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 801, Mon Mar 19 22:53:00 2001 UTC revision 987, Wed Nov 21 21:14:44 2001 UTC
# Line 14  Line 14 
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16  Name: Matthias Blume  Name: Matthias Blume
17    Date: 2001/11/21 16:14:00 EST
18    Tag: blume-20011121-forgottenfile
19    Description:
20    
21    Forgot to add a file.  (Just a .tex-file -- part of
22    the CM manual source.)
23    
24    ----------------------------------------------------------------------
25    Name: Matthias Blume
26    Date: 2001/11/21 16:10:00 EST
27    Tag: Release_110_37
28    Description:
29    
30    1. Modifications to the SML/NJ code generator and to the runtime system
31       so that code object name strings are directly inserted into code
32       objects at code generation time.  The only business the runtime system
33       has with this is now to read the name strings on occasions.
34       (The encoding of the name string has also changed somewhat.)
35    
36    2. CM now implements a simple "set calculus" for specifying export lists.
37       In particular, it is now possible to refer to the export lists of
38       other libraries/groups/sources and form unions as well as differences.
39       See the latest CM manual for details.
40    
41    3. An separate notion of "proxy" libraries has again be eliminated from
42       CM's model.  (Proxy libraries are now simply a special case of using
43       the export list calculus.)
44    
45    4. Some of the existing libraries now take advantage of the new set
46       calculus.
47       (Notice that not all libraries have been converted because some
48       of the existing .cm-files are supposed to be backward compatible
49       with 110.0.x.)
50    
51    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
52       "General.exnMessage"!  The former relies on a certain hook to be
53       initialized, and that often does not happen in the stand-alone case.)
54    
55    ----------------------------------------------------------------------
56    Name: Lal George
57    Date: 2001/11/21  13:56:18 EST
58    Tag: george-2001121-pseudo-ops
59    Description:
60    
61      Implemented a complete redesign of MLRISC pseudo-ops. Now there
62      ought to never be any question of incompatabilities with
63      pseudo-op syntax expected by host assemblers.
64    
65      For now, only modules supporting GAS syntax are implemented
66      but more should follow, such as MASM, and vendor assembler
67      syntax, e.g. IBM as, Sun as, etc.
68    
69    ----------------------------------------------------------------------
70    Name: Matthias Blume
71    Date: 2001/11/14 11:52:00 EST
72    Tag: blume-20011114-srcname
73    Description:
74    
75    1. Routed the name of the current source file to mlriscgen where it
76       should be directly emitted into the code object.  (This last part
77       is yet to be done.)
78    
79    2. Some cleanup of the pgraph code to make it match the proposal that
80       I put out the other day.  (The proposal notwithstanding, things are
81       still in flux here.)
82    
83    ----------------------------------------------------------------------
84    Name: Lal George
85    Date: 2001/11/14 09:44:04 EST
86    Tag:
87    Description:
88    
89      Fix for a backpatching bug reported by Allen.
90    
91      Because the boundary between short and long span-dependent
92      instructions is +/- 128, there are an astounding number of
93      span-dependent instructions whose size is over estimated.
94    
95      Allen came up with the idea of letting the size of span
96      dependent instructions be non-monotonic, for a maxIter
97      number of times, after which the size must be monotonically
98      increasing.
99    
100      This table shows the number of span-dependent instructions
101      whose size was over-estimated as a function of maxIter, for the
102      file Parse/parse/ml.grm.sml:
103    
104         maxIter            # of instructions:
105            10                      687
106            20                      438
107            30                      198
108            40                        0
109    
110      In compiling the compiler, there is no significant difference in
111      compilation speed between maxIter=10 and maxIter=40. Actually,
112      my measurements showed that maxIter=40 was a tad faster than
113      maxIter=10! Also 96% of the  files in the compiler reach a fix
114      point within 13 iterations, so fixing maxIter at 40, while high,
115      is okay.
116    
117    ----------------------------------------------------------------------
118    Name: Matthias Blume
119    Date: 2001/10/31 15:25:00 EST
120    Tag: blume-20011031-pgraph
121    Description:
122    
123    CKIT:
124    * Changed the "Function" constructor of type Ast.ctype to carry optional
125      argument identifiers.
