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 594, Tue Apr 4 23:41:47 2000 UTC revision 939, Tue Sep 25 19:25:48 2001 UTC
# Line 11  Line 11 
11  Date:  Date:
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15    ----------------------------------------------------------------------
16    Name: Matthias Blume
17    Date: 2001/09/25 15:25:00 EDT
18    Tag: blume-20010925-exninfo
19    Description:
20    
21    I provided a non-hook implementation of exnName (at the toplevel) and
22    made the "dummy" implementation of exnMessage (at the toplevel) more
23    useful: if nothing gets "hooked in", then at least you are going to
24    see the exception name and a message indicating why you don't see more.
25    
26    [For the time being, programs that need exnMessage and want to use
27    ml-build should either use General.exnMessage (strongly recommended) or
28    refer to structure General at some other point so that CM sees a
29    static dependency.]
30    
31    [Similar remarks go for "print" and "use":  If you want to use their
32    functionality in stand-alone programs generated by ml-build, then use
33    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
34    
35    ----------------------------------------------------------------------
36    Name: Matthias Blume
37    Date: 2001/09/20 17:28:00 EDT
38    Tag: blume-20010920-slowfp
39    Description:
40    
41    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
42    
43    ----------------------------------------------------------------------
44    Name: Matthias Blume
45    Date: 2001/09/20 17:20:00 EDT
46    Tag: blume-20010920-canonicalpaths
47    Description:
48    
49    0. Updated the BOOT file (something that I forgot to do earlier).
50    
51    1. Small internal change to CM so that it avoids "/../" in filenames
52       as much as possible (but only where it is safe).
53    
54    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
55       that arguments that contain delimiters are passed through correctly.
56       This change also means that all "special" arguments of the form
57       @SMLxxx... must come first.
58    
59    3. Changed install script to put relative anchor names for tool commands
60       into pathconfig.
61    
62    ----------------------------------------------------------------------
63    Name: Matthias Blume
64    Date: 2001/09/18 15:35:00 EDT
65    Tag: blume-20010918-readme11036
66    Description:
67    
68    Added README files.
69    
70    ----------------------------------------------------------------------
71    Name: Matthias Blume
72    Date: 2001/09/18 11:45:00 EDT
73    Tag: Release_110_36 (retag)
74    Description:
75    
76    Fixed mistake in config/preloads. Retagged as 110.36.
77    
78    ----------------------------------------------------------------------
79    Name: Matthias Blume
80    Date: 2001/09/18 09:40:00 EDT
81    Tag: Release_110_36_orig (tag changed)
82    Description:
83    
84    New version (110.36).  New bootfiles.
85    
86    ----------------------------------------------------------------------
87    Name: Matthias Blume
88    Date: 2001/09/14 16:15:00 EDT
89    Tag: blume-20010914-x86fastfp
90    Description:
91    
92    John committed some changes that Allen made, in particular a (hopefully)
93    correctly working version of the x86-fp module.
94    
95    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
96    flag to "true".  Everything seems to compile to a fixpoint ok, and
97    "mandelbrot" speeds up by about 15%.
98    
99    ----------------------------------------------------------------------
100    Name: Matthias Blume
101    Date: 2001/09/13 11:20:00 EDT
102    Tag: blume-20010913-minimal
103    Description:
104    
105    1. Stefan Monnier's patch to fix a miscompilation problem that
106       was brought to light by John Reppy's work on Moby.
107    
108    2. Implemented a minimal "structure Compiler" that contains just
109       "version" and "architecture".  The minimal version will be
110       available when the full version is not.  This is for backward-
111       compatibility with code that wants to test Compiler.version.
112    
113    ----------------------------------------------------------------------
114    Name: Matthias Blume
115    Date: 2001/08/28 14:03:00 EDT
116    Tag: blume-20010828-ml-lex
117    Description:
118    
119    Fix for bug 1581, received from Neophytos Michael.
120    
121    ----------------------------------------------------------------------
122    Name: Matthias Blume
123    Date: 2001/08/27 11:20:00 EDT
124    Tag: blume-20010827-readme11035
125    Description:
126    
127    Fleshed out the README file for 110.35.
128    
129    ----------------------------------------------------------------------
130    Name: Matthias Blume
131    Date: 2001/08/24 17:10:00 EDT
132    Tag: Release_110_35
133    Description:
134    
135    New version number (110.35).  New bootfiles.
136    
137    ----------------------------------------------------------------------
138    Name: Lal George
139    Date: 2001/08/24 13:47:18 EDT 2001
140    Tag: george-20010824-MLRISC-graphs
141    Description:
142    
143     removed clusters from MLRISC completely and replaced with graphs.
144    
145    ----------------------------------------------------------------------
146    Name: Matthias Blume
147    Date: 2001/08/23 17:50:00 EDT
148    Tag: blume-20010823-toplevel
149    Description:
150    
151    - some reorganization of the code that implements various kinds of
152      environments in the compiler (static, dynamic, symbolic, combined)
153    - re-implemented the EnvRef module so that evalStream works properly
154      (if the stream contains references to "use", "CM.make", etc.)
155    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
156    
157    ----------------------------------------------------------------------
158    Name: Matthias Blume
159    Date: 2001/08/20 15:50 EDT
160    Tag: blume20010820-slipup
161    Description:
162    
163    I forgot to commit a few files.  Here they are...
164    
165    ----------------------------------------------------------------------
166    Name: Matthias Blume
167    Date: 2001/08/20 15:35:00 EDT
168    Tag: blume-20010820-debugprof
169    Description:
170    
171    !!!! NEW BOOTFILES !!!!
172    
173    This is another round of reorganizing the compiler sources.  This
174    time the main goal was to factor out all the "instrumentation"
175    passes (for profiling and backtracing) into their own library.
176    The difficulty was to do it in such a way that it does not depend
177    on elaborate.cm but only on elabdata.cm.
178    
179    Therefore there have been further changes to both elaborate.cm and
180    elabdata.cm -- more "generic" things have been moved from the former
181    to the latter.  As a result, I was forced to split the assignment
182    of numbers indicating "primtyc"s into two portions: SML-generic and
183    SML/NJ-specific.  Since it would have been awkward to maintain,
184    I bit the bullet and actually _changed_ the mapping between these
185    numbers and primtycs.  The bottom line of this is that you need
186    a new set of bin- and bootfiles.
187    
188    I have built new bootfiles for all architectures, so doing a fresh
189    checkout and config/install.sh should be all you need.
190    
191    The newly created library's name is
192    
193        $smlnj/viscomp/debugprof.cm
194    
195    and its sources live under
196    
197        src/compiler/DebugProf
198    
199    ----------------------------------------------------------------------
200    Name: Matthias Blume
201    Date: 2001/08/15 17:15:00 EDT
202    Tag: blume-20010815-compreorg
203    Description:
204    
205    This is a first cut at reorganizing the CM libraries that make up the
206    core of the compiler.  The idea is to separate out pieces that could
207    be used independently by tools, e.g., the parser, the typechecker, etc.
208    
209    The current status is a step in this direction, but it is not quite
210    satisfactory yet.  Expect more changes in the future.
211    
212    Here is the current (new) organization...
213    
214        What used to be $smlnj/viscomp/core.cm is now divided into
215        six CM libraries:
216    
217             $smlnj/viscomp/basics.cm
218                           /parser.cm
219                           /elabdata.cm
220                           /elaborate.cm
221                           /execute.cm
222                           /core.cm
223    
224        The CM files for these libraries live under src/system/smlnj/viscomp.
225        All these libraries are proxy libraries that contain precisely
226        one CM library component.  Here are the locations of the components
227        (all within the src/compiler tree):
228    
229             Basics/basics.cm
230             Parse/parser.cm
231             ElabData/elabdata.cm
232             Elaborator/elaborate.cm
233             Execution/execute.cm
234             core.cm
235    
236         [This organization is the same that has been used already
237         for a while for the architecture-specific parts of the visible
238         compiler and for the old version of core.cm.]
239    
240         As you will notice, many source files have been moved from their
241         respective original locations to a new home in one of the above
242         subtrees.
243    
244         The division of labor between the new libraries is the following:
245    
246             basics.cm:
247                - Simple, basic definitions that pertain to many (or all) of
248                  the other libraries.
249             parser.cm:
250                - The SML parser, producing output of type Ast.dec.
251                - The type family for Ast is also defined and exported here.
252             elabdata.cm:
253                - The datatypes that describe input and output of the elaborator.
254                  This includes types, absyn, and static environments.
255             elaborator.cm:
256                - The SML/NJ type checker and elaborator.
257                  This maps an Ast.dec (with a given static environment) to
258                  an Absyn.dec (with a new static environment).
259                - This libraries implements certain modules that used to be
260                  structures as functors (to remove dependencies on FLINT).
261             execute.cm:
262                - Everything having to do with executing binary code objects.
263                - Dynamic environments.
264             core.cm:
265                - SML/NJ-specific instantiations of the elaborator and MLRISC.
266                - Top-level modules.
267                - FLINT (this should eventually become its own library)
268    
269    Notes:
270    
271    I am not 100% happy with the way I separated the elaborator (and its
272    data structures) from FLINT.  Two instances of the same problem:
273    
274        1. Data structures contain certain fields that carry FLINT-specific
275           information.  I hacked around this using exn and the property list
276           module from smlnj-lib.  But the fact that there are middle-end
277           specific fields around at all is a bit annoying.
278    
279        2. The elaborator calculates certain FLINT-related information.  I tried
280           to make this as abstract as I could using functorization, but, again,
281           the fact that the elaborator has to perform calculations on behalf
282           of the middle-end at all is not nice.
283    
284        3. Having to used exn and property lists is unfortunate because it
285           weakens type checking.  The other alternative (parameterizing
286           nearly *everything*) is not appealing, though.
287    
288    I removed the "rebinding =" warning hack because due to the new organization
289    it was awkward to maintain it.  As a result, the compiler now issues some of
290    these warnings when compiling init.cmi during bootstrap compilation. On
291    the plus side, you also get a warning when you do, for example:
292       val op = = Int32.+
293    which was not the case up to now.
294    
295    I placed "assign" and "deref" into the _Core structure so that the
296    code that deals with the "lazy" keyword can find them there.  This
297    removes the need for having access to the primitive environment
298    during elaboration.
299    
300    ----------------------------------------------------------------------
301    Name: Matthias Blume
302    Date: 2001/08/13
303    Tag: blume-20010813-closures
304    Description:
305    
306    This fix was sent to us by Zhong Shao.  It is supposed to improve the
307    performance of certain loops by avoiding needless closure allocation.
308    
309    ----------------------------------------------------------------------
310    Name: Lal George
311    Date: 2001/07/31 10:03:23 EDT 2001
312    Tag: george-20010731-x86-fmalloc
313    Description: Fixed bug in x86 calls
314    
315        There was a bug where call instructions would mysteriously
316        vanish. The call instruction had to be one that returned
317        a floating point value.
318    
319    ----------------------------------------------------------------------
320    Name: Lal George
321    Date: 2001/07/19 16:36:29 EDT 2001
322    Tag: george-20010719-simple-cells
323    Description:
324    
325    I have dramatically simplified the interface for CELLS in MLRISC.
326    
327    In summary, the cells interface is broken up into three parts:
328    
329      1. CellsBasis : CELLS_BASIS
330    
331            CellsBasis is a top level structure and common for all
332            architectures.  it contains the definitions of basic datatypes
333            and utility  functions over these types.
334    
335      2. functor Cells() : CELLS
336    
337            Cells generates an interface for CELLS that incorporates the
338            specific resources on the target architecture, such as the
339            presence of special register classes, their number and size,
340            and various useful substructures.
341    
342      3. <ARCH>CELLS
343    
344            e.g. SparcCells: SPARCCELLS
345    
346            <ARCH>CELLS usually contains additional bindings for special
347            registers  on the architecture, such as:
348    
349                    val r0 : cell           (* register zero *)
350                    val y : cell            (* Y register *)
351                    val psr : cell          (* processor status register *)
352                    ...
353    
354            The structure returned by applying the Cells functor is opened
355            in this interface.
356    
357    The main implication of all this is that the datatypes for cells is
358    split between CellsBasis and CELLS -- a fairly simple change for user
359    code.
360    
361    In the old scheme the CELLS interface had a definitional binding of
362    the form:
363    
364            signature CELLS = sig
365    
366               structure CellsBasis = CellsBasis
367    
368               ...
369    
370            end
371    
372    With all the sharing constraints that goes on in MLRISC, this old
373    design  quickly leads to errors such as:
374    
375            "structure definition spec inside of sharing ... "
376    
377    
378    and appears to require an unacceptable amount of sharing and where
379    constraint hackery.
380    
381    I think this error message (the interaction of definitional specs and
382    sharing) requires more explanation on our web page.
