the stream interface to allow annotations updates.
981    
982    ----------------------------------------------------------------------
983    Name: Matthias Blume
984    Date: 2001/05/01 11:45:00 EDT
985    Tag: blume-20010501-pcedittmp
986    Description:
987    
988    Changed install.sh to use the current working directory instead of
989    /usr/tmp for a temporary file (pcedittmp).  The previous choice
990    of /usr/tmp caused trouble with MacOS X because of file premission
991    problems.
992    
993    ----------------------------------------------------------------------
994    Name: Matthias Blume
995    Date: 2001/04/20 11:10:00 EDT
996    Tag: blume-20010420-inMLflag
997    Description:
998    
999     - added vp_limitPtrMask to vproc-state.h
1000       (for use by the raw-C-calls mechanism to implement proper interrupt
1001        handling)
1002     - made the ML compiler aware of various data-structure offsets so it
1003       can generate code for accessing the vp_inML flag and vp_limitPtrMask
1004     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
1005       raw C-calls
1006    
1007    ----------------------------------------------------------------------
1008    Name: Lal George
1009    Date: 2001/04/20 09:15:28 EDT
1010    Tag: george-20010420-macosX
1011    Description:
1012    
1013     - Changes to port to Mac OS X; Darwin.
1014    
1015     - In the process I found that sqrt was broken on the PPC, because the
1016       fsqrt instruction is not implemented.
1017    
1018    ----------------------------------------------------------------------
1019    Name: Matthias Blume
1020    Date: 2001/04/18 12:45:00 EDT
1021    Tag: blume-20010418-ccalls
1022    Description:
1023    
1024     - fixed two off-by-4 errors in the x86-specific c-calls implementation
1025       (this bug prevented structure arguments containing pointers from being
1026        passed correctly)
1027     - changed the raw-C-call code in mlriscGen.sml in such a way that
1028       structure arguments are represented as a pointer to the beginning
1029       of the structure (instead of having a series of synthesized arguments,
1030       one for each structure member)
1031    
1032     - made makeml script's verbosity level configurable via environment
1033       variable (MAKEML_VERBOSITY)
1034    
1035     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
1036       in rawmem-x86.sml; we are now using the real thing
1037    
1038    ----------------------------------------------------------------------
1039    Name: Matthias Blume
1040    Date: 2001/03/22 16:25:00 EST
1041    Tag: blume-20010322-bootfiles
1042    Description:
1043    
1044    Created a new set of bootfiles (for your automatic installation convenience).
1045    
1046    ----------------------------------------------------------------------
1047    Name: Matthias Blume
1048    Date: 2001/03/22 15:10:00 EST
1049    Tag: blume-20010322-rawmem-parcm
1050    Description:
1051    
1052    1. All "raw memory access" primitives for the new FFI are implemented now
1053       (at least on the x86).
1054    2. Some further cleanup of CM's parallel make mechanism.
1055    
1056    ----------------------------------------------------------------------
1057    Name: Matthias Blume
1058    Date: 2001/03/19 17:53:00 EST
1059    Tag: blume-20010319-parallel
1060    Description:
1061    
1062    Parallel make (using compile servers) now works again.
1063    
1064    To this end, CM.stabilize and CMB.make have been modified to work in
1065    two passes when compile servers are attached:
1066       1. Compile everything, do not perform stabilization; this pass
1067          uses compile servers
1068       2. Stabilize everything; this pass does not use compile servers
1069    If there are no compile servers, the two passes are combined into one
1070    (as before).  Splitting the passes increases the inherent parallelism
1071    in the dependency graph because the entire graph including all
1072    libraries is available at the same time.  This, in turn, improves
1073    server utilization.  The downside is that the master process will
1074    have to do some extra work after compilation is done (because for
1075    technical reasons it must re-read all the binfiles during stabilization).
1076    
1077    ----------------------------------------------------------------------
1078    Name: Matthias Blume
1079    Date: 2001/03/16 12:22:00 EST
1080    Tag: blume-20010316-bootfiles
1081    Description:
1082    
1083    Created a new set of bootfiles (for your automatic installation convenience).
1084    
1085    ----------------------------------------------------------------------
1086    Name: Matthias Blume
1087    Date: 2001/03/16 11:00:00 EST
1088    Tag: blume-20010316-MLTREE-fixup
1089    Description:
1090    
1091    This is a minor fixup for an (untagged) earlier commit by Allen.
1092    (A file was missing).
1093    
1094    ----------------------------------------------------------------------
1095    Name: Allen Leung
1096    Date: Mon Mar  5 18:54:57 EST 2001
1097    Tag: leunga-20010305-cut-support
1098    
1099    1. New support for alternative control-flow in MLTREE.
1100       Currently we support
1101    
1102          FLOW_TO(CALL ...., [k1,...,kn])
1103    
1104       This is needed for 'cuts to' in C-- and try/handle-like constructs
1105       in Moby
1106    
1107       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
1108    
1109    2. Register Allocator
1110    
1111       Changes in interface [from Fermin, John]
1112    
1113    3. Alpha 8-bit SLL support [Fermin]
1114    
1115    4. All architectures
1116    
1117       A new module (ClusterExpandCopies) for expanding parallel copies.
1118    
1119    ----------------------------------------------------------------------
1120    Name: Allen Leung
1121    Date: 2001/02/27 23:07:00 EST
1122    Tag: leunga-20010227-minor-stuff
1123    
1124    1. Alpha bug fix for CMOVNE
1125    2. Handle mltree COND(..,FCMP ...,...)
1126    3. Bug fix in simplifier
1127    
1128    ----------------------------------------------------------------------
1129    Name: Matthias Blume
1130    Date: 2001/01/30 17:50:00 EST
1131    Tag: blume-20010130-sync
1132    Description:
1133    
1134    This is just a minor update to sync my devel branch with the main brach.
1135    The only visible change is the addition of some README files.
1136    
1137    ----------------------------------------------------------------------
1138    Name: Matthias Blume
1139    Date: 2001/01/12 23:30:00 JST
1140    Tag: blume-20010112-bootfiles
1141    Description:
1142    
1143    Made a new set of bootfiles that goes with the current state of the
1144    repository.
1145    
1146    ----------------------------------------------------------------------
1147    Name: Matthias Blume
1148    Date: 2001/01/12 21:20:00 JST
1149    Tag: blume-20010112-sync
1150    Description:
1151    
1152    I am just flushing out some minor changes that had accumulated in
1153    my private branch in order to sync with the main tree.  (This is
1154    mainly because I had CVS trouble when trying to merge _into_ my
1155    private branch.)
1156    
1157    Most people should be completely unaffected by this.
1158    
1159    ----------------------------------------------------------------------
1160    Name: Allen Leung
1161    Date: Thu Jan 11 21:03:00 EST 2001
1162    Tag: leunga-20010111-labexp=mltree
1163    Description:
1164    
1165    1.  Removed the type LabelExp and replace it by MLTree.
1166    2.  Rewritten mltree-simplify with the pattern matcher tool.
1167    3.  There were some bugs in alpha code generator which would break
1168        64-bit code generation.
1169    4.  Redo the tools to generate code with the
1170    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
1171        are now generated by perl scripts.
1172    
1173    ----------------------------------------------------------------------
1174    Name: Matthias Blume
1175    Date: 2001/01/10 21:55:00 JST
1176    Tag: blume-20010110-rcc
1177    Description:
1178    
1179    The RCC stuff now seems to work (but only on the x86).
1180    This required hacking of the c-calls interface (and -implementation) in
1181    MLRISC.
1182    
1183    Normal compiler users should be unaffected.
1184    
1185    ----------------------------------------------------------------------
1186    Name: Matthias Blume
1187    Date: 2001/01/09 01:20:00 JST
1188    Tag: blume-20010109-rcc
1189    Description:
1190    
1191    This is a fairly big patch, flushing out a large number of pending
1192    changes that I made to my development copy over the last couple of days.
1193    
1194    Of practical relevance at this moment is a workaround for a pickling
1195    bug that Allen ran into the other day.  The cause of the bug itself is
1196    still unknown and it might be hard to fix it properly, but the
1197    workaround has some merits of its own (namely somewhat reducing pickling
1198    overhead for certain libraries).  Therefore, I think this solution should
1199    be satisfactory at this time.
1200    
1201    The rest of the changes (i.e., the vast majority) has to do with my
1202    ongoing efforts of providing direct support for C function calls from
1203    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
1204    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
1205    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
1206    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
1207    handle RCC.
1208    
1209    The last part (the changes to mlriscGen.sml) are still known to be
1210    wrong on the x86 and not implemented on all other architectures.  But
1211    the infrastructure is in place. I had to change a few functor
1212    signatures in the backend to be able to route the CCalls interface
1213    from MLRISC there, and I had to specialize the mltree type (on the
1214    x86) to include the necessary extensions. (The extensions themselves
1215    were already there and redy to go in MLRISC/x86).
1216    
1217    Everything should be very happy as soon as someone helps me with
1218    mlriscGen.sml...
1219    
1220    In any case, nothing of this should matter to anyone as long as the
1221    new primop is not being used (which is going to be the case unless you
1222    find it where I hid it :). The rest of the compiler is completely
1223    unaffected.