126    * Changed the return type of TypeUtil.getFunction accordingly.
127    * Type equality ignores the argument names.
128    * TypeUtil.composite tries to preserve argument names but gives up quickly
129      if there is a mismatch.
130    
131    installation script:
132    * attempts to use "curl" if available (unless "wget" is available as well)
133    
134    CM:
135    * has an experimental implementation of "portable graphs" which I will
136      soon propose as an implementation-independent library format
137    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
138    
139    NLFFI-LIB:
140    * some cleanup (all cosmetic)
141    
142    NLFFIGEN:
143    * temporarily disabled the mechanism that suppresses ML output for
144      C definitions whose identifiers start with an underscore character
145    * generate val bindings for enum constants
146    * user can request that only one style (light or heavy) is being used;
147      default is to use both (command-line arguments: -heavy and -light)
148    * fixed bug in handling of function types involving incomplete pointers
149    * generate ML entry points that take record arguments (i.e., using
150      named arguments) for C functions that have a prototype with named
151      arguments
152      (see changes to CKIT)
153    
154    ----------------------------------------------------------------------
155    Name: Allen Leung
156    Date: 2001/10/27 20:34:00 EDT
157    Tag: leunga-20011027-x86-fast-fp-call
158    Description:
159    
160       Fixed the bug described in blume-20010920-slowfp.
161    
162       The fix involves
163          1. generating FCOPYs in FSTP in ia32-svid
164          2. marking a CALL with the appropriate annotation
165    
166    ----------------------------------------------------------------------
167    Name: Matthias Blume
168    Date: 2001/10/16 11:32:00 EDT
169    Tag: blume-20011016-netbsd
170    Description:
171    
172    Underscore patch from Chris Richards (fixing problem with compiling
173    runtime system under recent NetBSD).
174    
175    ----------------------------------------------------------------------
176    Name: Allen Leung
177    Date: 2001/10/12 17:18:32 EDT 2001
178    Tag: leung-20011012-x86-printflowgraph
179    Description:
180    
181    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
182    
183    ----------------------------------------------------------------------
184    Name: Lal George
185    Date: 2001/10/11 23:51:34 EDT
186    Tag: george-20011011-too-many-instrs
187    Description:
188    
189    The representation of a program point never expected to see more
190    than 65536 instructions in a basic block!
191    
192    ----------------------------------------------------------------------
193    Name: Lal George
194    Date: 2001/10/09 09:41:37 EDT
195    Tag: george-20011008-mlrisc-labels
196    Description:
197    
198    Changed the machine description files to support printing of
199    local and global labels in assembly code, based on host assembler
200    conventions.
201    
202    ----------------------------------------------------------------------
203    Name: Matthias Blume
204    Date: 2001/09/25 15:25:00 EDT
205    Tag: blume-20010925-exninfo
206    Description:
207    
208    I provided a non-hook implementation of exnName (at the toplevel) and
209    made the "dummy" implementation of exnMessage (at the toplevel) more
210    useful: if nothing gets "hooked in", then at least you are going to
211    see the exception name and a message indicating why you don't see more.
212    
213    [For the time being, programs that need exnMessage and want to use
214    ml-build should either use General.exnMessage (strongly recommended) or
215    refer to structure General at some other point so that CM sees a
216    static dependency.]
217    
218    [Similar remarks go for "print" and "use":  If you want to use their
219    functionality in stand-alone programs generated by ml-build, then use
220    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
221    
222    ----------------------------------------------------------------------
223    Name: Matthias Blume
224    Date: 2001/09/20 17:28:00 EDT
225    Tag: blume-20010920-slowfp
226    Description:
227    
228    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
229    
230    ----------------------------------------------------------------------
231    Name: Matthias Blume
232    Date: 2001/09/20 17:20:00 EDT
233    Tag: blume-20010920-canonicalpaths
234    Description:
235    
236    0. Updated the BOOT file (something that I forgot to do earlier).
237    
238    1. Small internal change to CM so that it avoids "/../" in filenames
239       as much as possible (but only where it is safe).