383    
384    ----------------------------------------------------------------------
385    Name: Matthias Blume
386    Date: 2001/07/19 15:00:00 EDT
387    Tag: blume-20010719-libreorg
388    Description:
389    
390    This update puts together a fairly extensive but straightforward change
391    to the way the libraries that implement the interactive system are
392    organized:
393    
394       The biggest change is the elimination of structure Compiler.  As a
395       replacement for this structure, there is now a CM library
396       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
397       that exports all the substructures of the original structure Compiler
398       directly.  So instead of saying Compiler.Foo.bar one now simply
399       says Foo.bar.  (The CM libraries actually export a collection of
400       structures that is richer than the collection of substructures of
401       structure Compiler.)
402    
403       To make the transition smooth, there is a separate library called
404       $smlnj/compiler/compiler.cm which puts together and exports the
405       original structure Compiler (or at least something very close to it).
406    
407       There are five members of the original structure Compiler
408       that are not exported directly but which instead became members
409       of a new structure Backend (described by signature BACKEND).  These are:
410       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
411       Interact (: INTERACT), structure Machine (: MACHINE), and val
412       architecture (: string).
413    
414       Structure Compiler.Version has become structure CompilerVersion.
415    
416       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
417       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
418       or x86, respectively.
419       Each of these exports the same frontend structures that
420       $smlnj/compiler.cm exports.  But they do not have a structure Backend
421       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
422       Hppa, PPC, Sparc, or X86, respectively.
423    
424       Library $smlnj/compiler/all.cm exports the union of the exports of
425       $smlnj/compiler/<arch>.cm
426    
427       There are no structures <Arch>Compiler anymore, use
428       $smlnj/compiler/<arch>.cm instead.
429    
430       Library host-compiler-0.cm is gone.  Instead, the internal library
431       that instantiates CM is now called cm0.cm.  Selection of the host
432       compiler (backend) is no longer done here but. (Responsibility for it
433       now lies with $smlnj/compiler/current.cm.  This seems to be more
434       logical.)
435    
436       Many individual files have been moved or renamed.  Some files have
437       been split into multiple files, and some "dead" files have been deleted.
438    
439    Aside from these changes to library organization, there are also changes
440    to the way the code itself is organized:
441    
442       Structure Binfile has been re-implemented in such a way that it no
443       longer needs any knowledge of the compiler.  It exclusively deals
444       with the details of binfile layout.  It no longer invokes the
445       compiler (for the purpose of creating new prospective binfile
446       content), and it no longer has any knowledge of how to interpret
447       pickles.
448    
449       Structure Compile (: COMPILE) has been stripped down to the bare
450       essentials of compilation.  It no longer deals with linking/execution.
451       The interface has been cleaned up considerably.
452    
453       Utility routines for dealing with linking and execution have been
454       moved into their own substructures.
455    
456       (The ultimate goal of these changes is to provide a light-weight
457       binfile loader/linker (at least for, e.g., stable libraries) that
458       does not require CM or the compiler to be present.)
459    
460    CM documentation has been updated to reflect the changes to library
461    organization.
462    
463    ----------------------------------------------------------------------
464    Name: Matthias Blume
465    Date: 2001/07/10 17:30:00 EDT
466    Tag: Release_110_34
467    Description:
468    
469    Minor tweak to 110.34 (re-tagged):
470    
471      - README.html file added to CVS repository
472      - runtime compiles properly under FreeBSD 3.X and 4.X
473    
474    ----------------------------------------------------------------------
475    Name: Matthias Blume
476    Date: 2001/07/10 17:30:00 EDT
477    Tag: Release_110_34
478    Description:
479    
480    New version number (110.34). New bootfiles.
481    
482    ----------------------------------------------------------------------
483    Name: Matthias Blume
484    Date: 2001/07/09 16:00:00 EDT
485    Tag: blume-20010709-more-varargs
486    Description:
487    
488    I changed the handling of varargs in ml-nlffigen again:
489    The ellipsis ... will now simply be ignored (with an accompanying warning).
490    
491    The immediate effect is that you can actually call a varargs function
492    from ML -- but you can't actually supply any arguments beyond the ones
493    specified explicitly.  (For example, you can call printf with its format
494    string, but you cannot pass additional arguments.)
495    
496    This behavior is only marginally more useful than the one before, but
497    it has the advantage that a function or, more importantly, a function
498    type never gets dropped on the floor, thus avoiding follow-up problems with
499    other types that refer to the offending one.
500    
501    ----------------------------------------------------------------------
502    Name: Matthias Blume
503    Date: 2001/07/09 11:25:00 EDT
504    Tag: blume-20010709-varargs
505    Description:
506    
507    1. ckit-lib.cm now exports structure Error
508    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
509       with a warning accompanied by a source location.  Moreover, it
510       merely skips the offending function or type and proceeds with the
511       rest of its work.u  As a result, one can safely feed C code containing
512       "..." to ml-nlffigen.
513    3. There are some internal improvements to CM, providing slightly
514       more general string substitutions in the tools subsystem.
515    
516    ----------------------------------------------------------------------
517    Name: Matthias Blume
518    Date: 2001/06/27 15:10:00 EDT
519    Tag: blume-20010627-concur
520    Description:
521    
522    Fixed a small bug in CM's handling of parallel compilation.
523    (You could observe the bug by Control-C-interrupting an ordinary
524    CMB.make or CM.stabilize and then attaching some compile servers.
525    The result was that all of a sudden the previously interrupted
526    compilation would continue on its own.  This was because of
527    an over-optimization: CM did not bother to clean out certain queues
528    when no servers were attached "anyway", resulting in the contents
529    of these queues to grab control when new servers did get attached.)
530    
531    There is also another minor update to the CM manual.
532    
533    ----------------------------------------------------------------------
534    Name: Matthias Blume
535    Date: 2001/06/26 16:15:00 EDT
536    Tag: blume-20010626-cmdoc
537    Description:
538    
539    Minor typo fixed in CM manual (syntax diagram for libraries).
540    
541    ----------------------------------------------------------------------
542    Name: Matthias Blume
543    Date: 2001/06/25 22:55:00 EDT
544    Tag: blume-20010625-x86pc
545    Description:
546    
547    Fixed a nasty bug in the X86 assembly code that caused signal
548    handlers to fail (crash) randomly.
549    
550    ----------------------------------------------------------------------
551    Name: Matthias Blume
552    Date: 2001/06/25 12:05:00 EDT
553    Tag: blume-20010625-nlffigen
554    Description:
555    
556    This update fixes a number of minor bugs in ml-nlffigen as reported by
557    Nick Carter <nbc@andrew.cmu.edu>.
558    
559      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
560      2. Default names for generated files are now derived from the name of
561         the C file *without its directory*.  In particular, this causes generated
562         files to be placed locally even if the C file is in some system directory.
563      3. Default names for generated signatures and structures are also derived
564         from the C file name without its directory.  This avoids silly things
565         like "structure GL/GL".
566         (Other silly names are still possible because ml-nlffigen does not do
567          a thorough check of whether generated names are legal ML identifiers.
568          When in doubt, use command line arguments to force particular names.)
569    
570    ----------------------------------------------------------------------
571    Name: Matthias Blume
572    Date: 2001/06/21 12:25:00 EDT
573    Tag: blume-20010621-eXene
574    Description:
575    
576    eXene now compiles and (sort of) works again.
577    
578    The library name (for version > 110.33) is $/eXene.cm.
579    
580    I also added an new example in src/eXene/examples/nbody.  See the
581    README file there for details.
582    
583    ----------------------------------------------------------------------
584    Name: Matthias Blume
585    Date: 2001/06/20 16:40:00 EDT
586    Tag: blume-20010620-cml
587    Description:
588    
589    CML now compiles and works again.
590    
591    Libraries (for version > 110.33):
592    
593      $cml/cml.cm            Main CML library.
594      $cml/basis.cm          CML's version of $/basis.cm.
595      $cml/cml-internal.cm   Internal helper library.
596      $cml/core-cml.cm       Internal helper library.
597      $cml-lib/trace-cml.cm  Tracing facility.
598      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
599    
600    The installer (config/install.sh) has been taught how to properly
601    install this stuff.
602    
603    ----------------------------------------------------------------------
604    Name: Matthias Blume
605    Date: 2001/06/19 17:55:00 EDT
606    Tag: blume-20010619-instantiate
607    Description:
608    
609    This un-breaks the fix for bug 1432.
610    (The bug was originally fixed in 110.9 but I broke it again some
611    time after that.)
612    
613    ----------------------------------------------------------------------
614    Name: Matthias Blume
615    Date: 2001/06/19 17:25:00 EDT
616    Tag: blume-20010619-signals
617    Description:
618    
619    This should (hopefully) fix the long-standing signal handling bug.
620    (The runtime system was constructing a continuation record with an
621    incorrect descriptor which would cause the GC to drop data on the floor...)
622    
623    ----------------------------------------------------------------------
624    Name: Matthias Blume
625    Date: 2001/06/15 15:05:00 EDT
626    Tag: blume-20010615-moresparc
627    Description:
628    
629    Here is a short late-hour update related to Sparc c-calls:
630    
631     -- made handling of double-word arguments a bit smarter
632    
633     -- instruction selection phase tries to collapse certain clumsily
634        constructed ML-Trees; typical example:
635    
636            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
637    
638        This currently has no further impact on SML/NJ since mlriscGen does
639        not seem to generate such patterns in the first place, and c-calls
640        (which did generate them in the beginning) has meanwhile been fixed
641        so as to avoid them as well.
642    
643    ----------------------------------------------------------------------
644    Name: Matthias Blume
645    Date: 2001/06/15 15:05:00 EDT
646    Tag: blume-20010615-sparc
647    Description:
648    
649    The purpose of this update is to provide an implementation of NLFFI
650    on Sparc machines.
651    
652    Here are the changes in detail:
653    
654       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
655       the Sparc implementation of the c-calls API.
656       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
657       for accessing the ML frame.  Thus, we have a real frame pointer and
658       can freely modify %sp without need for an omit-frame-ptr phase.
659       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
660       this case.
661       * ml-nlffigen has been taught to produce code for different architectures
662       and calling conventions.
663       * In a way similar to what was done in the x86 case, the Sparc
664       backend uses its own specific extension to mltree.  (For example,
665       it needs to be able to generate UNIMP instructions which are part
666       of the calling convention.)
667       * ml-nlffi-lib was reorganized to make it more modular (in particular,
668       to make it easier to plug in new machine- and os-dependent parts).
669    
670    There are some other fairly unrelated bug fixes and cleanups as well:
671    
672       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
673       that they properly share their libraries with existing SML/NJ libraries.
674       * I fixed a minor cosmetic bug in CM, supressing certain spurious
675       follow-up error messages.
676       * Updates to CM/CMB documentation.
677    
678    TODO items:
679    
680       * MLRISC should use a different register as its asmTemp on the Sparc.
681         (The current %o2 is a really bad choice because it is part of the
682         calling conventions, so things might interfere in unexpected ways.)
683    
684    ----------------------------------------------------------------------
685    Name: Matthias Blume
686    Date: 2001/06/07
687    Tag: blume-20010607-calls
688    Description:
689    
690    A number of internal changes related to C calls and calling conventions:
691    
692    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
693       number of bytes popped implicitly (by the callee).  In most cases
694       this field is 0 but on x86/win32 it is some non-zero value.  This
695       is information provided for the benefit of the "omit-frameptr" pass.
696    2. The CALL instruction on the x86 carries a similar "pops" field.
697       The instruction selection phase copies its value from the ML-Tree
698       CALL statement.
699    3. On all other architectures, the instruction selection phase checks
700       whether "pops=0" and complains if not.
701    4. The c-calls implementation for x86 now accepts two calling conventions:
702       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
703       up after the call and pops is set to 0.  For "stdcall", the caller
704       does nothing, leaving the cleanup to the callee; pops is set to
705       the number of bytes that were pushed onto the stack.
706    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
707       distinguish between "ccall" and "stdcall".
708    6. The UNIMP instruction has been added to the supported Sparc instruction
709       set. (This is needed for implementing the official C calling convention
710       on this architecture.)
711    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
712       work with the latest CM.
713    
714    ----------------------------------------------------------------------
715    Name: Matthias Blume
716    Date: 2001/06/05 15:10:00 EDT
717    Tag: blume-20010605-cm-index
718    Description:
719    
720    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
721    
722    1. CM can now generate "index files".  These are human-readable files
723       that list on a per-.cm-file basis each toplevel symbol defined or
724       imported.  The location of the index file for
725       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
726       To enable index-file generation, set CM.Control.generate_index to true
727       or export an environment-symbol: export CM_GENERATE_INDEX=true.
728    
729       The CM manual has been updated accordingly.
730    
731    2. I made some slight modifications to the c-calls API in MLRISC.