1224    
1225    ----------------------------------------------------------------------
1226    Name: Matthias Blume
1227    Date: 2001/01/05 00:30:00 JST
1228    Tag: blume-20010105-primops
1229    Description:
1230    
1231    Added some experimental support for work that I am doing right now.
1232    These changes mostly concern added primops, but there is also a new
1233    experimental C library in the runtime system (but currently not enabled
1234    anywhere except on Linux/X86).
1235    
1236    In the course of adding primops (and playing with them), I discovered that
1237    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
1238    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
1239    To recover, I made type information in INL_PRIM mandatory and changed
1240    prim.sml as well as built-in.sml accordingly.  The InLine structure now
1241    has complete, correct type information (i.e., no bottom types).
1242    
1243    Since all these changes mean that we need new binfiles, I also bumped the
1244    version number to 110.32.1.
1245    
1246    ----------------------------------------------------------------------
1247    Name: Matthias Blume
1248    Date: 2000/12/30 22:10:00 JST
1249    Tag: blume-20001230-various
1250    Description:
1251    
1252    Added proxy libraries for MLRISC and let MLRISC libraries refer
1253    to each other using path anchors.  (See CM manual for explanation.)
1254    
1255    Updated CM documentation.
1256    
1257    Fixed some bugs in CM.
1258    
1259    Implemented "proxy" libraries (= syntactic sugar for CM).
1260    
1261    Added "-quiet" option to makeml and changed runtime system accordingly.
1262    
1263    Added cleanup handler for exportML to reset timers and compiler stats.
1264    
1265    ----------------------------------------------------------------------
1266    Name: Lal George
1267    Date: 2000/12/22 22:22:58 EST 2000
1268    Tag: Release_110_32
1269    Description:
1270    
1271            Infinite precision used throughout MLRISC.
1272            see MLRISC/mltree/machine-int.sig
1273    
1274    ----------------------------------------------------------------------
1275    Name: Matthias Blume
1276    Date: 2000/12/22 23:16:00 JST
1277    Tag: blume-20001222-warn
1278    Description:
1279    
1280    Corrected wording and formatting of some CM warning message which I
1281    broke in my previous patch.
1282    
1283    ----------------------------------------------------------------------
1284    Name: Matthias Blume
1285    Date: 2000/12/22 21:20:00 JST
1286    Tag: blume-20001222-anchorenv
1287    Description:
1288    
1289    Fixed CM's handling of anchor environments in connection with CMB.make.
1290    
1291    ----------------------------------------------------------------------
1292    Name: Matthias Blume
1293    Date: 2000/12/22 13:15:00 JST
1294    Tag: blume-20001222-cleanup
1295    Description:
1296    
1297    Removed src/cm/ffi which does not (and did not) belong here.
1298    
1299    ----------------------------------------------------------------------
1300    Name: Matthias Blume
1301    Date: 2000/12/21 23:55:00 JST
1302    Tag: blume-20001221-exn
1303    Description:
1304    
1305    Probably most important: CM no longer silently swallows all exceptions
1306    in the compiler.
1307    Plus: some other minor CM changes.  For example, CM now reports some
1308    sizes for generated binfiles (code, data, envpickle, lambdapickle).
1309    
1310    ----------------------------------------------------------------------
1311    Name: Matthias Blume
1312    Date: 2000/12/15 00:01:05 JST
1313    Tag: blume-20001215-dirtool
1314    Description:
1315    
1316    - "dir" tool added.
1317    - improvements and cleanup to Tools structure
1318    - documentation updates
1319    
1320    ----------------------------------------------------------------------
1321    Name: Allen Leung
1322    Date: Thu Dec 14 03:45:24 EST 2000
1323    Description:
1324    Tag:  leunga-20001214-int-inf
1325    Description:
1326    
1327       In IntInf, added these standard functions, which are missing from our
1328    implementation:
1329    
1330        andb : int * int -> int
1331        xorb : int * int -> int
1332        orb  : int * int -> int
1333        notb : int -> int
1334         <<   : int * word -> int
1335        ~>>  : int * word -> int
1336    
1337       Not tested, I hope they are correct.
1338    
1339    ----------------------------------------------------------------------
1340    Name: Allen Leung
1341    Date: Fri Dec  8 19:23:26 EST 2000
1342    Description:
1343    Tag:  leunga-20001208-nowhere
1344    Description:
1345    
1346      Slight improvements to the 'nowhere' tool to handle OR-patterns,
1347    to generate better error messages etc.  Plus a brief manual.
1348    
1349    ----------------------------------------------------------------------
1350    Name: Lal George
1351    Date: 2000/12/08 09:54:02 EST 2000
1352    Tag: Release_110_31
1353    Description:
1354    
1355    - Version 110.31
1356    ----------------------------------------------------------------------
1357    Name: Allen Leung
1358    Date: Thu Dec  7 22:01:04 EST 2000
1359    Tag:  leunga-20001207-cell-monster-hack
1360    Description:
1361    
1362    Major MLRISC internal changes.  Affect all clients.
1363    Summary:
1364    
1365    1.  Type CELLS.cell = int is now replaced by a datatype.
1366        As a result, the old regmap is now gone.  Almost all interfaces
1367        in MLRISC change as a consequence.
1368    
1369    2.  A new brand version of machine description tool (v3.0) that generates
1370        modules expecting the new interface.  The old version is removed.
1371    
1372    3.  The RA interface has been further abstracted into two new functors.
1373        RISC_RA and X86RA.  These functors have much simpler interfaces.
1374        [See also directory MLRISC/demo.]
1375    
1376    4.  Some other new source->source code generation tools are available:
1377    
1378        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
1379        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
1380           I use this tool to generate the peephole optimizers---with the new
1381           cell type changes, peephole rules are becoming difficult to write
1382           without conditional pattern matching.
1383    
1384    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
1385        cover the entire MLRISC source tree so many things broke.
1386    
1387    6.  CM files have been moved to the subdirectory MLRISC/cm.
1388        They are moved because there are a lot of them and they clutter up the
1389        root dir.
1390    
1391    7.  More detailed documentation to come...
1392    
1393        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
1394        a makeml -rebuild first.  This is because of other other
1395        changes that Matthias has made (see below).
1396    
1397    
1398    ----------------------------------------------------------------------
1399    Name: Matthias Blume
1400    Date: 2000/11/30 23:12:00 JST
1401    Tag: blume-20001130-filereorg
1402    Description:
1403    
1404    Some manual updates and some file reorganizations in CM.
1405    
1406    ----------------------------------------------------------------------
1407    Name: Matthias Blume
1408    Date: 2000/11/24 17:45:00 JST
1409    Tag: blume-20001124-link
1410    Description:
1411    
1412    Drastically improved link traversal code for the case that the dynamic
1413    value was already loaded at bootstrap time.  As a result, CM and CMB
1414    now both load blazingly fast -- even on a very slow machine.  Also,
1415    memory consumption has been further reduced by this.
1416    
1417    Warning: The format of the PIDMAP file has changed.  THerefore, to
1418    bootstrap you have to do this:
1419    
1420    1. Run CMB.make
1421    2. Make a symbolic link for the boot directory:
1422         ln -s sml.boot.ARCH-OS xxx
1423    3. "Rebuild" the boot directory:
1424         ./makeml -boot xxx -rebuild sml ; rm xxx
1425    4. Boot normally:
1426          ./makeml
1427    
1428    ----------------------------------------------------------------------
1429    Name: Matthias Blume
1430    Date: 2000/11/21 21:20:00 JST
1431    Tag: blume-20001121-tools
1432    Description:
1433    
1434    Continued hacking on autoloading problem -- with success this time.
1435    Also changed tool-plugin mechanism.  See new CM manual.
1436    
1437    ----------------------------------------------------------------------
1438    Name: Matthias Blume
1439    Date: 2000/11/19 14:30:00 JST
1440    Tag:  blume-20001119-autoload
1441    Description:
1442    
1443    Some hacking to make autoloading faster.  Success for CMB, no success
1444    so far for CM.  There is a reduced structure CM' that autoloads faster.
1445    (This is a temporary, non-documented hack to be eliminated again when
1446    the general problem is solved.)
1447    
1448    ----------------------------------------------------------------------
1449    Name: Matthias Blume
1450    Date: 2000/11/17 14:10:00 JST
1451    Tag: blume-20001117-pickle-lib
1452    Description:
1453    
1454    1. Eliminated comp-lib.cm
1455    2. Made pickle-lib.cm
1456    3. Eliminated all uses of intset.sml (from comp-lib.cm)
1457    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
1458       equivalent constructs from smlnj-lib.cm (INtHashTable).
1459    5. Point 4. also goes for those uses of intmap.* in MLRISC.
1460       Duplicated intmap modules thrown out.
1461    6. Hunted down all duplicated SCC code and replaced it with
1462       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
1463    7. Rewrote Feedback module.
1464    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
1465       should be thrown out and equivalent modules from smlnj-lib.cm
1466       should be used (IntRedBlackSet, IntListSet, ...).