240    
241    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
242       that arguments that contain delimiters are passed through correctly.
243       This change also means that all "special" arguments of the form
244       @SMLxxx... must come first.
245    
246    3. Changed install script to put relative anchor names for tool commands
247       into pathconfig.
248    
249    ----------------------------------------------------------------------
250    Name: Matthias Blume
251    >>>>>>> 1.169
252    Date: 2001/09/18 15:35:00 EDT
253    Tag: blume-20010918-readme11036
254    Description:
255    
256    Added README files.
257    
258    ----------------------------------------------------------------------
259    Name: Matthias Blume
260    Date: 2001/09/18 11:45:00 EDT
261    Tag: Release_110_36 (retag)
262    Description:
263    
264    Fixed mistake in config/preloads. Retagged as 110.36.
265    
266    ----------------------------------------------------------------------
267    Name: Matthias Blume
268    Date: 2001/09/18 09:40:00 EDT
269    Tag: Release_110_36_orig (tag changed)
270    Description:
271    
272    New version (110.36).  New bootfiles.
273    
274    ----------------------------------------------------------------------
275    Name: Matthias Blume
276    Date: 2001/09/14 16:15:00 EDT
277    Tag: blume-20010914-x86fastfp
278    Description:
279    
280    John committed some changes that Allen made, in particular a (hopefully)
281    correctly working version of the x86-fp module.
282    
283    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
284    flag to "true".  Everything seems to compile to a fixpoint ok, and
285    "mandelbrot" speeds up by about 15%.
286    
287    ----------------------------------------------------------------------
288    Name: Matthias Blume
289    Date: 2001/09/13 11:20:00 EDT
290    Tag: blume-20010913-minimal
291    Description:
292    
293    1. Stefan Monnier's patch to fix a miscompilation problem that
294       was brought to light by John Reppy's work on Moby.
295    
296    2. Implemented a minimal "structure Compiler" that contains just
297       "version" and "architecture".  The minimal version will be
298       available when the full version is not.  This is for backward-
299       compatibility with code that wants to test Compiler.version.
300    
301    ----------------------------------------------------------------------
302    Name: Matthias Blume
303    Date: 2001/08/28 14:03:00 EDT
304    Tag: blume-20010828-ml-lex
305    Description:
306    
307    Fix for bug 1581, received from Neophytos Michael.
308    
309    ----------------------------------------------------------------------
310    Name: Matthias Blume
311    Date: 2001/08/27 11:20:00 EDT
312    Tag: blume-20010827-readme11035
313    Description:
314    
315    Fleshed out the README file for 110.35.
316    
317    ----------------------------------------------------------------------
318    Name: Matthias Blume
319    Date: 2001/08/24 17:10:00 EDT
320    Tag: Release_110_35
321    Description:
322    
323    New version number (110.35).  New bootfiles.
324    
325    ----------------------------------------------------------------------
326    Name: Lal George
327    Date: 2001/08/24 13:47:18 EDT 2001
328    Tag: george-20010824-MLRISC-graphs
329    Description:
330    
331     removed clusters from MLRISC completely and replaced with graphs.
332    
333    ----------------------------------------------------------------------
334    Name: Matthias Blume
335    Date: 2001/08/23 17:50:00 EDT
336    Tag: blume-20010823-toplevel
337    Description:
338    
339    - some reorganization of the code that implements various kinds of
340      environments in the compiler (static, dynamic, symbolic, combined)
341    - re-implemented the EnvRef module so that evalStream works properly
342      (if the stream contains references to "use", "CM.make", etc.)
343    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
344    
345    ----------------------------------------------------------------------
346    Name: Matthias Blume
347    Date: 2001/08/20 15:50 EDT
348    Tag: blume20010820-slipup
349    Description:
350    
351    I forgot to commit a few files.  Here they are...
352    
353    ----------------------------------------------------------------------
354    Name: Matthias Blume
355    Date: 2001/08/20 15:35:00 EDT
356    Tag: blume-20010820-debugprof
357    Description:
358    
359    !!!! NEW BOOTFILES !!!!