732    
733         a) There is now a callback to support saving/restoring of
734            dedicated but caller-save registers around the actual call
735            instruction.
736         b) One can optionally specify a comment-annotation for the
737            call instruction.
738    
739    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
740       (For example, the comment annotation shows the C prototype of
741        the function being called.)
742    
743    ----------------------------------------------------------------------
744    Name: Matthias Blume
745    Date: 2001/06/01 13:30:00 EDT
746    Tag: blume-20010601-nlffi-cleanup
747    Description:
748    
749    This is mostly a cleanup of MLFFI stuff:
750    
751       - some signature files have been put into a more exposed place
752       - the ugly 'f type parameter is gone (simplifies types tremendously!)
753       - ml-nlffigen changed accordingly
754       - tutorial updated
755    
756    Other changes:
757    
758       - author's affiliation in CM manual(s) updated
759       - some more recognized keywords added to Allen's sml.sty
760    
761    ----------------------------------------------------------------------
762    Name: Matthias Blume
763    Date: 2001/05/25 15:30:00 EDT
764    Tag: blume-20010525-iptr
765    Description:
766    
767      - put the official 110.33-README (as it appears on the ftp server) under
768        CVS
769      - fixed a small bug related to incomplete pointer types in
770        ml-nlffigen
771      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
772        (it does not need the 'f type parameter)
773    
774    ----------------------------------------------------------------------
775    Name: Matthias Blume
776    Date: 2001/05/23 14:30:00 EDT
777    Tag: Release_110_33
778    Description:
779    
780    New version number (110.33).  New bootfiles.
781    
782    ----------------------------------------------------------------------
783    Name: Matthias Blume
784    Date: 2001/05/22 18:06:00 EDT
785    Tag: blume-20010522-targets
786    Description:
787    
788    Made install.sh use file config/targets.customized if it exists, falling
789    back to config/targets if it doesn't.  This way one can have a customized
790    version of the targets file without touching the "real thing", thus
791    eliminating the constant fear of accidentally checking something bogus
792    back into the CVS repository...  (File config/targets.customized must
793    not be added to the repository!)
794    
795    ----------------------------------------------------------------------
796    Name: Matthias Blume
797    Date: 2001/05/22 16:30:00 EDT
798    Tag: blume-20010522-minitut
799    Description:
800    
801    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
802       struct returns.
803    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
804       incomplete, preliminary documentation for NLFFI.
805    
806    ----------------------------------------------------------------------
807    Name: Matthias Blume
808    Date: 2001/05/14 11:30:00 EDT
809    Tag: blume-20010514-script
810    Description:
811    
812    Some bugs in install script fixed.
813    
814    In addition to that I also made a slight change to the NLFFI API:
815    Functors generated by ml-nlffigen now take the dynamic library as a
816    straight functor argument, not as a suspended one.  (The original
817    functor code used to force the suspension right away anyway, so there
818    was nothing gained by this complication of the interface.)
819    
820    ----------------------------------------------------------------------
821    Name: Matthias Blume
822    Date: 2001/05/11 14:35:00 EDT
823    Tag: blume-20010511-ml-nlffi
824    Description:
825    
826    I finally took the plunge and added my new FFI code to the main
827    repository.  For x86-linux it is now ready for prime-time.
828    
829    There are two new subdirectories of "src":
830    
831      - ml-nlffi-lib:
832           The utility library for programs using the FFI interface.
833           Here is the implementation of $/c.cm and its associated low-level
834           partners $/c-int.cm and $/memory.cm.
835      - ml-nlffigen:
836           A stand-alone program for generating ML glue code from C source
837           code.
838    
839    Building ml-nlffigen requires $/ckit-lib.cm.
840    
841    The config/install.sh script has been updates to do the Right Thing
842    (hopefully).
843    
844    Notice that the source tree for the C-Kit will not be put under "src"
845    but directly under the installation root directory.  (This is the
846    structure that currently exists on the CVS server when you check out
847    module "sml".)  Fortunately, config/install.sh knows about this oddity.
848    
849    Bugs: No documentation yet.
850    
851    ----------------------------------------------------------------------
852    Name: Matthias Blume
853    Date: 2001/05/09 16:35:00 EDT
854    Tag: blume-20010509-cpscontract
855    Description:
856    
857    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
858    wrapper/unwrapper elimination did not decrement usage counts and some
859    dead variables got overlooked by the dead-up logic.)
860    
861    ----------------------------------------------------------------------
862    Name: Lal George
863    Date: 2001/05/08  17:26:09 EDT
864    Tag: george-20010508-omit-frameptr
865    Description:
866    
867    Changes to implement the omit-frame-pointer optimization to support
868    raw C calls. For now, there is only support on the Intel x86, but
869    other architectures will follow as more experience is gained with this.
870    
871    
872    ----------------------------------------------------------------------
873    Name: Matthias Blume
874    Date: 2001/05/07 14:40:00 EDT
875    Tag: blume-20010507-proxies
876    Description:
877    
878    I made into "proxy libraries" all libraries that qualify for such a
879    change.  (A qualifying library is a library that has another library or
880    groups as its sole member and repeats that member's export list
881    verbatim.  A proxy library avoids this repetition by omitting its export
882    list, effectively inheriting the list that its (only) member exports.
883    See the CM manual for more explanation.)
884    The main effect is that explicit export lists for these libraries
885    do not have to be kepts in sync, making maintenance a bit easier.
886    
887    I also added copyright notices to many .cm-files.
888    
889    Last but not least, I made a new set of bootfiles.
890    
891    ----------------------------------------------------------------------
892    Name: Matthias Blume
893    Date: 2001/05/04 17:00:00 EDT
894    Tag: blume-20010504-cm-lsplit
895    Description:
896    
897    0. John merged pending changes to $/smlnj-lib.cm
898    
899    1. Allen's previous change accidentally backed out of one of Lal's
900       earlier changes.  I undid this mistake (re-introducing Lal's change).
901    
902    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
903       within the compiler where applicable.  There is some code simplification
904       because of that.
905    
906    3. The "split" phase (in FLINT) is now part of the default list of phases.
907       Compiler.Control.LambdaSplitting.* can be used to globally control the
908       lambda-splitting (cross-module-inlining) engine.  In addition to that,
909       it can now also be controlled on a per-source basis: CM has been taught
910       a new tool parameter applicable to ML source files.
911    
912       - To turn lambda-splitting off completely:
913            local open Compiler.Control.LambdaSplitting in
914                val _ = set Off
915            end
916       - To make "no lambda-splitting" the global default (but allow per-source
917         overriding); this is the initial setting:
918            local open Compiler.Control.LambdaSplitting in
919                val _ = set (Default NONE)
920            end
921       - To make "lambda-splitting with aggressiveness a" the global default
922         (and allow per-source overriding):
923            local open Compiler.Control.LambdaSplitting in
924                val _ = set (Default (SOME a))
925            end
926    
927       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
928         write (in the respective .cm-file):
929            a.sml (lambdasplitting:off)
930       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
931            a.sml (lambdasplitting:on)
932       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
933         <a> is a decimal non-negative integer):
934            a.sml (lambdasplitting:<a>)
935       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
936            a.sml (lambdasplitting:infinity)
937       - To use the global default for a.sml:
938            a.sml (lambdasplitting:default)
939         or simply
940            a.sml
941    
942    ----------------------------------------------------------------------
943    Name: Allen Leung
944    Date: 2001/05/04 01:57:00 EDT
945    Tag: leunga-20010504-sync
946    Description:
947    
948      MLRISC features.
949    
950      1. Fix to CMPXCHG instructions.
951      2. Changed RA interface to allow annotations in callbacks.
952      3. Added a new method to the stream interface to allow annotations updates.
953    
954    ----------------------------------------------------------------------
955    Name: Matthias Blume
956    Date: 2001/05/01 11:45:00 EDT
957    Tag: blume-20010501-pcedittmp
958    Description:
959    
960    Changed install.sh to use the current working directory instead of
961    /usr/tmp for a temporary file (pcedittmp).  The previous choice
962    of /usr/tmp caused trouble with MacOS X because of file premission
963    problems.
964    
965    ----------------------------------------------------------------------
966    Name: Matthias Blume
967    Date: 2001/04/20 11:10:00 EDT
968    Tag: blume-20010420-inMLflag
969    Description:
970    
971     - added vp_limitPtrMask to vproc-state.h
972       (for use by the raw-C-calls mechanism to implement proper interrupt
973        handling)
974     - made the ML compiler aware of various data-structure offsets so it
975       can generate code for accessing the vp_inML flag and vp_limitPtrMask
976     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
977       raw C-calls
978    
979    ----------------------------------------------------------------------
980    Name: Lal George
981    Date: 2001/04/20 09:15:28 EDT
982    Tag: george-20010420-macosX
983    Description:
984    
985     - Changes to port to Mac OS X; Darwin.
986    
987     - In the process I found that sqrt was broken on the PPC, because the
988       fsqrt instruction is not implemented.
989    
990    ----------------------------------------------------------------------
991    Name: Matthias Blume
992    Date: 2001/04/18 12:45:00 EDT
993    Tag: blume-20010418-ccalls
994    Description:
995    
996     - fixed two off-by-4 errors in the x86-specific c-calls implementation
997       (this bug prevented structure arguments containing pointers from being
998        passed correctly)
999     - changed the raw-C-call code in mlriscGen.sml in such a way that
1000       structure arguments are represented as a pointer to the beginning
1001       of the structure (instead of having a series of synthesized arguments,
1002       one for each structure member)
1003    
1004     - made makeml script's verbosity level configurable via environment
1005       variable (MAKEML_VERBOSITY)
1006    
1007     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
1008       in rawmem-x86.sml; we are now using the real thing
1009    
1010    ----------------------------------------------------------------------
1011    Name: Matthias Blume
1012    Date: 2001/03/22 16:25:00 EST
1013    Tag: blume-20010322-bootfiles
1014    Description:
1015    
1016    Created a new set of bootfiles (for your automatic installation convenience).
1017    
1018    ----------------------------------------------------------------------
1019    Name: Matthias Blume
1020    Date: 2001/03/22 15:10:00 EST
1021    Tag: blume-20010322-rawmem-parcm
1022    Description:
1023    
1024    1. All "raw memory access" primitives for the new FFI are implemented now
1025       (at least on the x86).
1026    2. Some further cleanup of CM's parallel make mechanism.
1027    
1028    ----------------------------------------------------------------------
1029    Name: Matthias Blume
1030    Date: 2001/03/19 17:53:00 EST
1031    Tag: blume-20010319-parallel
1032    Description:
1033    
1034    Parallel make (using compile servers) now works again.
1035    
1036    To this end, CM.stabilize and CMB.make have been modified to work in
1037    two passes when compile servers are attached:
1038       1. Compile everything, do not perform stabilization; this pass
1039          uses compile servers
1040       2. Stabilize everything; this pass does not use compile servers
1041    If there are no compile servers, the two passes are combined into one
1042    (as before).  Splitting the passes increases the inherent parallelism
1043    in the dependency graph because the entire graph including all
1044    libraries is available at the same time.  This, in turn, improves
1045    server utilization.  The downside is that the master process will
1046    have to do some extra work after compilation is done (because for
1047    technical reasons it must re-read all the binfiles during stabilization).
1048    
1049    ----------------------------------------------------------------------
1050    Name: Matthias Blume
1051    Date: 2001/03/16 12:22:00 EST
1052    Tag: blume-20010316-bootfiles
1053    Description:
1054    
1055    Created a new set of bootfiles (for your automatic installation convenience).
1056    
1057    ----------------------------------------------------------------------
1058    Name: Matthias Blume
1059    Date: 2001/03/16 11:00:00 EST
1060    Tag: blume-20010316-MLTREE-fixup
1061    Description:
1062    
1063    This is a minor fixup for an (untagged) earlier commit by Allen.
1064    (A file was missing).
1065    
1066    ----------------------------------------------------------------------
1067    Name: Allen Leung
1068    Date: Mon Mar  5 18:54:57 EST 2001
1069    Tag: leunga-20010305-cut-support
1070    
1071    1. New support for alternative control-flow in MLTREE.
1072       Currently we support
1073    
1074          FLOW_TO(CALL ...., [k1,...,kn])
1075    
1076       This is needed for 'cuts to' in C-- and try/handle-like constructs
1077       in Moby
1078    
1079       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
1080    
1081    2. Register Allocator
1082    
1083       Changes in interface [from Fermin, John]
1084    
1085    3. Alpha 8-bit SLL support [Fermin]
1086    
1087    4. All architectures
1088    
1089       A new module (ClusterExpandCopies) for expanding parallel copies.
1090    
1091    ----------------------------------------------------------------------
1092    Name: Allen Leung
1093    Date: 2001/02/27 23:07:00 EST
1094    Tag: leunga-20010227-minor-stuff
1095    
1096    1. Alpha bug fix for CMOVNE
1097    2. Handle mltree COND(..,FCMP ...,...)