1467    
1468    Confirmed that compiler compiles to fixpoint.
1469    
1470    ----------------------------------------------------------------------
1471    Name: Allen Leung
1472    Date: 2000/11/10 18:00:00
1473    Tag: leunga-20001110-new-x86-fp
1474    
1475    A new x86 floating point code generator has been added.
1476    By default this is turned off.  To turn this on, do:
1477    
1478        CM.autoload "$smlnj/compiler.cm";
1479        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
1480    
1481    Changes:
1482    
1483    1.  Changed FTAN to FPTAN so that the assembly output is correct.
1484    2.  Changed the extension callback for FTANGENT to generate:
1485    
1486              fptan
1487              fstp  %st(0)
1488        instead of
1489              fptan
1490              fstpl ftempmem
1491    
1492    3.  Numerous assembly fixes for x86.
1493    
1494    5.  Cleaned up the machine code output module x86/x86MC.sml and added
1495        support for a whole bunch of instructions and addressing modes:
1496    
1497          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
1498          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
1499          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
1500          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
1501          fxch %st(n)
1502          fld %st(n)
1503          fst %st(n)
1504          fst mem
1505          fstp %st(n)
1506          fucom %st(n)
1507          fucomp %st(n)
1508    
1509        All these are now generated when the fast fp mode is turned on.
1510    
1511    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
1512    
1513    ----------------------------------------------------------------------
1514    Name: Matthias Blume
1515    Date: 2000/11/09 11:20:00 JST
1516    Tag: blume-20001109-scc
1517    Description:
1518    
1519    Eliminated some code duplication:
1520    
1521    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
1522       (Otherwise the functor is useless.)
1523    2. Used GraphSCCFn where SCCUtilFun was used previously.
1524    3. Got rid of SCCUtilFun (in comp-lib.cm).
1525    
1526    ----------------------------------------------------------------------
1527    Name: Lal George
1528    Date: 2000/11/06 09:02:21 EST 2000
1529    Tag: Release_110_30
1530    Description:
1531    
1532    - Version 110.30
1533    ----------------------------------------------------------------------
1534    Name: Matthias Blume
1535    Date: 2000/11/04 14:45:00
1536    Tag: blume-20001104-mlbuild
1537    Description:
1538    
1539    - Made ml-build faster on startup.
1540    - Documentation fixes.
1541    
1542    ----------------------------------------------------------------------
1543    Name: Matthias Blume
1544    Date: 2000/11/02 17:00:00 JST
1545    Tag: blume-20001102-condcomp
1546    Description:
1547    
1548    - Small tweaks to pickler -- new BOOTFILES!
1549    - Version bumped to 110.29.2.
1550    - Added conditional compilation facility to init.cmi (see comment there).
1551    ----------------------------------------------------------------------
1552    Name: Allen Leung
1553    Date: 2000/10/23 19:31:00
1554    Tag: leunga-20001023-demo-ra
1555    
1556    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
1557    2. Test programs for the graph library updated
1558    3. Some new MLRISC demo programs added
1559    
1560    ----------------------------------------------------------------------
1561    Name: Matthias Blume
1562    Date: 2000/08/31 22:15:00 JST
1563    Tag: blume-20001017-errmsg
1564    Description:
1565    
1566    More error message grief: Where there used to be no messages, there
1567    now were some that had bogus error regions.  Fixed.
1568    
1569    ----------------------------------------------------------------------
1570    Name: Matthias Blume
1571    Date: 2000/08/31 17:30:00 JST
1572    Tag: blume-20001017-v110p29p1
1573    Description:
1574    
1575    I made a version 110.29.1 with new bootfiles.
1576    
1577    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
1578              CM documentation changes and a small bugfix in CM's error reporting.
1579    
1580    ----------------------------------------------------------------------
1581    Name: Lal George
1582    Date: 2000/09/27 14:42:35 EDT
1583    Tag: george-20000927-nodestatus
1584    Description:
1585    
1586    Changed the type of the nodestatus, so that:
1587    
1588            SPILLED(~1)             is now SPILLED
1589            SPILLED(m) where m>=0   is now MEMREG(m)
1590            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
1591    
1592    ----------------------------------------------------------------------
1593    Name: Matthias Blume
1594    Date: 2000/09/07 14:45:00 JST
1595    Tag: blume-20000907-cmerrmsg
1596    Description:
1597    
1598    Small tweak to CM to avoid getting ML syntax error messages twice.
1599    
1600    ----------------------------------------------------------------------
1601    Name: Matthias Blume
1602    Date: 2000/08/31 18:00:00 JST
1603    Tag: blume-20000831-cvsbootfiles
1604    Description:
1605    
1606    New URL for boot files (because the 110.29 files on the BL server do
1607    now work correctly with my updated install scripts for yacc and lex).
1608    
1609    ----------------------------------------------------------------------
1610    Name: Matthias Blume
1611    Date: 2000/08/08 12:33:00 JST
1612    Tag: blume-20000808-manual
1613    Description:
1614    
1615    Tiny update to CM manual.
1616    
1617    ----------------------------------------------------------------------
1618    Name: Allen Leung
1619    Date: 2000/08/7 19:31:00
1620    Tag: leunga-20000807-a-whole-bunch-of-stuff
1621    
1622      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
1623    and Mobdy.
1624    
1625    1.  x86
1626    
1627       a.  Fixes to peephole module by John and Dan.
1628       b.  Assembly fix to SETcc by Allen.
1629       c.  Fix to c-call by John.
1630       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
1631       e.  Instruction selection optimization to SETcc as suggested by John.
1632    
1633           For example,
1634    
1635            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
1636    
1637           should generate:
1638    
1639            MOVL a, x
1640            SUBL b, x
1641            SHRL 31, x
1642    
1643    2.  IR stuff
1644    
1645         A bunch of new DJ-graph related algorithms added.  These
1646         speed up SSA construction.
1647    
1648    3.  SSA + Scheduling
1649    
1650         Added code for SSA and scheduling to the repository
1651    
1652    ----------------------------------------------------------------------
1653    Name: Lal George
1654    Date: 2000/07/27 11:53:14 EDT
1655    
1656    Tag: lal-20000727-linux-ppc
1657    Description:
1658    
1659     Made changes to support Linux PPC.
1660     p.s. I have confirmation that the 110.29 boot files work fine.
1661    
1662    ----------------------------------------------------------------------
1663    Name: Matthias Blume
1664    Date: 2000/07/27 17:40:00 JST
1665    Tag: blume-20000727-scripts
1666    Description:
1667    
1668    !!!! WARNING !!!!
1669    You must recompile the runtime system!
1670    !!!! WARNING !!!!
1671    
1672    This is basically another round of script-enhancements:
1673    
1674    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
1675       and undefine CM preprocessor symbols.
1676    
1677    2. ml-build avoids generating a new heap image if it finds that the
1678       existing one is still ok.  (The condition is that no ML file had to
1679       be recompiled and all ML files are found to be older that the heap
1680       file.)
1681    
1682       To make this work smoothly, I also hacked the runtime system as
1683       well as SMLofNJ.SysInfo to get access to the heap image suffix
1684       (.sparc-solaris, ...) that is currently being used.
1685    
1686       Moreover, the signature of CM.mk_standalone has changed.  See the
1687       CM manual.
1688    
1689    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
1690       CM manual for details.)
1691    
1692    4. More CM manual updates:
1693        - all of the above has been documented.
1694        - there is now a section describing the (CM-related) command line
1695          arguments that are accepted by the "sml" command
1696    
1697    ----------------------------------------------------------------------
1698    Name: Matthias Blume
1699    Date: 2000/07/25 16:20:00 JST
1700    Tag: blume-20000725-makedepend
1701    Description:
1702    
1703    Added a script called ml-makedepend.  This can be used in makefiles
1704    for Unix' make in a way very similar to the "makedepend" command for
1705    C.
1706    
1707    The script internally uses function CM.sources.
1708    
1709    Synopsis:
1710    
1711        ml-makedepend [-f makefile] cmfile targetname
1712    
1713    The default for the makefile is "makefile" (or "Makefile" should
1714    "makefile" not exist).
1715    
1716    ml-makedepend adds a cmfile/targetname-specific section to this
1717    makefile (after removing the previous version of this section).  The
1718    section contains a single dependency specification with targetname on
1719    the LHS (targetname is an arbitrary name), and a list of files derived
1720    from the cmfile on the RHS.  Some of the files on the RHS are
1721    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
1722    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
1723    path names.  The makefile writer is responsible for making sure that
1724    these variables have correct at the time "make" is invoked.
1725    
1726    ----------------------------------------------------------------------
1727    Name: Matthias Blume
1728    Date: 2000/07/22 23:30:00 JST
1729    Tag: blume-20000722-urlupdate
1730    Description:
1731    
1732    Changed BOOT and config/srcarchiveurl to point to BL server:
1733    
1734        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
1735    
1736    ----------------------------------------------------------------------
1737    Name: Matthias Blume
1738    Date: 2000/07/18 18:00:00 JST
1739    Tag: blume-20000718-Version_110_29
1740    Description:
1741    
1742    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
1743    
1744    2. Updated config/version to 110.29
1745    
1746    3. Updated config/srcarchiveurl
1747    
1748    3. New boot files!