360    
361    This is another round of reorganizing the compiler sources.  This
362    time the main goal was to factor out all the "instrumentation"
363    passes (for profiling and backtracing) into their own library.
364    The difficulty was to do it in such a way that it does not depend
365    on elaborate.cm but only on elabdata.cm.
366    
367    Therefore there have been further changes to both elaborate.cm and
368    elabdata.cm -- more "generic" things have been moved from the former
369    to the latter.  As a result, I was forced to split the assignment
370    of numbers indicating "primtyc"s into two portions: SML-generic and
371    SML/NJ-specific.  Since it would have been awkward to maintain,
372    I bit the bullet and actually _changed_ the mapping between these
373    numbers and primtycs.  The bottom line of this is that you need
374    a new set of bin- and bootfiles.
375    
376    I have built new bootfiles for all architectures, so doing a fresh
377    checkout and config/install.sh should be all you need.
378    
379    The newly created library's name is
380    
381        $smlnj/viscomp/debugprof.cm
382    
383    and its sources live under
384    
385        src/compiler/DebugProf
386    
387    ----------------------------------------------------------------------
388    Name: Matthias Blume
389    Date: 2001/08/15 17:15:00 EDT
390    Tag: blume-20010815-compreorg
391    Description:
392    
393    This is a first cut at reorganizing the CM libraries that make up the
394    core of the compiler.  The idea is to separate out pieces that could
395    be used independently by tools, e.g., the parser, the typechecker, etc.
396    
397    The current status is a step in this direction, but it is not quite
398    satisfactory yet.  Expect more changes in the future.
399    
400    Here is the current (new) organization...
401    
402        What used to be $smlnj/viscomp/core.cm is now divided into
403        six CM libraries:
404    
405             $smlnj/viscomp/basics.cm
406                           /parser.cm
407                           /elabdata.cm
408                           /elaborate.cm
409                           /execute.cm
410                           /core.cm
411    
412        The CM files for these libraries live under src/system/smlnj/viscomp.
413        All these libraries are proxy libraries that contain precisely
414        one CM library component.  Here are the locations of the components
415        (all within the src/compiler tree):
416    
417             Basics/basics.cm
418             Parse/parser.cm
419             ElabData/elabdata.cm
420             Elaborator/elaborate.cm
421             Execution/execute.cm
422             core.cm
423    
424         [This organization is the same that has been used already
425         for a while for the architecture-specific parts of the visible
426         compiler and for the old version of core.cm.]
427    
428         As you will notice, many source files have been moved from their
429         respective original locations to a new home in one of the above
430         subtrees.
431    
432         The division of labor between the new libraries is the following:
433    
434             basics.cm:
435                - Simple, basic definitions that pertain to many (or all) of
436                  the other libraries.
437             parser.cm:
438                - The SML parser, producing output of type Ast.dec.
439                - The type family for Ast is also defined and exported here.
440             elabdata.cm:
441                - The datatypes that describe input and output of the elaborator.
442                  This includes types, absyn, and static environments.
443             elaborator.cm:
444                - The SML/NJ type checker and elaborator.
445                  This maps an Ast.dec (with a given static environment) to
446                  an Absyn.dec (with a new static environment).
447                - This libraries implements certain modules that used to be
448                  structures as functors (to remove dependencies on FLINT).
449             execute.cm:
450                - Everything having to do with executing binary code objects.
451                - Dynamic environments.
452             core.cm:
453                - SML/NJ-specific instantiations of the elaborator and MLRISC.
454                - Top-level modules.
455                - FLINT (this should eventually become its own library)
456    
457    Notes:
458    
459    I am not 100% happy with the way I separated the elaborator (and its
460    data structures) from FLINT.  Two instances of the same problem:
461    
462        1. Data structures contain certain fields that carry FLINT-specific
463           information.  I hacked around this using exn and the property list
464           module from smlnj-lib.  But the fact that there are middle-end
465           specific fields around at all is a bit annoying.
466    
467        2. The elaborator calculates certain FLINT-related information.  I tried
468           to make this as abstract as I could using functorization, but, again,
469           the fact that the elaborator has to perform calculations on behalf
470           of the middle-end at all is not nice.