1098    3. Bug fix in simplifier
1099    
1100    ----------------------------------------------------------------------
1101    Name: Matthias Blume
1102    Date: 2001/01/30 17:50:00 EST
1103    Tag: blume-20010130-sync
1104    Description:
1105    
1106    This is just a minor update to sync my devel branch with the main brach.
1107    The only visible change is the addition of some README files.
1108    
1109    ----------------------------------------------------------------------
1110    Name: Matthias Blume
1111    Date: 2001/01/12 23:30:00 JST
1112    Tag: blume-20010112-bootfiles
1113    Description:
1114    
1115    Made a new set of bootfiles that goes with the current state of the
1116    repository.
1117    
1118    ----------------------------------------------------------------------
1119    Name: Matthias Blume
1120    Date: 2001/01/12 21:20:00 JST
1121    Tag: blume-20010112-sync
1122    Description:
1123    
1124    I am just flushing out some minor changes that had accumulated in
1125    my private branch in order to sync with the main tree.  (This is
1126    mainly because I had CVS trouble when trying to merge _into_ my
1127    private branch.)
1128    
1129    Most people should be completely unaffected by this.
1130    
1131    ----------------------------------------------------------------------
1132    Name: Allen Leung
1133    Date: Thu Jan 11 21:03:00 EST 2001
1134    Tag: leunga-20010111-labexp=mltree
1135    Description:
1136    
1137    1.  Removed the type LabelExp and replace it by MLTree.
1138    2.  Rewritten mltree-simplify with the pattern matcher tool.
1139    3.  There were some bugs in alpha code generator which would break
1140        64-bit code generation.
1141    4.  Redo the tools to generate code with the
1142    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
1143        are now generated by perl scripts.
1144    
1145    ----------------------------------------------------------------------
1146    Name: Matthias Blume
1147    Date: 2001/01/10 21:55:00 JST
1148    Tag: blume-20010110-rcc
1149    Description:
1150    
1151    The RCC stuff now seems to work (but only on the x86).
1152    This required hacking of the c-calls interface (and -implementation) in
1153    MLRISC.
1154    
1155    Normal compiler users should be unaffected.
1156    
1157    ----------------------------------------------------------------------
1158    Name: Matthias Blume
1159    Date: 2001/01/09 01:20:00 JST
1160    Tag: blume-20010109-rcc
1161    Description:
1162    
1163    This is a fairly big patch, flushing out a large number of pending
1164    changes that I made to my development copy over the last couple of days.
1165    
1166    Of practical relevance at this moment is a workaround for a pickling
1167    bug that Allen ran into the other day.  The cause of the bug itself is
1168    still unknown and it might be hard to fix it properly, but the
1169    workaround has some merits of its own (namely somewhat reducing pickling
1170    overhead for certain libraries).  Therefore, I think this solution should
1171    be satisfactory at this time.
1172    
1173    The rest of the changes (i.e., the vast majority) has to do with my
1174    ongoing efforts of providing direct support for C function calls from
1175    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
1176    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
1177    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
1178    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
1179    handle RCC.
1180    
1181    The last part (the changes to mlriscGen.sml) are still known to be
1182    wrong on the x86 and not implemented on all other architectures.  But
1183    the infrastructure is in place. I had to change a few functor
1184    signatures in the backend to be able to route the CCalls interface
1185    from MLRISC there, and I had to specialize the mltree type (on the
1186    x86) to include the necessary extensions. (The extensions themselves
1187    were already there and redy to go in MLRISC/x86).
1188    
1189    Everything should be very happy as soon as someone helps me with
1190    mlriscGen.sml...
1191    
1192    In any case, nothing of this should matter to anyone as long as the
1193    new primop is not being used (which is going to be the case unless you
1194    find it where I hid it :). The rest of the compiler is completely
1195    unaffected.
1196    
1197    ----------------------------------------------------------------------
1198    Name: Matthias Blume
1199    Date: 2001/01/05 00:30:00 JST
1200    Tag: blume-20010105-primops
1201    Description:
1202    
1203    Added some experimental support for work that I am doing right now.
1204    These changes mostly concern added primops, but there is also a new
1205    experimental C library in the runtime system (but currently not enabled
1206    anywhere except on Linux/X86).
1207    
1208    In the course of adding primops (and playing with them), I discovered that
1209    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
1210    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
1211    To recover, I made type information in INL_PRIM mandatory and changed
1212    prim.sml as well as built-in.sml accordingly.  The InLine structure now
1213    has complete, correct type information (i.e., no bottom types).
1214    
1215    Since all these changes mean that we need new binfiles, I also bumped the
1216    version number to 110.32.1.
1217    
1218    ----------------------------------------------------------------------
1219    Name: Matthias Blume
1220    Date: 2000/12/30 22:10:00 JST
1221    Tag: blume-20001230-various
1222    Description:
1223    
1224    Added proxy libraries for MLRISC and let MLRISC libraries refer
1225    to each other using path anchors.  (See CM manual for explanation.)
1226    
1227    Updated CM documentation.
1228    
1229    Fixed some bugs in CM.
1230    
1231    Implemented "proxy" libraries (= syntactic sugar for CM).
1232    
1233    Added "-quiet" option to makeml and changed runtime system accordingly.
1234    
1235    Added cleanup handler for exportML to reset timers and compiler stats.
1236    
1237    ----------------------------------------------------------------------
1238    Name: Lal George
1239    Date: 2000/12/22 22:22:58 EST 2000
1240    Tag: Release_110_32
1241    Description:
1242    
1243            Infinite precision used throughout MLRISC.
1244            see MLRISC/mltree/machine-int.sig
1245    
1246    ----------------------------------------------------------------------
1247    Name: Matthias Blume
1248    Date: 2000/12/22 23:16:00 JST
1249    Tag: blume-20001222-warn
1250    Description:
1251    
1252    Corrected wording and formatting of some CM warning message which I
1253    broke in my previous patch.
1254    
1255    ----------------------------------------------------------------------
1256    Name: Matthias Blume
1257    Date: 2000/12/22 21:20:00 JST
1258    Tag: blume-20001222-anchorenv
1259    Description:
1260    
1261    Fixed CM's handling of anchor environments in connection with CMB.make.
1262    
1263    ----------------------------------------------------------------------
1264    Name: Matthias Blume
1265    Date: 2000/12/22 13:15:00 JST
1266    Tag: blume-20001222-cleanup
1267    Description:
1268    
1269    Removed src/cm/ffi which does not (and did not) belong here.
1270    
1271    ----------------------------------------------------------------------
1272    Name: Matthias Blume
1273    Date: 2000/12/21 23:55:00 JST
1274    Tag: blume-20001221-exn
1275    Description:
1276    
1277    Probably most important: CM no longer silently swallows all exceptions
1278    in the compiler.
1279    Plus: some other minor CM changes.  For example, CM now reports some
1280    sizes for generated binfiles (code, data, envpickle, lambdapickle).
1281    
1282    ----------------------------------------------------------------------
1283    Name: Matthias Blume
1284    Date: 2000/12/15 00:01:05 JST
1285    Tag: blume-20001215-dirtool
1286    Description:
1287    
1288    - "dir" tool added.
1289    - improvements and cleanup to Tools structure
1290    - documentation updates
1291    
1292    ----------------------------------------------------------------------
1293    Name: Allen Leung
1294    Date: Thu Dec 14 03:45:24 EST 2000
1295    Description:
1296    Tag:  leunga-20001214-int-inf
1297    Description:
1298    
1299       In IntInf, added these standard functions, which are missing from our
1300    implementation:
1301    
1302        andb : int * int -> int
1303        xorb : int * int -> int
1304        orb  : int * int -> int
1305        notb : int -> int
1306         <<   : int * word -> int
1307        ~>>  : int * word -> int
1308    
1309       Not tested, I hope they are correct.
1310    
1311    ----------------------------------------------------------------------
1312    Name: Allen Leung
1313    Date: Fri Dec  8 19:23:26 EST 2000
1314    Description:
1315    Tag:  leunga-20001208-nowhere
1316    Description:
1317    
1318      Slight improvements to the 'nowhere' tool to handle OR-patterns,
1319    to generate better error messages etc.  Plus a brief manual.
1320    
1321    ----------------------------------------------------------------------
1322    Name: Lal George
1323    Date: 2000/12/08 09:54:02 EST 2000
1324    Tag: Release_110_31
1325    Description:
1326    
1327    - Version 110.31
1328    ----------------------------------------------------------------------
1329    Name: Allen Leung
1330    Date: Thu Dec  7 22:01:04 EST 2000
1331    Tag:  leunga-20001207-cell-monster-hack
1332    Description:
1333    
1334    Major MLRISC internal changes.  Affect all clients.
1335    Summary:
1336    
1337    1.  Type CELLS.cell = int is now replaced by a datatype.
1338        As a result, the old regmap is now gone.  Almost all interfaces
1339        in MLRISC change as a consequence.
1340    
1341    2.  A new brand version of machine description tool (v3.0) that generates
1342        modules expecting the new interface.  The old version is removed.
1343    
1344    3.  The RA interface has been further abstracted into two new functors.
1345        RISC_RA and X86RA.  These functors have much simpler interfaces.
1346        [See also directory MLRISC/demo.]
1347    
1348    4.  Some other new source->source code generation tools are available:
1349    
1350        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
1351        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
1352           I use this tool to generate the peephole optimizers---with the new
1353           cell type changes, peephole rules are becoming difficult to write
1354           without conditional pattern matching.
1355    
1356    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
1357        cover the entire MLRISC source tree so many things broke.
1358    
1359    6.  CM files have been moved to the subdirectory MLRISC/cm.
1360        They are moved because there are a lot of them and they clutter up the
1361        root dir.
1362    
1363    7.  More detailed documentation to come...
1364    
1365        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
1366        a makeml -rebuild first.  This is because of other other
1367        changes that Matthias has made (see below).
1368    
1369    
1370    ----------------------------------------------------------------------
1371    Name: Matthias Blume
1372    Date: 2000/11/30 23:12:00 JST
1373    Tag: blume-20001130-filereorg
1374    Description:
1375    
1376    Some manual updates and some file reorganizations in CM.
1377    
1378    ----------------------------------------------------------------------
1379    Name: Matthias Blume
1380    Date: 2000/11/24 17:45:00 JST
1381    Tag: blume-20001124-link
1382    Description:
1383    
1384    Drastically improved link traversal code for the case that the dynamic
1385    value was already loaded at bootstrap time.  As a result, CM and CMB
1386    now both load blazingly fast -- even on a very slow machine.  Also,
1387    memory consumption has been further reduced by this.
1388    
1389    Warning: The format of the PIDMAP file has changed.  THerefore, to
1390    bootstrap you have to do this:
1391    
1392    1. Run CMB.make
1393    2. Make a symbolic link for the boot directory:
1394         ln -s sml.boot.ARCH-OS xxx
1395    3. "Rebuild" the boot directory:
1396         ./makeml -boot xxx -rebuild sml ; rm xxx
1397    4. Boot normally:
1398          ./makeml
1399    
1400    ----------------------------------------------------------------------
1401    Name: Matthias Blume
1402    Date: 2000/11/21 21:20:00 JST
1403    Tag: blume-20001121-tools
1404    Description:
1405    
1406    Continued hacking on autoloading problem -- with success this time.
1407    Also changed tool-plugin mechanism.  See new CM manual.
1408    
1409    ----------------------------------------------------------------------
1410    Name: Matthias Blume
1411    Date: 2000/11/19 14:30:00 JST
1412    Tag:  blume-20001119-autoload
1413    Description:
1414    
1415    Some hacking to make autoloading faster.  Success for CMB, no success
1416    so far for CM.  There is a reduced structure CM' that autoloads faster.
1417    (This is a temporary, non-documented hack to be eliminated again when
1418    the general problem is solved.)
1419    
1420    ----------------------------------------------------------------------
1421    Name: Matthias Blume
1422    Date: 2000/11/17 14:10:00 JST
1423    Tag: blume-20001117-pickle-lib
1424    Description:
1425    
1426    1. Eliminated comp-lib.cm
1427    2. Made pickle-lib.cm
1428    3. Eliminated all uses of intset.sml (from comp-lib.cm)
1429    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
1430       equivalent constructs from smlnj-lib.cm (INtHashTable).
1431    5. Point 4. also goes for those uses of intmap.* in MLRISC.
1432       Duplicated intmap modules thrown out.
1433    6. Hunted down all duplicated SCC code and replaced it with
1434       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
1435    7. Rewrote Feedback module.