1749       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
1750    
1751    ----------------------------------------------------------------------
1752    Name: Matthias Blume
1753    Date: 2000/07/11 13:58:00 JST
1754    Tag: blume-20000711-doctypo
1755    Description:
1756    
1757    Fixed a few typos in CM manual.
1758    
1759    ----------------------------------------------------------------------
1760    Name: Allen Leung
1761    Date: 2000/06/15 00:38:00
1762    Tag: leunga-20000704-sparc-x86
1763    
1764    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
1765    2. fix to x86 RET bug [found by Dan Grossman]
1766    3. sparc assembly bug fix for ticc instructions [found by Fermin]
1767    
1768       Affects c-- and moby only
1769    
1770    ----------------------------------------------------------------------
1771    Name: Matthias Blume
1772    Date: 2000/07/04 15:26:00
1773    Tag: blume-20000704-trigger
1774    Description:
1775    
1776    1. Improvements to CM manual.
1777    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
1778       of getting a back-trace.  The function, when called, raises an
1779       internal exception which explicitly carries the full back-trace history,
1780       so it is unaffected by any intervening handle-raise pairs ("trivial"
1781       or not).  The interactive loop will print that history once it arrives
1782       at top level.
1783       Short of having all exceptions implicitly carry the full history, the
1784       recommended way of using this facility is:
1785         - compile your program with instrumentation "on"
1786         - run it, when it raises an exception, look at the history
1787         - if the history is "cut off" because of some handler, go and modify
1788           your program so that it explicitly calls BTrace.trigger
1789         - recompile (still instrumented), and rerun; look at the full history
1790    
1791    ----------------------------------------------------------------------
1792    Name: Matthias Blume
1793    Date: 2000/07/03 15:36:00 JST
1794    Tag: blume-20000702-manual
1795    Description:
1796    
1797    Small corrections and updates to CM manual.
1798    
1799    ----------------------------------------------------------------------
1800    Name: Matthias Blume
1801    Date: 2000/06/29 16:04:00 JST
1802    Tag: blume-20000629-yacctool
1803    Description:
1804    
1805    Changes:
1806    
1807    1. Class "mlyacc" now takes separate arguments to pass options to
1808       generated .sml- and .sig-files independently.
1809    2. Corresponding CM manual updates.
1810    3. BTrace module now also reports call sites.  (However, for loop clusters
1811       it only shows from where the cluster was entered.)  There are associated
1812       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1813    
1814    ----------------------------------------------------------------------
1815    Name: Matthias Blume
1816    Date: 2000/06/27 16:51:00 JST
1817    Tag: blume-20000627-noweb
1818    Description:
1819    
1820    Changes:
1821    
1822     1. Implemented "subdir" and "witness" options for noweb tool.
1823        This caused some slight internal changes in CM's tool implementation.
1824     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
1825        some remaining issues from earlier path anchor changes.
1826     3. Updated CM manual accordingly.
1827    
1828     4. Changed implementation of back-tracing so that I now consider it
1829        ready for prime-time.
1830    
1831        In particular, you don't have to explicitly trigger the back-trace
1832        anymore.  Instead, if you are running BTrace-instrumented code and
1833        there is an uncaught exception (regardless of whether or not it was
1834        raised in instrumented code), the top-level evalloop will print
1835        the back-trace.
1836    
1837        Features:
1838    
1839          - Instrumented and uninstrumented code work together seemlessly.
1840            (Of course, uninstrumented code is never mentioned in actual
1841             back-traces.)
1842    
1843          - Asymptotic time- and space-complexity of instrumented code is
1844            equal to that of uninstrumented code.  (This means that
1845            tail-recursion is preserved by the instrumentation phase.)
1846    
1847          - Modules whose code has been instrumented in different sessions
1848            work together without problem.
1849    
1850          - There is no penalty whatsoever on uninstrumented code.
1851    
1852          - There is no penalty on "raise" expressions, even in
1853            instrumented code.
1854    
1855        A potential bug (or perhaps it is a feature, too):
1856    
1857          A back-trace reaches no further than the outermost instrumented
1858          non-trivial "raise".  Here, a "trivial" raise is one that is the
1859          sole RHS of a "handle" rule.  Thus, back-traces reach trough
1860    
1861               <exp> handle e => raise e
1862    
1863          and even
1864    
1865               <exp> handle Foo => raise Bar
1866    
1867          and, of course, through
1868    
1869               <exp> handle Foo => ...
1870    
1871         if the exception was not Foo.
1872    
1873         Back-traces always reach right through any un-instrumented code
1874         including any of its "handle" expressions, trivial or not.
1875    
1876       To try this out, do the following:
1877    
1878         - Erase all existing binfiles for your program.
1879           (You may keep binfiles for those modules where you think you
1880            definitely don't need back-tracing.)
1881         - Turn on back-trace instrumentation:
1882              SMLofNJ.Internals.BTrace.mode (SOME true);
1883         - Recompile your program.  (I.e., run "CM.make" or "use".)
1884         - You may now turn instrumentation off again (if you want):
1885              SMLofNJ.Internals.BTrace.mode (SOME false);
1886         - Run your program as usual.  If it raises an exception that
1887           reaches the interactive toplevel, then a back-trace will
1888           automatically be printed.  After that, the toplevel loop
1889           will print the exception history as usual.
1890    
1891    ----------------------------------------------------------------------
1892    Name: Matthias Blume
1893    Date: 2000/06/26 09:56:46 JST
1894    Tag: blume-20000626-setup
1895    Description:
1896    
1897    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1898          ML code before and after compiling a file (e.g., to set compiler
1899          flags)
1900    
1901    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1902              - associated changes to btrace.sml (BTrace instrumentation pass)
1903              - cleaner implementation of btimp.sml (BTrace tracing and report
1904                module)
1905    
1906    CM manual: * new path encoding documented
1907               * description of setup-parameter to "sml" added
1908    
1909    The biggest user-visible change to back-tracing is that it is no
1910    longer necessary to compile all traced modules within the same
1911    session.  (This was a real limitation.)
1912    
1913    ----------------------------------------------------------------------
1914    Name: Matthias Blume
1915    Date: 2000/06/24 12:40:00 JST
1916    Tag: blume-20000624-startup
1917    Description:
1918    
1919    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1920    bit too often -- to put it mildly. :)
1921    
1922    ----------------------------------------------------------------------
1923    Name: Matthias Blume
1924    Date: 2000/06/23 18:20:00 JST
1925    Tag: blume-20000623-btrace
1926    Description:
1927    
1928    This updates adds a backtrace facility to aid programmers in debugging
1929    their programs.  This involves the following changes:
1930    
1931    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1932       keeping track of the current call stack.  When programs are compiled
1933       in a special mode, the compiler will insert calls to these hooks
1934       into the user program.
1935       "Hook" means that it is possible for different implementations of
1936       back-tracing to register themselves (at different times).
1937    
1938    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1939       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1940       It can be turned on using this call:
1941         SMLofNJ.Internals.BTrace.mode (SOME true);
1942       Turning it off again:
1943         SMLofNJ.Internals.BTrace.mode (SOME false);
1944       Querying the current status:
1945         SMLofNJ.Internals.BTrace.mode NONE;
1946       Annotated programs are about twice as big as normal ones, and they
1947       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1948       where all hooks do nothing).  The slowdown with a plugin that is
1949       actually useful (such as the one supplied by default) is even greater,
1950       but in the case of the default plugin it is still only an constant
1951       factor (amortized).
1952    
1953    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1954       with a sub-structure BTrace for controlling back-tracing.  In particular,
1955       the above-mentioned function "mode" controls whether the annotation
1956       phase is invoked by the compiler.  Another important function is
1957       "trigger": when called it aborts the current execution and causes
1958       the top-level loop to print a full back-trace.
1959    
1960    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1961       for back-tracing.  It keeps track of the dynamic call stack and in
1962       addition to that it keeps a partial history at each "level" of that
1963       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1964       d and b (at separate times, dynamically), then the report will show:
1965    
1966       GOTO   d
1967             /c
1968       GOTO  \b
1969       CALL   a
1970    
1971       This shows that there was an initial non-tail call of a, then a
1972       tail-call to b or c, looping behavior in a cluster of functions that
1973       consist of b and c, and then a goto from that cluster (i.e., either from
1974       b or from c) to d.
1975    
1976       Note that (depending on the user program) the amount of information
1977       that the back-trace module has to keep track of at each level is bounded
1978       by a constant.  Thus, the whole implementation has the same asymptotical
1979       complexity as the original program (both in space and in time).
1980    
1981    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1982       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1983       which is raised by the "trigger" function mentioned above.
1984    
1985    Notes on usage:
1986    
1987    - Annotated code works well together with unannotated code:
1988    Unannotated calls simply do not show up at all in the backtrace.
1989    
1990    - It is not a good idea to let modules that were annotated during
1991    different sessions run at the same time.  This is because the compiler
1992    chooses small integers to identify individual functions, and there
1993    will be clashes if different modules were compiled in separate sessions.
1994    (Nothing will crash, and you will even be told about the clashes, but
1995    back-trace information will in general not be useful.)
1996    
1997    - Back-tracing can be confused by callcc and capture.