471    
472        3. Having to used exn and property lists is unfortunate because it
473           weakens type checking.  The other alternative (parameterizing
474           nearly *everything*) is not appealing, though.
475    
476    I removed the "rebinding =" warning hack because due to the new organization
477    it was awkward to maintain it.  As a result, the compiler now issues some of
478    these warnings when compiling init.cmi during bootstrap compilation. On
479    the plus side, you also get a warning when you do, for example:
480       val op = = Int32.+
481    which was not the case up to now.
482    
483    I placed "assign" and "deref" into the _Core structure so that the
484    code that deals with the "lazy" keyword can find them there.  This
485    removes the need for having access to the primitive environment
486    during elaboration.
487    
488    ----------------------------------------------------------------------
489    Name: Matthias Blume
490    Date: 2001/08/13
491    Tag: blume-20010813-closures
492    Description:
493    
494    This fix was sent to us by Zhong Shao.  It is supposed to improve the
495    performance of certain loops by avoiding needless closure allocation.
496    
497    ----------------------------------------------------------------------
498    Name: Lal George
499    Date: 2001/07/31 10:03:23 EDT 2001
500    Tag: george-20010731-x86-fmalloc
501    Description: Fixed bug in x86 calls
502    
503        There was a bug where call instructions would mysteriously
504        vanish. The call instruction had to be one that returned
505        a floating point value.
506    
507    ----------------------------------------------------------------------
508    Name: Lal George
509    Date: 2001/07/19 16:36:29 EDT 2001
510    Tag: george-20010719-simple-cells
511    Description:
512    
513    I have dramatically simplified the interface for CELLS in MLRISC.
514    
515    In summary, the cells interface is broken up into three parts:
516    
517      1. CellsBasis : CELLS_BASIS
518    
519            CellsBasis is a top level structure and common for all
520            architectures.  it contains the definitions of basic datatypes
521            and utility  functions over these types.
522    
523      2. functor Cells() : CELLS
524    
525            Cells generates an interface for CELLS that incorporates the
526            specific resources on the target architecture, such as the
527            presence of special register classes, their number and size,
528            and various useful substructures.
529    
530      3. <ARCH>CELLS
531    
532            e.g. SparcCells: SPARCCELLS
533    
534            <ARCH>CELLS usually contains additional bindings for special
535            registers  on the architecture, such as:
536    
537                    val r0 : cell           (* register zero *)
538                    val y : cell            (* Y register *)
539                    val psr : cell          (* processor status register *)
540                    ...
541    
542            The structure returned by applying the Cells functor is opened
543            in this interface.
544    
545    The main implication of all this is that the datatypes for cells is
546    split between CellsBasis and CELLS -- a fairly simple change for user
547    code.
548    
549    In the old scheme the CELLS interface had a definitional binding of
550    the form:
551    
552            signature CELLS = sig
553    
554               structure CellsBasis = CellsBasis
555    
556               ...
557    
558            end
559    
560    With all the sharing constraints that goes on in MLRISC, this old
561    design  quickly leads to errors such as:
562    
563            "structure definition spec inside of sharing ... "
564    
565    
566    and appears to require an unacceptable amount of sharing and where
567    constraint hackery.
568    
569    I think this error message (the interaction of definitional specs and
570    sharing) requires more explanation on our web page.
571    
572    ----------------------------------------------------------------------
573    Name: Matthias Blume
574    Date: 2001/07/19 15:00:00 EDT
575    Tag: blume-20010719-libreorg
576    Description:
577    
578    This update puts together a fairly extensive but straightforward change
579    to the way the libraries that implement the interactive system are
580    organized:
581    
582       The biggest change is the elimination of structure Compiler.  As a
583       replacement for this structure, there is now a CM library
584       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
585       that exports all the substructures of the original structure Compiler
586       directly.  So instead of saying Compiler.Foo.bar one now simply
587       says Foo.bar.  (The CM libraries actually export a collection of
588       structures that is richer than the collection of substructures of
589       structure Compiler.)