1436    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
1437       should be thrown out and equivalent modules from smlnj-lib.cm
1438       should be used (IntRedBlackSet, IntListSet, ...).
1439    
1440    Confirmed that compiler compiles to fixpoint.
1441    
1442    ----------------------------------------------------------------------
1443    Name: Allen Leung
1444    Date: 2000/11/10 18:00:00
1445    Tag: leunga-20001110-new-x86-fp
1446    
1447    A new x86 floating point code generator has been added.
1448    By default this is turned off.  To turn this on, do:
1449    
1450        CM.autoload "$smlnj/compiler.cm";
1451        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
1452    
1453    Changes:
1454    
1455    1.  Changed FTAN to FPTAN so that the assembly output is correct.
1456    2.  Changed the extension callback for FTANGENT to generate:
1457    
1458              fptan
1459              fstp  %st(0)
1460        instead of
1461              fptan
1462              fstpl ftempmem
1463    
1464    3.  Numerous assembly fixes for x86.
1465    
1466    5.  Cleaned up the machine code output module x86/x86MC.sml and added
1467        support for a whole bunch of instructions and addressing modes:
1468    
1469          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
1470          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
1471          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
1472          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
1473          fxch %st(n)
1474          fld %st(n)
1475          fst %st(n)
1476          fst mem
1477          fstp %st(n)
1478          fucom %st(n)
1479          fucomp %st(n)
1480    
1481        All these are now generated when the fast fp mode is turned on.
1482    
1483    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
1484    
1485    ----------------------------------------------------------------------
1486    Name: Matthias Blume
1487    Date: 2000/11/09 11:20:00 JST
1488    Tag: blume-20001109-scc
1489    Description:
1490    
1491    Eliminated some code duplication:
1492    
1493    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
1494       (Otherwise the functor is useless.)
1495    2. Used GraphSCCFn where SCCUtilFun was used previously.
1496    3. Got rid of SCCUtilFun (in comp-lib.cm).
1497    
1498    ----------------------------------------------------------------------
1499    Name: Lal George
1500    Date: 2000/11/06 09:02:21 EST 2000
1501    Tag: Release_110_30
1502    Description:
1503    
1504    - Version 110.30
1505    ----------------------------------------------------------------------
1506    Name: Matthias Blume
1507    Date: 2000/11/04 14:45:00
1508    Tag: blume-20001104-mlbuild
1509    Description:
1510    
1511    - Made ml-build faster on startup.
1512    - Documentation fixes.
1513    
1514    ----------------------------------------------------------------------
1515    Name: Matthias Blume
1516    Date: 2000/11/02 17:00:00 JST
1517    Tag: blume-20001102-condcomp
1518    Description:
1519    
1520    - Small tweaks to pickler -- new BOOTFILES!
1521    - Version bumped to 110.29.2.
1522    - Added conditional compilation facility to init.cmi (see comment there).
1523    ----------------------------------------------------------------------
1524    Name: Allen Leung
1525    Date: 2000/10/23 19:31:00
1526    Tag: leunga-20001023-demo-ra
1527    
1528    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
1529    2. Test programs for the graph library updated
1530    3. Some new MLRISC demo programs added
1531    
1532    ----------------------------------------------------------------------
1533    Name: Matthias Blume
1534    Date: 2000/08/31 22:15:00 JST
1535    Tag: blume-20001017-errmsg
1536    Description:
1537    
1538    More error message grief: Where there used to be no messages, there
1539    now were some that had bogus error regions.  Fixed.
1540    
1541    ----------------------------------------------------------------------
1542    Name: Matthias Blume
1543    Date: 2000/08/31 17:30:00 JST
1544    Tag: blume-20001017-v110p29p1
1545    Description:
1546    
1547    I made a version 110.29.1 with new bootfiles.
1548    
1549    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
1550              CM documentation changes and a small bugfix in CM's error reporting.
1551    
1552    ----------------------------------------------------------------------
1553    Name: Lal George
1554    Date: 2000/09/27 14:42:35 EDT
1555    Tag: george-20000927-nodestatus
1556    Description:
1557    
1558    Changed the type of the nodestatus, so that:
1559    
1560            SPILLED(~1)             is now SPILLED
1561            SPILLED(m) where m>=0   is now MEMREG(m)
1562            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
1563    
1564    ----------------------------------------------------------------------
1565    Name: Matthias Blume
1566    Date: 2000/09/07 14:45:00 JST
1567    Tag: blume-20000907-cmerrmsg
1568    Description:
1569    
1570    Small tweak to CM to avoid getting ML syntax error messages twice.
1571    
1572    ----------------------------------------------------------------------
1573    Name: Matthias Blume
1574    Date: 2000/08/31 18:00:00 JST
1575    Tag: blume-20000831-cvsbootfiles
1576    Description:
1577    
1578    New URL for boot files (because the 110.29 files on the BL server do
1579    now work correctly with my updated install scripts for yacc and lex).
1580    
1581    ----------------------------------------------------------------------
1582    Name: Matthias Blume
1583    Date: 2000/08/08 12:33:00 JST
1584    Tag: blume-20000808-manual
1585    Description:
1586    
1587    Tiny update to CM manual.
1588    
1589    ----------------------------------------------------------------------
1590    Name: Allen Leung
1591    Date: 2000/08/7 19:31:00
1592    Tag: leunga-20000807-a-whole-bunch-of-stuff
1593    
1594      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
1595    and Mobdy.
1596    
1597    1.  x86
1598    
1599       a.  Fixes to peephole module by John and Dan.
1600       b.  Assembly fix to SETcc by Allen.
1601       c.  Fix to c-call by John.
1602       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
1603       e.  Instruction selection optimization to SETcc as suggested by John.
1604    
1605           For example,
1606    
1607            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
1608    
1609           should generate:
1610    
1611            MOVL a, x
1612            SUBL b, x
1613            SHRL 31, x
1614    
1615    2.  IR stuff
1616    
1617         A bunch of new DJ-graph related algorithms added.  These
1618         speed up SSA construction.
1619    
1620    3.  SSA + Scheduling
1621    
1622         Added code for SSA and scheduling to the repository
1623    
1624    ----------------------------------------------------------------------
1625    Name: Lal George
1626    Date: 2000/07/27 11:53:14 EDT
1627    
1628    Tag: lal-20000727-linux-ppc
1629    Description:
1630    
1631     Made changes to support Linux PPC.
1632     p.s. I have confirmation that the 110.29 boot files work fine.
1633    
1634    ----------------------------------------------------------------------
1635    Name: Matthias Blume
1636    Date: 2000/07/27 17:40:00 JST
1637    Tag: blume-20000727-scripts
1638    Description:
1639    
1640    !!!! WARNING !!!!
1641    You must recompile the runtime system!
1642    !!!! WARNING !!!!
1643    
1644    This is basically another round of script-enhancements:
1645    
1646    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
1647       and undefine CM preprocessor symbols.
1648    
1649    2. ml-build avoids generating a new heap image if it finds that the
1650       existing one is still ok.  (The condition is that no ML file had to
1651       be recompiled and all ML files are found to be older that the heap
1652       file.)
1653    
1654       To make this work smoothly, I also hacked the runtime system as
1655       well as SMLofNJ.SysInfo to get access to the heap image suffix
1656       (.sparc-solaris, ...) that is currently being used.
1657    
1658       Moreover, the signature of CM.mk_standalone has changed.  See the
1659       CM manual.
1660    
1661    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
1662       CM manual for details.)
1663    
1664    4. More CM manual updates:
1665        - all of the above has been documented.
1666        - there is now a section describing the (CM-related) command line
1667          arguments that are accepted by the "sml" command
1668    
1669    ----------------------------------------------------------------------
1670    Name: Matthias Blume
1671    Date: 2000/07/25 16:20:00 JST
1672    Tag: blume-20000725-makedepend
1673    Description:
1674    
1675    Added a script called ml-makedepend.  This can be used in makefiles
1676    for Unix' make in a way very similar to the "makedepend" command for
1677    C.
1678    
1679    The script internally uses function CM.sources.
1680    
1681    Synopsis:
1682    
1683        ml-makedepend [-f makefile] cmfile targetname
1684    
1685    The default for the makefile is "makefile" (or "Makefile" should
1686    "makefile" not exist).
1687    
1688    ml-makedepend adds a cmfile/targetname-specific section to this
1689    makefile (after removing the previous version of this section).  The
1690    section contains a single dependency specification with targetname on
1691    the LHS (targetname is an arbitrary name), and a list of files derived
1692    from the cmfile on the RHS.  Some of the files on the RHS are
1693    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
1694    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
1695    path names.  The makefile writer is responsible for making sure that
1696    these variables have correct at the time "make" is invoked.
1697    
1698    ----------------------------------------------------------------------
1699    Name: Matthias Blume
1700    Date: 2000/07/22 23:30:00 JST
1701    Tag: blume-20000722-urlupdate
1702    Description:
1703    
1704    Changed BOOT and config/srcarchiveurl to point to BL server:
1705    
1706        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
1707    
1708    ----------------------------------------------------------------------
1709    Name: Matthias Blume
1710    Date: 2000/07/18 18:00:00 JST
1711    Tag: blume-20000718-Version_110_29
1712    Description:
1713    
1714    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
1715    
1716    2. Updated config/version to 110.29
1717    
1718    3. Updated config/srcarchiveurl
1719    
1720    3. New boot files!
1721       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
1722    
1723    ----------------------------------------------------------------------
1724    Name: Matthias Blume
1725    Date: 2000/07/11 13:58:00 JST
1726    Tag: blume-20000711-doctypo
1727    Description:
1728    
1729    Fixed a few typos in CM manual.
1730    
1731    ----------------------------------------------------------------------
1732    Name: Allen Leung
1733    Date: 2000/06/15 00:38:00
1734    Tag: leunga-20000704-sparc-x86
1735    
1736    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
1737    2. fix to x86 RET bug [found by Dan Grossman]
1738    3. sparc assembly bug fix for ticc instructions [found by Fermin]
1739    
1740       Affects c-- and moby only
1741    
1742    ----------------------------------------------------------------------
1743    Name: Matthias Blume
1744    Date: 2000/07/04 15:26:00
1745    Tag: blume-20000704-trigger
1746    Description:
1747    
1748    1. Improvements to CM manual.
1749    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
1750       of getting a back-trace.  The function, when called, raises an
1751       internal exception which explicitly carries the full back-trace history,
1752       so it is unaffected by any intervening handle-raise pairs ("trivial"
1753       or not).  The interactive loop will print that history once it arrives
1754       at top level.
1755       Short of having all exceptions implicitly carry the full history, the
1756       recommended way of using this facility is:
1757         - compile your program with instrumentation "on"
1758         - run it, when it raises an exception, look at the history
1759         - if the history is "cut off" because of some handler, go and modify
1760           your program so that it explicitly calls BTrace.trigger
1761         - recompile (still instrumented), and rerun; look at the full history
1762    
1763    ----------------------------------------------------------------------
1764    Name: Matthias Blume
1765    Date: 2000/07/03 15:36:00 JST
1766    Tag: blume-20000702-manual
1767    Description:
1768    
1769    Small corrections and updates to CM manual.
1770    
1771    ----------------------------------------------------------------------
1772    Name: Matthias Blume
1773    Date: 2000/06/29 16:04:00 JST
1774    Tag: blume-20000629-yacctool
1775    Description:
1776    
1777    Changes:
1778    
1779    1. Class "mlyacc" now takes separate arguments to pass options to
1780       generated .sml- and .sig-files independently.
1781    2. Corresponding CM manual updates.
1782    3. BTrace module now also reports call sites.  (However, for loop clusters
1783       it only shows from where the cluster was entered.)  There are associated
1784       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1785    
1786    ----------------------------------------------------------------------
1787    Name: Matthias Blume
1788    Date: 2000/06/27 16:51:00 JST
1789    Tag: blume-20000627-noweb
1790    Description:
1791    
1792    Changes:
1793    
1794     1. Implemented "subdir" and "witness" options for noweb tool.
1795        This caused some slight internal changes in CM's tool implementation.
1796     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
1797        some remaining issues from earlier path anchor changes.
1798     3. Updated CM manual accordingly.
1799    
1800     4. Changed implementation of back-tracing so that I now consider it
1801        ready for prime-time.
1802    
1803        In particular, you don't have to explicitly trigger the back-trace
1804        anymore.  Instead, if you are running BTrace-instrumented code and
1805        there is an uncaught exception (regardless of whether or not it was
1806        raised in instrumented code), the top-level evalloop will print
1807        the back-trace.
1808    
1809        Features:
1810    
1811          - Instrumented and uninstrumented code work together seemlessly.
1812            (Of course, uninstrumented code is never mentioned in actual
1813             back-traces.)