1998    
1999    - The only way of getting a back-trace right now is to explicitly
2000    invoke the "trigger" function from your user program.  Eventually, we
2001    should make every exception carry back-trace information (if
2002    available).  But since this creates more overhead at "raise"-time
2003    (similar to the current exnHistory overhead), I have not yet
2004    implemented this.  (The implementation will be rather easy.)  With
2005    exceptions carrying back-trace information, this facility will be even
2006    more useful because users don't need to modify their programs...
2007    
2008    - While it is possible to compile the compiler with back-trace
2009    annotations turned on (I did it to get some confidence in
2010    correctness), you must make absolutely sure that core.sml and
2011    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
2012    be compiled with annotation because there is no core access yet, but
2013    if you compile btimp.sml with annotation, then the system will go into
2014    an infinite recursion and crash.)
2015    Since CM currently does not know about BTrace, the only way to turn
2016    annotations on and off for different modules of the compiler is to
2017    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
2018    this is awkward and clumsy.
2019    
2020    Sample sessions:
2021    
2022    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
2023    - SMLofNJ.Internals.BTrace.mode (SOME true);
2024    [autoloading]
2025    [autoloading done]
2026    val it = false : bool
2027    - structure X = struct
2028    -     fun main n = let
2029    -         fun a (x, 0) = d x
2030    -           | a (x, n) = b (x, n - 1)
2031    -         and b (x, n) = c (x, n)
2032    -         and c (x, n) = a (x, n)
2033    -         and d x = e (x, 3)
2034    -         and e (x, 0) = f x
2035    -           | e (x, n) = e (x, n - 1)
2036    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
2037    -           | f n = n * g (n - 1)
2038    -         and g n = a (n, 3)
2039    -     in
2040    -         f n
2041    -     end
2042    - end;
2043    structure X : sig val main : int -> int end
2044    - X.main 3;
2045    *** BACK-TRACE ***
2046    GOTO   stdIn:4.2-13.20: X.main[2].f
2047    GOTO-( stdIn:4.2-13.20: X.main[2].e
2048    GOTO   stdIn:4.2-13.20: X.main[2].d
2049         / stdIn:4.2-13.20: X.main[2].a
2050         | stdIn:4.2-13.20: X.main[2].b
2051    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2052    CALL   stdIn:4.2-13.20: X.main[2].g
2053    GOTO   stdIn:4.2-13.20: X.main[2].f
2054    GOTO-( stdIn:4.2-13.20: X.main[2].e
2055    GOTO   stdIn:4.2-13.20: X.main[2].d
2056         / stdIn:4.2-13.20: X.main[2].a
2057         | stdIn:4.2-13.20: X.main[2].b
2058    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2059    CALL   stdIn:4.2-13.20: X.main[2].g
2060    GOTO   stdIn:4.2-13.20: X.main[2].f
2061    GOTO-( stdIn:4.2-13.20: X.main[2].e
2062    GOTO   stdIn:4.2-13.20: X.main[2].d
2063         / stdIn:4.2-13.20: X.main[2].a
2064         | stdIn:4.2-13.20: X.main[2].b
2065    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2066    CALL   stdIn:4.2-13.20: X.main[2].g
2067    GOTO   stdIn:4.2-13.20: X.main[2].f
2068    CALL   stdIn:2.15-17.4: X.main[2]
2069    -
2070    
2071    (Note that because of a FLINt bug the above code currently does not
2072    compile without BTrace turned on.)
2073    
2074    Here is another example, using my modified Tiger compiler:
2075    
2076    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
2077    - SMLofNJ.Internals.BTrace.mode (SOME true);
2078    [autoloading]
2079    [autoloading done]
2080    val it = false : bool
2081    - CM.make "sources.cm";
2082    [autoloading]
2083    ...
2084    [autoloading done]
2085    [scanning sources.cm]
2086    [parsing (sources.cm):parse.sml]
2087    [creating directory CM/SKEL ...]
2088    [parsing (sources.cm):tiger.lex.sml]
2089    ...
2090    [wrote CM/sparc-unix/semant.sml]
2091    [compiling (sources.cm):main.sml]
2092    [wrote CM/sparc-unix/main.sml]
2093    [New bindings added.]
2094    val it = true : bool
2095    - Main.compile ("../testcases/merge.tig", "foo.out");
2096    *** BACK-TRACE ***
2097    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
2098    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2099    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2100    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2101    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2102    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2103    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
2104         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
2105    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
2106    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2107    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
2108    CALL   main.sml:1.18-118.4: Main.compile[2]
2109    -
2110    
2111    ----------------------------------------------------------------------
2112    Name: Matthias Blumen
2113    Date: 2000/06/21 18:00:00 JST
2114    Tag: blume-20000621-manual
2115    Description:
2116    
2117    CM manual update: Path environments documented.
2118    
2119    ----------------------------------------------------------------------
2120    Name: Matthias Blume
2121    Date: 2000/06/19 13:40:00
2122    Tag: blume-20000619-manual
2123    Description:
2124    
2125    CM manual and system/README update.  This only covers the fact that
2126    there are no more implicit anchors.  (Path environments and the "bind"
2127    option to "cm" have yet to be documented.)
2128    
2129    ----------------------------------------------------------------------
2130    Name: Matthias Blume
2131    Date: 2000/06/19 11:05:00 JST
2132    Tag: blume-20000619-chdir-bugfix
2133    Description:
2134    
2135    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
2136    
2137    ----------------------------------------------------------------------
2138    Name: Matthias Blume
2139    Date: 2000/06/18 22:00:10 JST
2140    Tag: blume-20000618-implicit-anchors-really-gone
2141    Description:
2142    
2143    I updates the previous HISTORY entry where I forgot to mention that
2144    implicit anchors are no longer with us.
2145    
2146    The current update also gets rid of the (now useless) controller
2147    CM.Control.implicit_anchors.
2148    
2149    ----------------------------------------------------------------------
2150    Name: Matthias Blume
2151    Date: 2000/06/16 17:30:00 JST
2152    Tag: blume-20000616-anchorenv
2153    Description:
2154    
2155    This patch implements the long anticipated (just kidding :) "anchor
2156    environment" mechanism.  In the course of doing this, I also
2157    re-implemented CM's internal "SrcPath" module from scratch.  The new
2158    one should be more robust in certain boundary cases.  In any case, it
2159    is a lot cleaner than its predecessor (IMHO).
2160    
2161    This time, although there is yet another boot file format change, I
2162    kept the unpickler backward-compatible.  As a result, no new bootfiles
2163    are necessary and bootstrapping is straightforward.  (You cannot read
2164    new bootfiles into an old system, but the other way around is no
2165    problem.)
2166    
2167    Visible changes:
2168    
2169    ** 0. Implicit path anchors (without the leading $-symbol) are no
2170    longer recognized at all. This means that such path names are not
2171    illegal either.  For example, the name basis.cm simply refers to a
2172    local file called "basis.cm" (i.e, the name is an ordinary path
2173    relative to .cm-files directory).  Or, to put it differently, only
2174    names that start with $ are anchored paths.
2175    
2176    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
2177    vanished.
2178    
2179    John (Reppy) had critizised this as soon as I originally proposed and
2180    implemented it, but at that time I did not really deeply believe
2181    him. :) Now I came full-circle because I need the $<singlearc> syntax
2182    in another place where it cannot be seen as an abbreviation for
2183    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
2184    seems to mean (i.e., it "expands" into the corresponding anchor
2185    value).
2186    
2187    However, when paths are used as members in CM description files, it
2188    continues to be true that there must be at least another arc after the
2189    anchor.  This is now enforced separately during semantic analysis
2190    (i.e., from a lexical/syntactical point of view, the notation is ok.)
2191    
2192    ** 2. The "cm" class now accepts an option "bind".  The option's value
2193    is a sub-option list of precisely two items -- one labeled "anchor"
2194    and the other one labeled "value".  As you might expect, "anchor" is
2195    used to specify an anchor name to be bound, and "value" specifies what
2196    the anchor is being bound to.
2197    
2198    The value must be a directory name and can be given in either standard
2199    syntax (including the possibility that it is itself an anchored path)
2200    or native syntax.
2201    
2202    Examples:
2203    
2204       foo.cm (bind:(anchor:bar value:$mystuff/bar))
2205       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
2206    
2207    and so on.
2208    
2209    The meaning of this is that the .cm-file will be processed with an
2210    augmented anchor environment where the given anchor(s) is/are bound to
2211    the given values(s).
2212    
2213    The rationale for having this feature is this: Suppose you are trying
2214    to use two different (already stable) libraries a.cm and b.cm (that
2215    you perhaps didn't write yourself).  Further, suppose each of these
2216    two libraries internally uses its own auxiliary library $aux/lib.cm.