590    
591       To make the transition smooth, there is a separate library called
592       $smlnj/compiler/compiler.cm which puts together and exports the
593       original structure Compiler (or at least something very close to it).
594    
595       There are five members of the original structure Compiler
596       that are not exported directly but which instead became members
597       of a new structure Backend (described by signature BACKEND).  These are:
598       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
599       Interact (: INTERACT), structure Machine (: MACHINE), and val
600       architecture (: string).
601    
602       Structure Compiler.Version has become structure CompilerVersion.
603    
604       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
605       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
606       or x86, respectively.
607       Each of these exports the same frontend structures that
608       $smlnj/compiler.cm exports.  But they do not have a structure Backend
609       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
610       Hppa, PPC, Sparc, or X86, respectively.
611    
612       Library $smlnj/compiler/all.cm exports the union of the exports of
613       $smlnj/compiler/<arch>.cm
614    
615       There are no structures <Arch>Compiler anymore, use
616       $smlnj/compiler/<arch>.cm instead.
617    
618       Library host-compiler-0.cm is gone.  Instead, the internal library
619       that instantiates CM is now called cm0.cm.  Selection of the host
620       compiler (backend) is no longer done here but. (Responsibility for it
621       now lies with $smlnj/compiler/current.cm.  This seems to be more
622       logical.)
623    
624       Many individual files have been moved or renamed.  Some files have
625       been split into multiple files, and some "dead" files have been deleted.
626    
627    Aside from these changes to library organization, there are also changes
628    to the way the code itself is organized:
629    
630       Structure Binfile has been re-implemented in such a way that it no
631       longer needs any knowledge of the compiler.  It exclusively deals
632       with the details of binfile layout.  It no longer invokes the
633       compiler (for the purpose of creating new prospective binfile
634       content), and it no longer has any knowledge of how to interpret
635       pickles.
636    
637       Structure Compile (: COMPILE) has been stripped down to the bare
638       essentials of compilation.  It no longer deals with linking/execution.
639       The interface has been cleaned up considerably.
640    
641       Utility routines for dealing with linking and execution have been
642       moved into their own substructures.
643    
644       (The ultimate goal of these changes is to provide a light-weight
645       binfile loader/linker (at least for, e.g., stable libraries) that
646       does not require CM or the compiler to be present.)
647    
648    CM documentation has been updated to reflect the changes to library
649    organization.
650    
651    ----------------------------------------------------------------------
652    Name: Matthias Blume
653    Date: 2001/07/10 17:30:00 EDT
654    Tag: Release_110_34
655    Description:
656    
657    Minor tweak to 110.34 (re-tagged):
658    
659      - README.html file added to CVS repository
660      - runtime compiles properly under FreeBSD 3.X and 4.X
661    
662    ----------------------------------------------------------------------
663    Name: Matthias Blume
664    Date: 2001/07/10 17:30:00 EDT
665    Tag: Release_110_34
666    Description:
667    
668    New version number (110.34). New bootfiles.
669    
670    ----------------------------------------------------------------------
671    Name: Matthias Blume
672    Date: 2001/07/09 16:00:00 EDT
673    Tag: blume-20010709-more-varargs
674    Description:
675    
676    I changed the handling of varargs in ml-nlffigen again:
677    The ellipsis ... will now simply be ignored (with an accompanying warning).
678    
679    The immediate effect is that you can actually call a varargs function
680    from ML -- but you can't actually supply any arguments beyond the ones
681    specified explicitly.  (For example, you can call printf with its format
682    string, but you cannot pass additional arguments.)
683    
684    This behavior is only marginally more useful than the one before, but
685    it has the advantage that a function or, more importantly, a function
686    type never gets dropped on the floor, thus avoiding follow-up problems with
687    other types that refer to the offending one.
688    
689    ----------------------------------------------------------------------
690    Name: Matthias Blume
691    Date: 2001/07/09 11:25:00 EDT
692    Tag: blume-20010709-varargs
693    Description:
694    
695    1. ckit-lib.cm now exports structure Error
696    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
697       with a warning accompanied by a source location.  Moreover, it
698       merely skips the offending function or type and proceeds with the
699       rest of its work.u  As a result, one can safely feed C code containing
700       "..." to ml-nlffigen.