1814    
1815          - Asymptotic time- and space-complexity of instrumented code is
1816            equal to that of uninstrumented code.  (This means that
1817            tail-recursion is preserved by the instrumentation phase.)
1818    
1819          - Modules whose code has been instrumented in different sessions
1820            work together without problem.
1821    
1822          - There is no penalty whatsoever on uninstrumented code.
1823    
1824          - There is no penalty on "raise" expressions, even in
1825            instrumented code.
1826    
1827        A potential bug (or perhaps it is a feature, too):
1828    
1829          A back-trace reaches no further than the outermost instrumented
1830          non-trivial "raise".  Here, a "trivial" raise is one that is the
1831          sole RHS of a "handle" rule.  Thus, back-traces reach trough
1832    
1833               <exp> handle e => raise e
1834    
1835          and even
1836    
1837               <exp> handle Foo => raise Bar
1838    
1839          and, of course, through
1840    
1841               <exp> handle Foo => ...
1842    
1843         if the exception was not Foo.
1844    
1845         Back-traces always reach right through any un-instrumented code
1846         including any of its "handle" expressions, trivial or not.
1847    
1848       To try this out, do the following:
1849    
1850         - Erase all existing binfiles for your program.
1851           (You may keep binfiles for those modules where you think you
1852            definitely don't need back-tracing.)
1853         - Turn on back-trace instrumentation:
1854              SMLofNJ.Internals.BTrace.mode (SOME true);
1855         - Recompile your program.  (I.e., run "CM.make" or "use".)
1856         - You may now turn instrumentation off again (if you want):
1857              SMLofNJ.Internals.BTrace.mode (SOME false);
1858         - Run your program as usual.  If it raises an exception that
1859           reaches the interactive toplevel, then a back-trace will
1860           automatically be printed.  After that, the toplevel loop
1861           will print the exception history as usual.
1862    
1863    ----------------------------------------------------------------------
1864    Name: Matthias Blume
1865    Date: 2000/06/26 09:56:46 JST
1866    Tag: blume-20000626-setup
1867    Description:
1868    
1869    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1870          ML code before and after compiling a file (e.g., to set compiler
1871          flags)
1872    
1873    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1874              - associated changes to btrace.sml (BTrace instrumentation pass)
1875              - cleaner implementation of btimp.sml (BTrace tracing and report
1876                module)
1877    
1878    CM manual: * new path encoding documented
1879               * description of setup-parameter to "sml" added
1880    
1881    The biggest user-visible change to back-tracing is that it is no
1882    longer necessary to compile all traced modules within the same
1883    session.  (This was a real limitation.)
1884    
1885    ----------------------------------------------------------------------
1886    Name: Matthias Blume
1887    Date: 2000/06/24 12:40:00 JST
1888    Tag: blume-20000624-startup
1889    Description:
1890    
1891    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1892    bit too often -- to put it mildly. :)
1893    
1894    ----------------------------------------------------------------------
1895    Name: Matthias Blume
1896    Date: 2000/06/23 18:20:00 JST
1897    Tag: blume-20000623-btrace
1898    Description:
1899    
1900    This updates adds a backtrace facility to aid programmers in debugging
1901    their programs.  This involves the following changes:
1902    
1903    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1904       keeping track of the current call stack.  When programs are compiled
1905       in a special mode, the compiler will insert calls to these hooks
1906       into the user program.
1907       "Hook" means that it is possible for different implementations of
1908       back-tracing to register themselves (at different times).
1909    
1910    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1911       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1912       It can be turned on using this call:
1913         SMLofNJ.Internals.BTrace.mode (SOME true);
1914       Turning it off again:
1915         SMLofNJ.Internals.BTrace.mode (SOME false);
1916       Querying the current status:
1917         SMLofNJ.Internals.BTrace.mode NONE;
1918       Annotated programs are about twice as big as normal ones, and they
1919       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1920       where all hooks do nothing).  The slowdown with a plugin that is
1921       actually useful (such as the one supplied by default) is even greater,
1922       but in the case of the default plugin it is still only an constant
1923       factor (amortized).
1924    
1925    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1926       with a sub-structure BTrace for controlling back-tracing.  In particular,
1927       the above-mentioned function "mode" controls whether the annotation
1928       phase is invoked by the compiler.  Another important function is
1929       "trigger": when called it aborts the current execution and causes
1930       the top-level loop to print a full back-trace.
1931    
1932    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1933       for back-tracing.  It keeps track of the dynamic call stack and in
1934       addition to that it keeps a partial history at each "level" of that
1935       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1936       d and b (at separate times, dynamically), then the report will show:
1937    
1938       GOTO   d
1939             /c
1940       GOTO  \b
1941       CALL   a
1942    
1943       This shows that there was an initial non-tail call of a, then a
1944       tail-call to b or c, looping behavior in a cluster of functions that
1945       consist of b and c, and then a goto from that cluster (i.e., either from
1946       b or from c) to d.
1947    
1948       Note that (depending on the user program) the amount of information
1949       that the back-trace module has to keep track of at each level is bounded
1950       by a constant.  Thus, the whole implementation has the same asymptotical
1951       complexity as the original program (both in space and in time).
1952    
1953    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1954       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1955       which is raised by the "trigger" function mentioned above.
1956    
1957    Notes on usage:
1958    
1959    - Annotated code works well together with unannotated code:
1960    Unannotated calls simply do not show up at all in the backtrace.
1961    
1962    - It is not a good idea to let modules that were annotated during
1963    different sessions run at the same time.  This is because the compiler
1964    chooses small integers to identify individual functions, and there
1965    will be clashes if different modules were compiled in separate sessions.
1966    (Nothing will crash, and you will even be told about the clashes, but
1967    back-trace information will in general not be useful.)
1968    
1969    - Back-tracing can be confused by callcc and capture.
1970    
1971    - The only way of getting a back-trace right now is to explicitly
1972    invoke the "trigger" function from your user program.  Eventually, we
1973    should make every exception carry back-trace information (if
1974    available).  But since this creates more overhead at "raise"-time
1975    (similar to the current exnHistory overhead), I have not yet
1976    implemented this.  (The implementation will be rather easy.)  With
1977    exceptions carrying back-trace information, this facility will be even
1978    more useful because users don't need to modify their programs...
1979    
1980    - While it is possible to compile the compiler with back-trace
1981    annotations turned on (I did it to get some confidence in
1982    correctness), you must make absolutely sure that core.sml and
1983    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1984    be compiled with annotation because there is no core access yet, but
1985    if you compile btimp.sml with annotation, then the system will go into
1986    an infinite recursion and crash.)
1987    Since CM currently does not know about BTrace, the only way to turn
1988    annotations on and off for different modules of the compiler is to
1989    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1990    this is awkward and clumsy.
1991    
1992    Sample sessions:
1993    
1994    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1995    - SMLofNJ.Internals.BTrace.mode (SOME true);
1996    [autoloading]
1997    [autoloading done]
1998    val it = false : bool
1999    - structure X = struct
2000    -     fun main n = let
2001    -         fun a (x, 0) = d x
2002    -           | a (x, n) = b (x, n - 1)
2003    -         and b (x, n) = c (x, n)
2004    -         and c (x, n) = a (x, n)
2005    -         and d x = e (x, 3)
2006    -         and e (x, 0) = f x
2007    -           | e (x, n) = e (x, n - 1)
2008    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
2009    -           | f n = n * g (n - 1)
2010    -         and g n = a (n, 3)
2011    -     in
2012    -         f n
2013    -     end
2014    - end;
2015    structure X : sig val main : int -> int end
2016    - X.main 3;
2017    *** BACK-TRACE ***
2018    GOTO   stdIn:4.2-13.20: X.main[2].f
2019    GOTO-( stdIn:4.2-13.20: X.main[2].e
2020    GOTO   stdIn:4.2-13.20: X.main[2].d
2021         / stdIn:4.2-13.20: X.main[2].a
2022         | stdIn:4.2-13.20: X.main[2].b
2023    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2024    CALL   stdIn:4.2-13.20: X.main[2].g
2025    GOTO   stdIn:4.2-13.20: X.main[2].f
2026    GOTO-( stdIn:4.2-13.20: X.main[2].e
2027    GOTO   stdIn:4.2-13.20: X.main[2].d
2028         / stdIn:4.2-13.20: X.main[2].a
2029         | stdIn:4.2-13.20: X.main[2].b
2030    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2031    CALL   stdIn:4.2-13.20: X.main[2].g
2032    GOTO   stdIn:4.2-13.20: X.main[2].f
2033    GOTO-( stdIn:4.2-13.20: X.main[2].e
2034    GOTO   stdIn:4.2-13.20: X.main[2].d
2035         / stdIn:4.2-13.20: X.main[2].a
2036         | stdIn:4.2-13.20: X.main[2].b
2037    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2038    CALL   stdIn:4.2-13.20: X.main[2].g
2039    GOTO   stdIn:4.2-13.20: X.main[2].f
2040    CALL   stdIn:2.15-17.4: X.main[2]
2041    -
2042    
2043    (Note that because of a FLINt bug the above code currently does not
2044    compile without BTrace turned on.)
2045    
2046    Here is another example, using my modified Tiger compiler:
2047    
2048    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
2049    - SMLofNJ.Internals.BTrace.mode (SOME true);
2050    [autoloading]
2051    [autoloading done]
2052    val it = false : bool
2053    - CM.make "sources.cm";
2054    [autoloading]
2055    ...
2056    [autoloading done]
2057    [scanning sources.cm]
2058    [parsing (sources.cm):parse.sml]
2059    [creating directory CM/SKEL ...]
2060    [parsing (sources.cm):tiger.lex.sml]
2061    ...
2062    [wrote CM/sparc-unix/semant.sml]
2063    [compiling (sources.cm):main.sml]
2064    [wrote CM/sparc-unix/main.sml]
2065    [New bindings added.]
2066    val it = true : bool
2067    - Main.compile ("../testcases/merge.tig", "foo.out");
2068    *** BACK-TRACE ***
2069    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
2070    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2071    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2072    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2073    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2074    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2075    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
2076         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
2077    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
2078    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2079    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
2080    CALL   main.sml:1.18-118.4: Main.compile[2]
2081    -
2082    
2083    ----------------------------------------------------------------------
2084    Name: Matthias Blumen
2085    Date: 2000/06/21 18:00:00 JST
2086    Tag: blume-20000621-manual
2087    Description:
2088    
2089    CM manual update: Path environments documented.
2090    
2091    ----------------------------------------------------------------------
2092    Name: Matthias Blume
2093    Date: 2000/06/19 13:40:00
2094    Tag: blume-20000619-manual
2095    Description:
2096    
2097    CM manual and system/README update.  This only covers the fact that
2098    there are no more implicit anchors.  (Path environments and the "bind"
2099    option to "cm" have yet to be documented.)
2100    
2101    ----------------------------------------------------------------------
2102    Name: Matthias Blume
2103    Date: 2000/06/19 11:05:00 JST
2104    Tag: blume-20000619-chdir-bugfix
2105    Description:
2106    
2107    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
2108    
2109    ----------------------------------------------------------------------
2110    Name: Matthias Blume
2111    Date: 2000/06/18 22:00:10 JST
2112    Tag: blume-20000618-implicit-anchors-really-gone
2113    Description:
2114    
2115    I updates the previous HISTORY entry where I forgot to mention that
2116    implicit anchors are no longer with us.
2117    
2118    The current update also gets rid of the (now useless) controller
2119    CM.Control.implicit_anchors.
2120    
2121    ----------------------------------------------------------------------
2122    Name: Matthias Blume
2123    Date: 2000/06/16 17:30:00 JST
2124    Tag: blume-20000616-anchorenv
2125    Description:
2126    
2127    This patch implements the long anticipated (just kidding :) "anchor
2128    environment" mechanism.  In the course of doing this, I also
2129    re-implemented CM's internal "SrcPath" module from scratch.  The new
2130    one should be more robust in certain boundary cases.  In any case, it
2131    is a lot cleaner than its predecessor (IMHO).
2132    
2133    This time, although there is yet another boot file format change, I
2134    kept the unpickler backward-compatible.  As a result, no new bootfiles
2135    are necessary and bootstrapping is straightforward.  (You cannot read
2136    new bootfiles into an old system, but the other way around is no
2137    problem.)
2138    
2139    Visible changes:
2140    
2141    ** 0. Implicit path anchors (without the leading $-symbol) are no
2142    longer recognized at all. This means that such path names are not
2143    illegal either.  For example, the name basis.cm simply refers to a
2144    local file called "basis.cm" (i.e, the name is an ordinary path
2145    relative to .cm-files directory).  Or, to put it differently, only
2146    names that start with $ are anchored paths.
2147    
2148    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
2149    vanished.