2217    Normally you would now have a problem because the anchor "lib" can not
2218    be bound to more than one value globally.  Therefore, the project that
2219    uses both a.cm and b.cm must locally redirect the anchor to some other
2220    place:
2221    
2222       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
2223       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
2224    
2225    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
2226    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
2227    
2228    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
2229    you when you do so at the time of CM.stabilize).  Therefore, you can
2230    also use an anchored path as the value:
2231    
2232      a.cm (bind:(anchor:lib value:$a-lib))
2233      b.cm (bind:(anchor:lib value:$b-lib))
2234    
2235    Now you can globally configure (using the usual CM.Anchor.anchor or
2236    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
2237    itself is always locally bound, setting it globally is no longer
2238    meaningful or necessary (but it does not hurt either).  In fact, "lib"
2239    can still be used as a global anchor for separate purposes.  As a
2240    matter of fact, one can locally define "lib" in terms of a global
2241    "lib":
2242    
2243      a.cm (bind:(anchor:lib value:$lib/a))
2244      b.cm (bind:(anchor:lib value:$lib/b))
2245    
2246    ** 3: The encoding of path names has changed.  This affects the way
2247    path names are shown in CM's progress report and also the internal
2248    protocol encoding used for parallel make.
2249    
2250    The encoding now uses one or more ':'-separated segments.  Each
2251    segments corresponds to a file that has been specified relative to the
2252    file given by its preceding segment.  The first segment is either
2253    relative to the CWD, absolute, or anchored.  Each segment itself is
2254    basically a Unix pathname; all segments but the first are relative.
2255    
2256    Example:
2257    
2258       $foo/bar/baz.cm:a/b/c.sml
2259    
2260    This path denotes the file bar/a/b/c.sml relative to the directory
2261    denoted by anchor "foo".  Notice that the encoding also includes
2262    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
2263    paths are resolved relative to the .cm-files directory, so baz.cm must
2264    be ignored to get the "real" pathname.
2265    
2266    To make this fact more obvious, CM puts the names of such "virtual
2267    arcs" into parentheses when they appear in progress reports. (No
2268    parentheses will appear in the internal protocol encoding.)  Thus,
2269    what you really see is:
2270    
2271      $foo/bar/(baz.cm):a/b/c.sml
2272    
2273    I find this notation to be much more informative than before.
2274    
2275    Another new feature of the encoding is that special characters
2276    including parentheses, colons, (back)slashes, and white space are
2277    written as \ddd (where ddd is the decimal encoding of the character).
2278    
2279    *** The CM manual still needs to be updated.
2280    
2281    ----------------------------------------------------------------------
2282    Name: Allen Leung
2283    Date: 2000/06/15 00:38:00
2284    Tag: leunga-20000615-x86-peephole
2285    
2286    x86 Peephole fix by Fermin.  Affects c-- and moby only.
2287    
2288    ----------------------------------------------------------------------
2289    Name: Matthias Blume
2290    Date: 2000/06/12 11:40:00
2291    Tag: blume-20000612-parmakefix
2292    Description:
2293    
2294    More cleanup after changing the file naming scheme: This time I
2295    repaired the parallel make mechanism for CMB.make which I broke earlier.
2296    
2297    ----------------------------------------------------------------------
2298    Name: Allen Leung
2299    Date: 2000/06/09 01:25:00
2300    Tag: leunga-20000609-various
2301    
2302    None of these things should affect normal SML/NJ operations
2303    
2304    1. Peephole improvements provided by Fermin (c--)
2305    2. New annotation DEFUSE for adding extra dependence (moby)
2306    3. New X86 LOCK instructions (moby)
2307    4. New machine description language for reservation tables (scheduling)
2308    5. Fixes to various optimization/analysis modules (branch chaining, dominator
2309       trees etc.)
2310    6. I've changed the CM files so that they can work with versions
2311       110.0.6, 110.25 and 110.28
2312    
2313    ----------------------------------------------------------------------
2314    Name: Matthias Blume
2315    Date: 2000/06/09 12:40:00
2316    Tag: blume-20000609-log
2317    Description:
2318    
2319    - Removed all(?) remaining RCS Log entries from sources.
2320    
2321    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
2322      anchored paths).
2323    
2324    ----------------------------------------------------------------------
2325    Name: Matthias Blume
2326    Date: 2000/06/07 17:00:00 JST
2327    Tag: blume-20000607-no-implicit-anchors
2328    Description:
2329    
2330    1. This update changes the default setting for
2331    CM.Control.implicit_anchors from true to false.  This means that
2332    implicit anchors are no longer permitted by default.  I also tried to
2333    make sure that nothing else still relies on implicit anchors.
2334    (This is the next step on the schedule towards a CM that does not even
2335    have the notion of implicit anchors anymore.)
2336    
2337    2. More CM manual updates.
2338    
2339    3. I managed to track down and fix the pickling bug I mentioned last
2340    time.  Because of the previously existing workaround, this entails no
2341    immediate practical changes.
2342    
2343    ----------------------------------------------------------------------
2344    Name: Matthias Blume
2345    Date: 2000/06/06 11:15:00 JST
2346    Tag: blume-20000606-lazierpickle
2347    Description:
2348    
2349    !!!! NEW BOOT FILES !!!!
2350    
2351    * The main purpose of this update is to make library pickles lazier in
2352    order to reduce the initial space penalty for autoloading a library.
2353    As a result, it is now possible to have $smlnj/compiler.cm
2354    pre-registered.  This should take care of the many complaints or
2355    inquiries about missing structure Compiler.  This required changes to
2356    CM's internal data structures and small tweaks to some algorithms.
2357    
2358    As a neat additional effect, it is no longer necessary (for the sake
2359    of lean heap image files) to distinguish between a "minimal" CM and a
2360    "full" CM.  Now, there is only one CM (i.e., the "full" version:
2361    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
2362    interactive top level. ($smlnj/cm/minimal.cm is gone.)
2363    
2364    To make the life of compiler-hackers easier, "makeml" now also
2365    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
2366    words, after you bootstrap a new sml for the first time, you will not
2367    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
2368    around you will still have to do it, though.)
2369    
2370    * A second change consists of major updates to the CM manual.  There
2371    are now several appendices with summary information and also a full
2372    specification of the CM description file syntax.
2373    
2374    * In directory src/system I added the script "allcross".  This script
2375    invokes sml and cross-compiles the compiler for all supported
2376    architectures.  (Useful when providing a new set of boot files.)
2377    
2378    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
2379    added a small tweak to pickle-util.sml to work around this problem,
2380    but it is not a proper fix yet.  I will investigate further.  (The
2381    effect of the bug was an inflation of library pickle size.)
2382    
2383    * Version number increased to 110.28.1 (to avoid compatibility problems).
2384    
2385    ----------------------------------------------------------------------
2386    Name: Allen Leung
2387    Date: 2000/05/25 17:28 EDT
2388    Tag: leunga-20000525-ra
2389    Description:
2390    
2391      Fixed a bug in freezing phase of the register allocator.
2392    
2393    ----------------------------------------------------------------------
2394    Name: Allen Leung
2395    Date: 2000/05/15 22:53 EDT
2396    Tag: leunga-20000515-alpha-x86-ra
2397    Description:
2398    
2399      1. Alpha
2400    
2401          Slight cleanup.  Removed the instruction SGNXL
2402    
2403      2. X86
2404    
2405          Added the following instructions to the instruction set:
2406    
2407            ROLx, RORx,
2408            BTx, BTSx, BTLx, BTRx,
2409            XCHGx, and variants with the LOCK prefix
2410    
2411      3. Register Allocation
2412    
2413          The module ra-rewrite-with-renaming has been improved.
2414    
2415      These have no effect on SML/NJ.
2416    
2417    ----------------------------------------------------------------------
2418    Name: Matthias Blume
2419    Date: 2000/05/15 16:20:00 JST
2420    Tag: blume-20000515-lightrebuild
2421    Description:
2422    
2423    1. I added an alternative to "-rebuild" to "makeml".  The difference is
2424       that prior to calling CMB.make' the CM-variable "LIGHT" will be
2425       defined.  In effect, the command will not build any cross-compiler
2426       backends and therefore finish more quickly.
2427    
2428       The "fixpt" script also takes a "-light" switch to be able to use
2429       this new facility while compiling for a fixpoint.
2430    
2431    2. I replaced all mentions of anchored paths in group owner specifications
2432       with simple relative paths (usually starting with "..").
2433       The rationale is that a library's internal workings should not be
2434       compromised by the lack of some anchor.  (An anchor is necessary
2435       for someone who wants to refer to the library by an anchored path,
2436       but it should not be necessary to build the same library in the first
2437       place.)
2438    
2439    3. I changed the way CM's tool mechanism determines the shell command
2440       string used for things like ml-yacc etc. so that it does not break
2441       when CM.Control.implicit_anchors is turned off.
2442    
2443    ----------------------------------------------------------------------
2444    Name: Matthias Blume
2445    Date: 2000/05/12 18:20:00 JST
2446    Tag: blume-20000512-ml-build
2447    Description:
2448    
2449    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
2450    getting installed properly (by config/install.sh).
2451    
2452    ----------------------------------------------------------------------
2453    Name: Matthias Blume
2454    Date: 2000/05/12 17:30:00 JST
2455    Tag: blume-20000512-anchors
2456    Description:
2457    
2458    !!! NEW BOOT FILES !!!