701    3. There are some internal improvements to CM, providing slightly
702       more general string substitutions in the tools subsystem.
703    
704    ----------------------------------------------------------------------
705    Name: Matthias Blume
706    Date: 2001/06/27 15:10:00 EDT
707    Tag: blume-20010627-concur
708    Description:
709    
710    Fixed a small bug in CM's handling of parallel compilation.
711    (You could observe the bug by Control-C-interrupting an ordinary
712    CMB.make or CM.stabilize and then attaching some compile servers.
713    The result was that all of a sudden the previously interrupted
714    compilation would continue on its own.  This was because of
715    an over-optimization: CM did not bother to clean out certain queues
716    when no servers were attached "anyway", resulting in the contents
717    of these queues to grab control when new servers did get attached.)
718    
719    There is also another minor update to the CM manual.
720    
721    ----------------------------------------------------------------------
722    Name: Matthias Blume
723    Date: 2001/06/26 16:15:00 EDT
724    Tag: blume-20010626-cmdoc
725    Description:
726    
727    Minor typo fixed in CM manual (syntax diagram for libraries).
728    
729    ----------------------------------------------------------------------
730    Name: Matthias Blume
731    Date: 2001/06/25 22:55:00 EDT
732    Tag: blume-20010625-x86pc
733    Description:
734    
735    Fixed a nasty bug in the X86 assembly code that caused signal
736    handlers to fail (crash) randomly.
737    
738    ----------------------------------------------------------------------
739    Name: Matthias Blume
740    Date: 2001/06/25 12:05:00 EDT
741    Tag: blume-20010625-nlffigen
742    Description:
743    
744    This update fixes a number of minor bugs in ml-nlffigen as reported by
745    Nick Carter <nbc@andrew.cmu.edu>.
746    
747      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
748      2. Default names for generated files are now derived from the name of
749         the C file *without its directory*.  In particular, this causes generated
750         files to be placed locally even if the C file is in some system directory.
751      3. Default names for generated signatures and structures are also derived
752         from the C file name without its directory.  This avoids silly things
753         like "structure GL/GL".
754         (Other silly names are still possible because ml-nlffigen does not do
755          a thorough check of whether generated names are legal ML identifiers.
756          When in doubt, use command line arguments to force particular names.)
757    
758    ----------------------------------------------------------------------
759    Name: Matthias Blume
760    Date: 2001/06/21 12:25:00 EDT
761    Tag: blume-20010621-eXene
762    Description:
763    
764    eXene now compiles and (sort of) works again.
765    
766    The library name (for version > 110.33) is $/eXene.cm.
767    
768    I also added an new example in src/eXene/examples/nbody.  See the
769    README file there for details.
770    
771    ----------------------------------------------------------------------
772    Name: Matthias Blume
773    Date: 2001/06/20 16:40:00 EDT
774    Tag: blume-20010620-cml
775    Description:
776    
777    CML now compiles and works again.
778    
779    Libraries (for version > 110.33):
780    
781      $cml/cml.cm            Main CML library.
782      $cml/basis.cm          CML's version of $/basis.cm.
783      $cml/cml-internal.cm   Internal helper library.
784      $cml/core-cml.cm       Internal helper library.
785      $cml-lib/trace-cml.cm  Tracing facility.
786      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
787    
788    The installer (config/install.sh) has been taught how to properly
789    install this stuff.
790    
791    ----------------------------------------------------------------------
792    Name: Matthias Blume
793    Date: 2001/06/19 17:55:00 EDT
794    Tag: blume-20010619-instantiate
795    Description:
796    
797    This un-breaks the fix for bug 1432.
798    (The bug was originally fixed in 110.9 but I broke it again some
799    time after that.)
800    
801    ----------------------------------------------------------------------
802    Name: Matthias Blume
803    Date: 2001/06/19 17:25:00 EDT
804    Tag: blume-20010619-signals
805    Description:
806    
807    This should (hopefully) fix the long-standing signal handling bug.
808    (The runtime system was constructing a continuation record with an
809    incorrect descriptor which would cause the GC to drop data on the floor...)