2150    
2151    John (Reppy) had critizised this as soon as I originally proposed and
2152    implemented it, but at that time I did not really deeply believe
2153    him. :) Now I came full-circle because I need the $<singlearc> syntax
2154    in another place where it cannot be seen as an abbreviation for
2155    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
2156    seems to mean (i.e., it "expands" into the corresponding anchor
2157    value).
2158    
2159    However, when paths are used as members in CM description files, it
2160    continues to be true that there must be at least another arc after the
2161    anchor.  This is now enforced separately during semantic analysis
2162    (i.e., from a lexical/syntactical point of view, the notation is ok.)
2163    
2164    ** 2. The "cm" class now accepts an option "bind".  The option's value
2165    is a sub-option list of precisely two items -- one labeled "anchor"
2166    and the other one labeled "value".  As you might expect, "anchor" is
2167    used to specify an anchor name to be bound, and "value" specifies what
2168    the anchor is being bound to.
2169    
2170    The value must be a directory name and can be given in either standard
2171    syntax (including the possibility that it is itself an anchored path)
2172    or native syntax.
2173    
2174    Examples:
2175    
2176       foo.cm (bind:(anchor:bar value:$mystuff/bar))
2177       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
2178    
2179    and so on.
2180    
2181    The meaning of this is that the .cm-file will be processed with an
2182    augmented anchor environment where the given anchor(s) is/are bound to
2183    the given values(s).
2184    
2185    The rationale for having this feature is this: Suppose you are trying
2186    to use two different (already stable) libraries a.cm and b.cm (that
2187    you perhaps didn't write yourself).  Further, suppose each of these
2188    two libraries internally uses its own auxiliary library $aux/lib.cm.
2189    Normally you would now have a problem because the anchor "lib" can not
2190    be bound to more than one value globally.  Therefore, the project that
2191    uses both a.cm and b.cm must locally redirect the anchor to some other
2192    place:
2193    
2194       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
2195       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
2196    
2197    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
2198    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
2199    
2200    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
2201    you when you do so at the time of CM.stabilize).  Therefore, you can
2202    also use an anchored path as the value:
2203    
2204      a.cm (bind:(anchor:lib value:$a-lib))
2205      b.cm (bind:(anchor:lib value:$b-lib))
2206    
2207    Now you can globally configure (using the usual CM.Anchor.anchor or
2208    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
2209    itself is always locally bound, setting it globally is no longer
2210    meaningful or necessary (but it does not hurt either).  In fact, "lib"
2211    can still be used as a global anchor for separate purposes.  As a
2212    matter of fact, one can locally define "lib" in terms of a global
2213    "lib":
2214    
2215      a.cm (bind:(anchor:lib value:$lib/a))
2216      b.cm (bind:(anchor:lib value:$lib/b))
2217    
2218    ** 3: The encoding of path names has changed.  This affects the way
2219    path names are shown in CM's progress report and also the internal
2220    protocol encoding used for parallel make.
2221    
2222    The encoding now uses one or more ':'-separated segments.  Each
2223    segments corresponds to a file that has been specified relative to the
2224    file given by its preceding segment.  The first segment is either
2225    relative to the CWD, absolute, or anchored.  Each segment itself is
2226    basically a Unix pathname; all segments but the first are relative.
2227    
2228    Example:
2229    
2230       $foo/bar/baz.cm:a/b/c.sml
2231    
2232    This path denotes the file bar/a/b/c.sml relative to the directory
2233    denoted by anchor "foo".  Notice that the encoding also includes
2234    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
2235    paths are resolved relative to the .cm-files directory, so baz.cm must
2236    be ignored to get the "real" pathname.
2237    
2238    To make this fact more obvious, CM puts the names of such "virtual
2239    arcs" into parentheses when they appear in progress reports. (No
2240    parentheses will appear in the internal protocol encoding.)  Thus,
2241    what you really see is:
2242    
2243      $foo/bar/(baz.cm):a/b/c.sml
2244    
2245    I find this notation to be much more informative than before.
2246    
2247    Another new feature of the encoding is that special characters
2248    including parentheses, colons, (back)slashes, and white space are
2249    written as \ddd (where ddd is the decimal encoding of the character).
2250    
2251    *** The CM manual still needs to be updated.
2252    
2253    ----------------------------------------------------------------------
2254    Name: Allen Leung
2255    Date: 2000/06/15 00:38:00
2256    Tag: leunga-20000615-x86-peephole
2257    
2258    x86 Peephole fix by Fermin.  Affects c-- and moby only.
2259    
2260    ----------------------------------------------------------------------
2261    Name: Matthias Blume
2262    Date: 2000/06/12 11:40:00
2263    Tag: blume-20000612-parmakefix
2264    Description:
2265    
2266    More cleanup after changing the file naming scheme: This time I
2267    repaired the parallel make mechanism for CMB.make which I broke earlier.
2268    
2269    ----------------------------------------------------------------------
2270    Name: Allen Leung
2271    Date: 2000/06/09 01:25:00
2272    Tag: leunga-20000609-various
2273    
2274    None of these things should affect normal SML/NJ operations
2275    
2276    1. Peephole improvements provided by Fermin (c--)
2277    2. New annotation DEFUSE for adding extra dependence (moby)
2278    3. New X86 LOCK instructions (moby)
2279    4. New machine description language for reservation tables (scheduling)
2280    5. Fixes to various optimization/analysis modules (branch chaining, dominator
2281       trees etc.)
2282    6. I've changed the CM files so that they can work with versions
2283       110.0.6, 110.25 and 110.28
2284    
2285    ----------------------------------------------------------------------
2286    Name: Matthias Blume
2287    Date: 2000/06/09 12:40:00
2288    Tag: blume-20000609-log
2289    Description:
2290    
2291    - Removed all(?) remaining RCS Log entries from sources.
2292    
2293    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
2294      anchored paths).
2295    
2296    ----------------------------------------------------------------------
2297    Name: Matthias Blume
2298    Date: 2000/06/07 17:00:00 JST
2299    Tag: blume-20000607-no-implicit-anchors
2300    Description:
2301    
2302    1. This update changes the default setting for
2303    CM.Control.implicit_anchors from true to false.  This means that
2304    implicit anchors are no longer permitted by default.  I also tried to
2305    make sure that nothing else still relies on implicit anchors.
2306    (This is the next step on the schedule towards a CM that does not even
2307    have the notion of implicit anchors anymore.)
2308    
2309    2. More CM manual updates.
2310    
2311    3. I managed to track down and fix the pickling bug I mentioned last
2312    time.  Because of the previously existing workaround, this entails no
2313    immediate practical changes.
2314    
2315    ----------------------------------------------------------------------
2316    Name: Matthias Blume
2317    Date: 2000/06/06 11:15:00 JST
2318    Tag: blume-20000606-lazierpickle
2319    Description:
2320    
2321    !!!! NEW BOOT FILES !!!!
2322    
2323    * The main purpose of this update is to make library pickles lazier in
2324    order to reduce the initial space penalty for autoloading a library.
2325    As a result, it is now possible to have $smlnj/compiler.cm
2326    pre-registered.  This should take care of the many complaints or
2327    inquiries about missing structure Compiler.  This required changes to
2328    CM's internal data structures and small tweaks to some algorithms.
2329    
2330    As a neat additional effect, it is no longer necessary (for the sake
2331    of lean heap image files) to distinguish between a "minimal" CM and a
2332    "full" CM.  Now, there is only one CM (i.e., the "full" version:
2333    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
2334    interactive top level. ($smlnj/cm/minimal.cm is gone.)
2335    
2336    To make the life of compiler-hackers easier, "makeml" now also
2337    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
2338    words, after you bootstrap a new sml for the first time, you will not
2339    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
2340    around you will still have to do it, though.)
2341    
2342    * A second change consists of major updates to the CM manual.  There
2343    are now several appendices with summary information and also a full
2344    specification of the CM description file syntax.
2345    
2346    * In directory src/system I added the script "allcross".  This script
2347    invokes sml and cross-compiles the compiler for all supported
2348    architectures.  (Useful when providing a new set of boot files.)
2349    
2350    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
2351    added a small tweak to pickle-util.sml to work around this problem,
2352    but it is not a proper fix yet.  I will investigate further.  (The
2353    effect of the bug was an inflation of library pickle size.)
2354    
2355    * Version number increased to 110.28.1 (to avoid compatibility problems).
2356    
2357    ----------------------------------------------------------------------
2358    Name: Allen Leung
2359    Date: 2000/05/25 17:28 EDT
2360    Tag: leunga-20000525-ra
2361    Description:
2362    
2363      Fixed a bug in freezing phase of the register allocator.
2364    
2365    ----------------------------------------------------------------------
2366    Name: Allen Leung
2367    Date: 2000/05/15 22:53 EDT
2368    Tag: leunga-20000515-alpha-x86-ra
2369    Description:
2370    
2371      1. Alpha
2372    
2373          Slight cleanup.  Removed the instruction SGNXL
2374    
2375      2. X86
2376    
2377          Added the following instructions to the instruction set:
2378    
2379            ROLx, RORx,
2380            BTx, BTSx, BTLx, BTRx,
2381            XCHGx, and variants with the LOCK prefix
2382    
2383      3. Register Allocation
2384    
2385          The module ra-rewrite-with-renaming has been improved.
2386    
2387      These have no effect on SML/NJ.
2388    
2389    ----------------------------------------------------------------------
2390    Name: Matthias Blume
2391    Date: 2000/05/15 16:20:00 JST
2392    Tag: blume-20000515-lightrebuild
2393    Description:
2394    
2395    1. I added an alternative to "-rebuild" to "makeml".  The difference is
2396       that prior to calling CMB.make' the CM-variable "LIGHT" will be
2397       defined.  In effect, the command will not build any cross-compiler
2398       backends and therefore finish more quickly.
2399    
2400       The "fixpt" script also takes a "-light" switch to be able to use
2401       this new facility while compiling for a fixpoint.
2402    
2403    2. I replaced all mentions of anchored paths in group owner specifications
2404       with simple relative paths (usually starting with "..").
2405       The rationale is that a library's internal workings should not be
2406       compromised by the lack of some anchor.  (An anchor is necessary
2407       for someone who wants to refer to the library by an anchored path,
2408       but it should not be necessary to build the same library in the first
2409       place.)
2410    
2411    3. I changed the way CM's tool mechanism determines the shell command
2412       string used for things like ml-yacc etc. so that it does not break
2413       when CM.Control.implicit_anchors is turned off.
2414    
2415    ----------------------------------------------------------------------
2416    Name: Matthias Blume
2417    Date: 2000/05/12 18:20:00 JST
2418    Tag: blume-20000512-ml-build
2419    Description:
2420    
2421    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
2422    getting installed properly (by config/install.sh).
2423    
2424    ----------------------------------------------------------------------
2425    Name: Matthias Blume
2426    Date: 2000/05/12 17:30:00 JST
2427    Tag: blume-20000512-anchors
2428    Description:
2429    
2430    !!! NEW BOOT FILES !!!
2431    
2432    This change is in preparation of fading out support for "implicitly
2433    anchored path names".  I went through all sources and used the
2434    explicit (and relatively new) $-notation.  See system/README and the
2435    CM manual for more info on this.
2436    
2437    I also modified the anchoring scheme for some things such as "smlnj",
2438    "MLRISC", "cm", etc. to take advantage of the fact that explicit
2439    anchors are more expressive: anchor name and first arc do not have to
2440    coincide.  This entails the following user-visible change:
2441    
2442    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
2443    particular, when you fire up sml with a command-line argument, say,
2444    e.g.:
2445    
2446       sml '$smlnj/cmb.cm'
2447    
2448    At the ML toplevel prompt:
2449    
2450       CM.autoload "$smlnj/cmb.cm";
2451    
2452    There is also a new controller in CM.Control that can be used to turn
2453    off all remaining support for implicit anchors by saying:
2454    
2455        CM.autoload "$smlnj/
2456        #set CM.Control.implicit_anchors false;
2457    
2458    This causes CM to reject implicitly anchored paths.  This is (for the
2459    time being) less permissive than the "final" version where there will
2460    be no more such implicit anchors and relative paths will be just that:
2461    relative.
2462    
2463    The next step (version after next version?) will be to make the
2464    default for CM.Control.implicit_anchors false.  After the dust has
2465    settled, I can then produce the "final" version of this...
2466    
2467    Note: Since bootstrapping is a bit tricky, I provided new boot files.
2468    
2469    ----------------------------------------------------------------------
2470    Name: Matthias Blume
2471    Date: 2000/05/11 16:30:00 JST
2472    Tag: blume-20000511-sources
2473    Description:
2474    
2475    The main change is that I added function CM.sources as a generalized
2476    version of the earlier CM.makedepend.  This entails the following
2477    additional changes:
2478    
2479      - CM.makedepend has been dropped.
2480    
2481      - CM manual has been updated.
2482    
2483      - TOOLS signature and API have been changed.