2459    
2460    This change is in preparation of fading out support for "implicitly
2461    anchored path names".  I went through all sources and used the
2462    explicit (and relatively new) $-notation.  See system/README and the
2463    CM manual for more info on this.
2464    
2465    I also modified the anchoring scheme for some things such as "smlnj",
2466    "MLRISC", "cm", etc. to take advantage of the fact that explicit
2467    anchors are more expressive: anchor name and first arc do not have to
2468    coincide.  This entails the following user-visible change:
2469    
2470    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
2471    particular, when you fire up sml with a command-line argument, say,
2472    e.g.:
2473    
2474       sml '$smlnj/cmb.cm'
2475    
2476    At the ML toplevel prompt:
2477    
2478       CM.autoload "$smlnj/cmb.cm";
2479    
2480    There is also a new controller in CM.Control that can be used to turn
2481    off all remaining support for implicit anchors by saying:
2482    
2483        CM.autoload "$smlnj/
2484        #set CM.Control.implicit_anchors false;
2485    
2486    This causes CM to reject implicitly anchored paths.  This is (for the
2487    time being) less permissive than the "final" version where there will
2488    be no more such implicit anchors and relative paths will be just that:
2489    relative.
2490    
2491    The next step (version after next version?) will be to make the
2492    default for CM.Control.implicit_anchors false.  After the dust has
2493    settled, I can then produce the "final" version of this...
2494    
2495    Note: Since bootstrapping is a bit tricky, I provided new boot files.
2496    
2497    ----------------------------------------------------------------------
2498    Name: Matthias Blume
2499    Date: 2000/05/11 16:30:00 JST
2500    Tag: blume-20000511-sources
2501    Description:
2502    
2503    The main change is that I added function CM.sources as a generalized
2504    version of the earlier CM.makedepend.  This entails the following
2505    additional changes:
2506    
2507      - CM.makedepend has been dropped.
2508    
2509      - CM manual has been updated.
2510    
2511      - TOOLS signature and API have been changed.
2512    
2513    ----------------------------------------------------------------------
2514    Name: Allen Leung
2515    Date: 2000/05/10 21:17 EDT
2516    Tag: leunga-20000510-moby-c--ssa
2517    Description:
2518    
2519      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
2520    None of these affect SML/NJ.
2521    
2522    1. Register Allocation
2523    
2524        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
2525           This module tries to remove local (i.e. basic block level) redundancies
2526           during spilling.
2527    
2528        b. A new framework for performing region based register allocation.
2529           Not yet entirely functional.
2530    
2531    2. X86
2532    
2533       a. DefUse for POP was missing the stack pointer [found by Lal]
2534       b. Reload for CALL was incorrect in X86Spill [found by John]
2535       c. Various fixes in X86Spill so that it can be used correctly for
2536          the new spilling module.
2537    
2538    3. SSA/IR
2539    
2540       a. New module ir/dj-dataflow.sml implements elimination based
2541          data flow analysis.
2542    
2543    4. MLRiscGen
2544    
2545       a. Fix for gc type annotation
2546    
2547    5. MDGen
2548    
2549       Various fixes for machine description -> ml code translation.  For ssa
2550       only.
2551    
2552    ----------------------------------------------------------------------
2553    Name: Allen Leung
2554    Date: 2000/05/08 22:17 EDT
2555    Tag: leunga-20000508-labexp
2556    Description:
2557    
2558      Fermin has found a few assembly problems with constant expressions
2559      generated in LabelExp.  Mostly, the problems involve extra parentheses,
2560      which choke on dumb assemblers.  This is his fix.
2561    
2562    ----------------------------------------------------------------------
2563    Name: Dave MacQueen
2564    Date: 2000/04/09 14:00 EDT
2565    Tag: dbm-20000502-Version_110_28
2566    Description:
2567    
2568    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
2569    
2570    2. Updated config/version to 110.28
2571    
2572    3. Updated config/srcarchiveurl
2573    
2574    3. New boot files!
2575       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
2576    
2577    ----------------------------------------------------------------------
2578    Name: Matthias Blume
2579    Date: 2000/05/01 19:05:00 JST
2580    Tag: blume-20000501-noweb
2581    Description:
2582    
2583    A new noweb tool has been added.  The existing system is entirely
2584    unaffected by this, but some CM users have asked for renewed noweb
2585    support.  Everything is documented in the CM manual.
2586    
2587    New (plugin) libraries:
2588    
2589       noweb-tool.cm
2590       nw-ext.cm
2591    
2592    ----------------------------------------------------------------------
2593    Name: Dave MacQueen
2594    Date: 2000/04/30 12:40PM EDT
2595    Tag: dbm-20000430-bug_fixes
2596    Description:
2597    
2598    1. Fix for bug 1498
2599       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
2600       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
2601         added toRealArray function
2602       smlnj/src/compiler/MiscUtil/print/ppobj.sml
2603         added check for tag Obj.RealArray to array printing case in ppObj
2604    
2605    2. Fix for bug 1510
2606       smlnj/src/compiler/Semant/types/typesutil.sml
2607         fixed definition of dummyargs (used by equalTycon) so that
2608         dummy args are distinct types
2609    
2610    ----------------------------------------------------------------------
2611    Name: Matthias Blume
2612    Date: 2000/04/30 01:00:00 JST
2613    Tag: blume-20000430-versions
2614    Description:
2615    
2616    1. CM version numbering added.  This is an implementation of Lal's
2617       proposal for adding version numbers and version checking to .cm
2618       files.  Lal said that his proposal was just that -- a proposal.
2619       For the time being I went ahead and implemented it so that people
2620       can comment on it.  Everything is completely backward-compatible
2621       (except for the stable library format, i.e., new bootfiles!).
2622    
2623       As usual, see the CM manual for details.
2624    
2625    2. An alternative syntax for anchored paths has been implemented.
2626       Dave has recently voiced the same concerns that I had when I did
2627       this, so there should be some support.  My take is that eventually
2628       I will let support for the current syntax (where anchors are
2629       "implicit") fade out in favor of the new, explicit syntax.
2630       In order to be backward-compatible, both old and new syntax are
2631       currently supported.
2632    
2633       Again, see the CM manual for details.
2634    
2635    3. Parallel make is trying to be slightly smarter:  When the master
2636       process finds a "bottleneck", i.e., when there is only one
2637       compilation unit that can be compiled and everybody else is
2638       waiting on it, then it will simply compile it directly instead
2639       of clumsily telling one of the slaves to do it.
2640    
2641    4. Support for "unsharing" added.  This is necessary in order to be
2642       able to have two different versions of the same library running
2643       at the same time (e.g., for trying out a new MLRISC while still
2644       having the old MLRISC linked into the current compiler, etc.)
2645       See the CM manual.
2646    
2647    5. Simple "makedepend" functionality added for generating Makefile
2648       dependency information.  (This is rather crude at the moment.
2649       Expect some changes here in the future.)
2650    
2651    6. ".fun" added as a recognized suffix for ML files. Also documented
2652       explicitly in the manual that the fallback behavior (unknown suffix
2653       -> ML file) is not an official feature!
2654    
2655    7. Small changes to the pickler for stable libraries.
2656    
2657    8. Several internal changes to CM (for cleanup/improvement).
2658    
2659    
2660    !!!! NEW BINFILES !!!!
2661    
2662    ----------------------------------------------------------------------
2663    Name: Matthias Blume
2664    Date: 2000/04/28 17:30:00 JST
2665    Tag: blume-20000428-pathconfig
2666    Description:
2667    
2668    1. I changed config/install.sh to remove duplicate entries from the
2669       lib/pathconfig file at the end.  Moreover, the final version of
2670       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
2671       in src/system/installml.
2672    
2673    2. The config/install.sh script now consistently uses relative
2674       pathnames in lib/pathconfig whenever the anchor is in the lib
2675       directory.  (So far this was true for the libraries that come
2676       pre-compiled and bundled as part of the bootfiles but not for
2677       libraries that are compiled by the script itself.)
2678    
2679    ----------------------------------------------------------------------
2680    Name: Matthias Blume
2681    Date: 2000/04/26 13:10:00 JST
2682    Tag: blume-20000426-fun_suffix
2683    Description:
2684    
2685    Added ".fun" as a recognized file name suffix (for ML code).
2686    
2687    ----------------------------------------------------------------------
2688    Name: Allen Leung
2689    Date: 2000/04/25 17:00:00 EST
2690    Tag: leunga-20000425-alpha-ra
2691    Description:
2692    
2693    1. Alpha
2694    
2695        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
2696    in C--.
2697    
2698    2. RA
2699    
2700       Added a flag "ra-dump-size" to print out the size of the flowgraph
2701       and the interference graph.
2702    
2703    ----------------------------------------------------------------------
2704    Name: Dave MacQueen
2705    Date: 2000/04/25/
2706    Tag: dbm-20000425-mlyacc_doc_examples
2707    Description:
2708      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
2709      Updated all three examples in src/ml-yacc/examples to run
2710      under 110.* using CM.make.
2711    
2712    ----------------------------------------------------------------------
2713    Name: Allen Leung
2714    Date: 2000/04/20 23:04:00 EST
2715    Tag: leunga-20000420-ssa-c---stuff
2716    Description:
2717    
2718      This update synchronizes my repository with Yale's.  Most of these
2719    changes, however, do not affect SML/NJ at all (the RA is an exception).
2720    
2721    1. Register Allocator
2722    
2723       a. An improvement in the interference graph construction:
2724          Given a copy
2725    
2726                s <- t
2727    
2728          no interference edge between s and t is added for this definition of s.
2729    
2730       b. I've added two new spill heuristic modules that Fermin and I developed
2731          (in the new library RA.cm). These are unused in SML/NJ but maybe
2732          useful for others (Moby?)
2733    
2734    2. X86
2735    
2736       a. Various fixes in the backend provided by Fermin [C--] and Lal.
2737    
2738    3. Alpha
2739    
2740       a. Added the BSR instruction and code generation that goes with it [C--]
2741       b. Other fixes too numerous to recount provided by Fermin [C--]
2742    
2743    4. Regmaps
2744    
2745       a. The regmaps are not initialized with the identity physical bindings
2746          at creation time.  This is unneeded.
2747    
2748    5. MLRISC Optimizations
2749    
2750       a. The DJ-Graph module can now compute the iterated dominance frontiers
2751          intersects with liveness incrementally in linear time! Woohoo!
2752          This is now used in my new SSA construction algorithm.
2753    
2754       b. THe branch reorganization module is now smarter about linear chains of
2755          basic blocks.
2756    
2757    
2758    ----------------------------------------------------------------------
2759    Name: Matthias Blume
2760    Date: 2000/04/12 13:52:00 JST
2761    Tag: blume_main_v110p27_1
2762    Description:
2763    
2764    Changed install.sh script to handle archive files without version number
2765    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
2766    name of the boot file archive.
2767    
2768    ----------------------------------------------------------------------
2769    Name: Dave MacQueen
2770    Date: 2000/04/09 14:00 EDT
2771    Tag: dbm-20000410-Version_110_27
2772    Description:
2773    
2774    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
2775    
2776    2. Updated src/config/version to 110.27
2777    
2778    3. New boot files!
2779    
2780    ----------------------------------------------------------------------
2781    Name: Allen Leung
2782    Date: 2000/04/09 19:09:00 EST
2783    Tag: leunga-20000409-misc
2784    Description:
2785    
2786    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
2787    
2788    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
2789    
2790    ----------------------------------------------------------------------
2791    Name: Stefan
2792    Date: 2000/04/07 10:00:00 EDT
2793    Tag: monnier-20000406-branch-handling
2794    Description:
2795    
2796    Improved handling of branches (mostly those generated from
2797    polymorphic equality), removed switchoff and changed the
2798    default optimization settings (more cpsopt and less flintopt).
2799    
2800    ----------------------------------------------------------------------
2801    Name: Allen Leung
2802    Date: 2000/04/06 01:30:00 EST
2803    Tag: leunga-20000406-peephole-x86-SSA-2
2804    Description:
2805    
2806       Forgot a few files.
2807    
2808    ----------------------------------------------------------------------
2809    Name: Allen Leung
2810    Date: 2000/04/06 00:36:00 EST
2811    Tag: leunga-20000406-peephole-x86-SSA
2812    Description:
2813    
2814    1.  New Peephole code
2815    
2816    2.  Minor improvement to X86 instruction selection
2817    
2818    3.  Various fixes to SSA and machine description -> code translator
2819    
2820    ----------------------------------------------------------------------
2821    Name: Matthias Blume
2822    Date: 2000/04/05 12:30:00 JST
2823    Tag: blume_main_v110p26p2_3
2824    Description:
2825    
2826    This update just merges three minor cosmetic updates to CM's sources
2827    to get ready for the 110.27 code freeze on Friday.  No functionality
2828    has changed.
2829    
2830    ----------------------------------------------------------------------
2831    Name: Allen Leung
2832    Date: 2000/04/04 19:39:00 EST
2833    Tag: leunga-20000404-x86-asm
2834    Description:
2835    
2836    1.  Fixed a problem in X86 assembly.
2837    
2838        Things like
2839    
2840           jmp %eax
2841           jmp (%eax)
2842    
2843        should be output as
2844    
2845           jmp *%eax
2846           jmp *(%eax)
2847    
2848    2.  Assembly output
2849    
2850          Added a new flag
2851    
2852              "asm-indent-copies" (default to false)
2853    
2854          When this flag is on, parallel copies will be indented an extra level.
2855    
2856    ----------------------------------------------------------------------
2857    Name: Allen Leung
2858    Date: 2000/04/04 03:18:00 EST
2859    Tag: leunga-20000404-C--Moby
2860    Description:
2861    
2862        All of these fixes are related to C--, Moby, and my own optimization
2863        stuff; so they shouldn't affect SML/NJ.
2864    
2865    1.  X86
2866    
2867        Various fixes related floating point, and extensions.
2868    
2869    2.  Alpha
2870    
2871        Some extra patterns related to loads with signed/zero extension
2872        provided by Fermin.
2873    
2874    3.  Assembly
2875    
2876        When generating assembly, resolve the value of client defined constants,
2877        instead of generating symbolic values.  This is controlled by the
2878        new flag "asm-resolve-constants", which is default to true.
2879    
2880    4.  Machine Descriptions
2881    
2882        a. The precedence parser was slightly broken when parsing infixr symbols.
2883        b. The type generalizing code had the bound variables reversed, resulting
2884           in a problem during arity raising.
2885        c. Various fixes in machine descriptions.
2886    
2887    ----------------------------------------------------------------------
2888    Name: Matthias Blume
2889    Date: 2000/04/03 16:05:00 JST
2890    Tag: blume_main_v110p26p2_2
2891    Description:
2892    
2893    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
2894    now done via the ordinary static environment that is context to each
2895    compilation unit.
2896    
2897    To this end, I arranged that instead of "structure Core" as "structure
2898    _Core" is bound in the pervasive environment.  Core access is done via
2899    _Core (which can never be accidentally rebound because _Core is not a
2900    legal surface-syntax symbol).
2901    
2902    The current solution is much cleaner because the core environment is
2903    now simply part of the pervasive environment which is part of every
2904    compilation unit's context anyway.  In particular, this eliminates all
2905    special-case handling that was necessary until now in order to deal
2906    with dynamic and symbolic parts of the core environment.
2907    
2908    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2909    compilation manager's bootstrap compiler (actually: in the "init group"
2910    handling).  See the comments in src/system/smlnj/init/init.cmi for
2911    more details.
2912    
2913    I also tried to track down all mentions of "Core" (as string argument
2914    to Symbol.strSymbol) in the compiler and replaced them with a
2915    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
2916    name appears in one place only.
2917    
2918    Binfile and bootfile format have not changed, but the switchover from
2919    the old "init.cmi" to the new one is a bit tricky, so I supplied new
2920    bootfiles anyway.
2921    
2922    ----------------------------------------------------------------------
2923    Name: Allen Leung
2924    Date: 2000/04/02 21:17:00 EST
2925    Tag: leunga-20000402-mltree
2926    Description:
2927    
2928       1. Renamed the constructor CALL in MLTREE by popular demand.
2929       2. Added a bunch of files from my repository.  These are currently
2930          used by other non-SMLNJ backends.
2931    
2932    ----------------------------------------------------------------------
2933    Name: Allen Leung
2934    Date: 2000/03/31 21:15:00 EST
2935    Tag: leunga-20000331-aliasing
2936    Description:
2937    
2938    This update contains a rewritten (and hopefully more correct) module
2939    for extracting aliasing information from CPS.
2940    
2941       To turn on this feature:
2942    
2943            Compiler.Control.CG.memDisambiguate := true
2944    
2945       To pretty print the region information with assembly
2946    
2947           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
2948    
2949       To control how many levels of aliasing information are printed, use:
2950    
2951           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
2952    
2953       The default of n is 3.
2954    
2955    ----------------------------------------------------------------------
2956  Name: David MacQueen  Name: David MacQueen
2957  Date: 2000/03/31 11:15:00 EST  Date: 2000/03/31 11:15:00 EST
2958  Tag: dbm-20000331-runtime_fix  Tag: dbm-20000331-runtime_fix
# Line 81  Line 3021 
3021     (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
3022     used).     used).
3023    
3024     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
3025     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
3026     Makefiles or Shell scripts.     Makefiles or Shell scripts.
3027    
3028     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 94  Line 3034 
3034    
3035     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
3036    
3037     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
3038     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
3039    
3040  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 125  Line 3065 
3065       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3066       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3067       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
3068       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3069       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3070       Dave's hands. :)       Dave's hands. :)
3071    
# Line 228  Line 3168 
3168    
3169       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
3170       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
3171       testl is perferable to andl.       testl is preferable to andl.
3172    
3173  7. RA (x86 only)  7. RA (x86 only)
3174    
# Line 410  Line 3350 
3350    
3351  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
3352  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
3353  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
3354  explanation.  explanation.
3355    
3356  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
3357  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
3358  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
3359  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 457  Line 3397 
3397  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
3398  <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,
3399  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
3400  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
3401  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
3402  and, thus, saves a lot of time.  and, thus, saves a lot of time.
3403    
# Line 507  Line 3447 
3447    it from that remote directory.    it from that remote directory.
3448    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
3449    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
3450    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
3451    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
3452    
3453    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.589  
changed lines
  Added in v.959

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