2484    
2485    ----------------------------------------------------------------------
2486    Name: Allen Leung
2487    Date: 2000/05/10 21:17 EDT
2488    Tag: leunga-20000510-moby-c--ssa
2489    Description:
2490    
2491      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
2492    None of these affect SML/NJ.
2493    
2494    1. Register Allocation
2495    
2496        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
2497           This module tries to remove local (i.e. basic block level) redundancies
2498           during spilling.
2499    
2500        b. A new framework for performing region based register allocation.
2501           Not yet entirely functional.
2502    
2503    2. X86
2504    
2505       a. DefUse for POP was missing the stack pointer [found by Lal]
2506       b. Reload for CALL was incorrect in X86Spill [found by John]
2507       c. Various fixes in X86Spill so that it can be used correctly for
2508          the new spilling module.
2509    
2510    3. SSA/IR
2511    
2512       a. New module ir/dj-dataflow.sml implements elimination based
2513          data flow analysis.
2514    
2515    4. MLRiscGen
2516    
2517       a. Fix for gc type annotation
2518    
2519    5. MDGen
2520    
2521       Various fixes for machine description -> ml code translation.  For ssa
2522       only.
2523    
2524    ----------------------------------------------------------------------
2525    Name: Allen Leung
2526    Date: 2000/05/08 22:17 EDT
2527    Tag: leunga-20000508-labexp
2528    Description:
2529    
2530      Fermin has found a few assembly problems with constant expressions
2531      generated in LabelExp.  Mostly, the problems involve extra parentheses,
2532      which choke on dumb assemblers.  This is his fix.
2533    
2534    ----------------------------------------------------------------------
2535    Name: Dave MacQueen
2536    Date: 2000/04/09 14:00 EDT
2537    Tag: dbm-20000502-Version_110_28
2538    Description:
2539    
2540    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
2541    
2542    2. Updated config/version to 110.28
2543    
2544    3. Updated config/srcarchiveurl
2545    
2546    3. New boot files!
2547       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
2548    
2549    ----------------------------------------------------------------------
2550    Name: Matthias Blume
2551    Date: 2000/05/01 19:05:00 JST
2552    Tag: blume-20000501-noweb
2553    Description:
2554    
2555    A new noweb tool has been added.  The existing system is entirely
2556    unaffected by this, but some CM users have asked for renewed noweb
2557    support.  Everything is documented in the CM manual.
2558    
2559    New (plugin) libraries:
2560    
2561       noweb-tool.cm
2562       nw-ext.cm
2563    
2564    ----------------------------------------------------------------------
2565    Name: Dave MacQueen
2566    Date: 2000/04/30 12:40PM EDT
2567    Tag: dbm-20000430-bug_fixes
2568    Description:
2569    
2570    1. Fix for bug 1498
2571       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
2572       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
2573         added toRealArray function
2574       smlnj/src/compiler/MiscUtil/print/ppobj.sml
2575         added check for tag Obj.RealArray to array printing case in ppObj
2576    
2577    2. Fix for bug 1510
2578       smlnj/src/compiler/Semant/types/typesutil.sml
2579         fixed definition of dummyargs (used by equalTycon) so that
2580         dummy args are distinct types
2581    
2582    ----------------------------------------------------------------------
2583    Name: Matthias Blume
2584    Date: 2000/04/30 01:00:00 JST
2585    Tag: blume-20000430-versions
2586    Description:
2587    
2588    1. CM version numbering added.  This is an implementation of Lal's
2589       proposal for adding version numbers and version checking to .cm
2590       files.  Lal said that his proposal was just that -- a proposal.
2591       For the time being I went ahead and implemented it so that people
2592       can comment on it.  Everything is completely backward-compatible
2593       (except for the stable library format, i.e., new bootfiles!).
2594    
2595       As usual, see the CM manual for details.
2596    
2597    2. An alternative syntax for anchored paths has been implemented.
2598       Dave has recently voiced the same concerns that I had when I did
2599       this, so there should be some support.  My take is that eventually
2600       I will let support for the current syntax (where anchors are
2601       "implicit") fade out in favor of the new, explicit syntax.
2602       In order to be backward-compatible, both old and new syntax are
2603       currently supported.
2604    
2605       Again, see the CM manual for details.
2606    
2607    3. Parallel make is trying to be slightly smarter:  When the master
2608       process finds a "bottleneck", i.e., when there is only one
2609       compilation unit that can be compiled and everybody else is
2610       waiting on it, then it will simply compile it directly instead
2611       of clumsily telling one of the slaves to do it.
2612    
2613    4. Support for "unsharing" added.  This is necessary in order to be
2614       able to have two different versions of the same library running
2615       at the same time (e.g., for trying out a new MLRISC while still
2616       having the old MLRISC linked into the current compiler, etc.)
2617       See the CM manual.
2618    
2619    5. Simple "makedepend" functionality added for generating Makefile
2620       dependency information.  (This is rather crude at the moment.
2621       Expect some changes here in the future.)
2622    
2623    6. ".fun" added as a recognized suffix for ML files. Also documented
2624       explicitly in the manual that the fallback behavior (unknown suffix
2625       -> ML file) is not an official feature!
2626    
2627    7. Small changes to the pickler for stable libraries.
2628    
2629    8. Several internal changes to CM (for cleanup/improvement).
2630    
2631    
2632    !!!! NEW BINFILES !!!!
2633    
2634    ----------------------------------------------------------------------
2635    Name: Matthias Blume
2636    Date: 2000/04/28 17:30:00 JST
2637    Tag: blume-20000428-pathconfig
2638    Description:
2639    
2640    1. I changed config/install.sh to remove duplicate entries from the
2641       lib/pathconfig file at the end.  Moreover, the final version of
2642       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
2643       in src/system/installml.
2644    
2645    2. The config/install.sh script now consistently uses relative
2646       pathnames in lib/pathconfig whenever the anchor is in the lib
2647       directory.  (So far this was true for the libraries that come
2648       pre-compiled and bundled as part of the bootfiles but not for
2649       libraries that are compiled by the script itself.)
2650    
2651    ----------------------------------------------------------------------
2652    Name: Matthias Blume
2653    Date: 2000/04/26 13:10:00 JST
2654    Tag: blume-20000426-fun_suffix
2655    Description:
2656    
2657    Added ".fun" as a recognized file name suffix (for ML code).
2658    
2659    ----------------------------------------------------------------------
2660    Name: Allen Leung
2661    Date: 2000/04/25 17:00:00 EST
2662    Tag: leunga-20000425-alpha-ra
2663    Description:
2664    
2665    1. Alpha
2666    
2667        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
2668    in C--.
2669    
2670    2. RA
2671    
2672       Added a flag "ra-dump-size" to print out the size of the flowgraph
2673       and the interference graph.
2674    
2675    ----------------------------------------------------------------------
2676    Name: Dave MacQueen
2677    Date: 2000/04/25/
2678    Tag: dbm-20000425-mlyacc_doc_examples
2679    Description:
2680      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
2681      Updated all three examples in src/ml-yacc/examples to run
2682      under 110.* using CM.make.
2683    
2684    ----------------------------------------------------------------------
2685    Name: Allen Leung
2686    Date: 2000/04/20 23:04:00 EST
2687    Tag: leunga-20000420-ssa-c---stuff
2688    Description:
2689    
2690      This update synchronizes my repository with Yale's.  Most of these
2691    changes, however, do not affect SML/NJ at all (the RA is an exception).
2692    
2693    1. Register Allocator
2694    
2695       a. An improvement in the interference graph construction:
2696          Given a copy
2697    
2698                s <- t
2699    
2700          no interference edge between s and t is added for this definition of s.
2701    
2702       b. I've added two new spill heuristic modules that Fermin and I developed
2703          (in the new library RA.cm). These are unused in SML/NJ but maybe
2704          useful for others (Moby?)
2705    
2706    2. X86
2707    
2708       a. Various fixes in the backend provided by Fermin [C--] and Lal.
2709    
2710    3. Alpha
2711    
2712       a. Added the BSR instruction and code generation that goes with it [C--]
2713       b. Other fixes too numerous to recount provided by Fermin [C--]
2714    
2715    4. Regmaps
2716    
2717       a. The regmaps are not initialized with the identity physical bindings
2718          at creation time.  This is unneeded.
2719    
2720    5. MLRISC Optimizations
2721    
2722       a. The DJ-Graph module can now compute the iterated dominance frontiers
2723          intersects with liveness incrementally in linear time! Woohoo!
2724          This is now used in my new SSA construction algorithm.
2725    
2726       b. THe branch reorganization module is now smarter about linear chains of
2727          basic blocks.
2728    
2729    
2730    ----------------------------------------------------------------------
2731    Name: Matthias Blume
2732    Date: 2000/04/12 13:52:00 JST
2733    Tag: blume_main_v110p27_1
2734    Description:
2735    
2736    Changed install.sh script to handle archive files without version number
2737    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
2738    name of the boot file archive.
2739    
2740    ----------------------------------------------------------------------
2741    Name: Dave MacQueen
2742    Date: 2000/04/09 14:00 EDT
2743    Tag: dbm-20000410-Version_110_27
2744    Description:
2745    
2746    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
2747    
2748    2. Updated src/config/version to 110.27
2749    
2750    3. New boot files!
2751    
2752    ----------------------------------------------------------------------
2753    Name: Allen Leung
2754    Date: 2000/04/09 19:09:00 EST
2755    Tag: leunga-20000409-misc
2756    Description:
2757    
2758    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
2759    
2760    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
2761    
2762    ----------------------------------------------------------------------
2763    Name: Stefan
2764    Date: 2000/04/07 10:00:00 EDT
2765    Tag: monnier-20000406-branch-handling
2766    Description:
2767    
2768    Improved handling of branches (mostly those generated from
2769    polymorphic equality), removed switchoff and changed the
2770    default optimization settings (more cpsopt and less flintopt).
2771    
2772    ----------------------------------------------------------------------
2773    Name: Allen Leung
2774    Date: 2000/04/06 01:30:00 EST
2775    Tag: leunga-20000406-peephole-x86-SSA-2
2776    Description:
2777    
2778       Forgot a few files.
2779    
2780    ----------------------------------------------------------------------
2781    Name: Allen Leung
2782    Date: 2000/04/06 00:36:00 EST
2783    Tag: leunga-20000406-peephole-x86-SSA
2784    Description:
2785    
2786    1.  New Peephole code
2787    
2788    2.  Minor improvement to X86 instruction selection
2789    
2790    3.  Various fixes to SSA and machine description -> code translator
2791    
2792    ----------------------------------------------------------------------
2793    Name: Matthias Blume
2794    Date: 2000/04/05 12:30:00 JST
2795    Tag: blume_main_v110p26p2_3
2796    Description:
2797    
2798    This update just merges three minor cosmetic updates to CM's sources
2799    to get ready for the 110.27 code freeze on Friday.  No functionality
2800    has changed.
2801    
2802  ----------------------------------------------------------------------  ----------------------------------------------------------------------
2803  Name: Allen Leung  Name: Allen Leung
2804  Date: 2000/04/04 19:39:00 EST  Date: 2000/04/04 19:39:00 EST
# Line 57  Line 2845 
2845    
2846  3.  Assembly  3.  Assembly
2847    
2848      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
2849      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
2850      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
2851    
# Line 80  Line 2868 
2868    
2869  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
2870  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
2871  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
2872  legal surface-syntax symbol).  legal surface-syntax symbol).
2873    
2874  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 90  Line 2878 
2878  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
2879    
2880  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2881  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
2882  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
2883  more details.  more details.
2884    
# Line 205  Line 2993 
2993     (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
2994     used).     used).
2995    
2996     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
2997     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
2998     Makefiles or Shell scripts.     Makefiles or Shell scripts.
2999    
3000     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 218  Line 3006 
3006    
3007     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
3008    
3009     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
3010     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
3011    
3012  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 249  Line 3037 
3037       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3038       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3039       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
3040       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3041       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3042       Dave's hands. :)       Dave's hands. :)
3043    
# Line 352  Line 3140 
3140    
3141       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
3142       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
3143       testl is perferable to andl.       testl is preferable to andl.
3144    
3145  7. RA (x86 only)  7. RA (x86 only)
3146    
# Line 534  Line 3322 
3322    
3323  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
3324  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
3325  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
3326  explanation.  explanation.
3327    
3328  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
3329  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
3330  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
3331  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 581  Line 3369 
3369  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
3370  <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,
3371  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
3372  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
3373  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
3374  and, thus, saves a lot of time.  and, thus, saves a lot of time.
3375    
# Line 631  Line 3419 
3419    it from that remote directory.    it from that remote directory.
3420    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
3421    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
3422    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
3423    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
3424    
3425    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.594  
changed lines
  Added in v.939

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