810    
811    ----------------------------------------------------------------------
812    Name: Matthias Blume
813    Date: 2001/06/15 15:05:00 EDT
814    Tag: blume-20010615-moresparc
815    Description:
816    
817    Here is a short late-hour update related to Sparc c-calls:
818    
819     -- made handling of double-word arguments a bit smarter
820    
821     -- instruction selection phase tries to collapse certain clumsily
822        constructed ML-Trees; typical example:
823    
824            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
825    
826        This currently has no further impact on SML/NJ since mlriscGen does
827        not seem to generate such patterns in the first place, and c-calls
828        (which did generate them in the beginning) has meanwhile been fixed
829        so as to avoid them as well.
830    
831    ----------------------------------------------------------------------
832    Name: Matthias Blume
833    Date: 2001/06/15 15:05:00 EDT
834    Tag: blume-20010615-sparc
835    Description:
836    
837    The purpose of this update is to provide an implementation of NLFFI
838    on Sparc machines.
839    
840    Here are the changes in detail:
841    
842       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
843       the Sparc implementation of the c-calls API.
844       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
845       for accessing the ML frame.  Thus, we have a real frame pointer and
846       can freely modify %sp without need for an omit-frame-ptr phase.
847       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
848       this case.
849       * ml-nlffigen has been taught to produce code for different architectures
850       and calling conventions.
851       * In a way similar to what was done in the x86 case, the Sparc
852       backend uses its own specific extension to mltree.  (For example,
853       it needs to be able to generate UNIMP instructions which are part
854       of the calling convention.)
855       * ml-nlffi-lib was reorganized to make it more modular (in particular,
856       to make it easier to plug in new machine- and os-dependent parts).
857    
858    There are some other fairly unrelated bug fixes and cleanups as well:
859    
860       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
861       that they properly share their libraries with existing SML/NJ libraries.
862       * I fixed a minor cosmetic bug in CM, supressing certain spurious
863       follow-up error messages.
864       * Updates to CM/CMB documentation.
865    
866    TODO items:
867    
868       * MLRISC should use a different register as its asmTemp on the Sparc.
869         (The current %o2 is a really bad choice because it is part of the
870         calling conventions, so things might interfere in unexpected ways.)
871    
872    ----------------------------------------------------------------------
873    Name: Matthias Blume
874    Date: 2001/06/07
875    Tag: blume-20010607-calls
876    Description:
877    
878    A number of internal changes related to C calls and calling conventions:
879    
880    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
881       number of bytes popped implicitly (by the callee).  In most cases
882       this field is 0 but on x86/win32 it is some non-zero value.  This
883       is information provided for the benefit of the "omit-frameptr" pass.
884    2. The CALL instruction on the x86 carries a similar "pops" field.
885       The instruction selection phase copies its value from the ML-Tree
886       CALL statement.
887    3. On all other architectures, the instruction selection phase checks
888       whether "pops=0" and complains if not.
889    4. The c-calls implementation for x86 now accepts two calling conventions:
890       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
891       up after the call and pops is set to 0.  For "stdcall", the caller
892       does nothing, leaving the cleanup to the callee; pops is set to
893       the number of bytes that were pushed onto the stack.
894    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
895       distinguish between "ccall" and "stdcall".
896    6. The UNIMP instruction has been added to the supported Sparc instruction
897       set. (This is needed for implementing the official C calling convention
898       on this architecture.)
899    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
900       work with the latest CM.
901    
902    ----------------------------------------------------------------------
903    Name: Matthias Blume
904    Date: 2001/06/05 15:10:00 EDT
905    Tag: blume-20010605-cm-index
906    Description:
907    
908    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
909    
910    1. CM can now generate "index files".  These are human-readable files
911       that list on a per-.cm-file basis each toplevel symbol defined or
912       imported.  The location of the index file for
913       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
914       To enable index-file generation, set CM.Control.generate_index to true
915       or export an environment-symbol: export CM_GENERATE_INDEX=true.
916    
917       The CM manual has been updated accordingly.
918    
919    2. I made some slight modifications to the c-calls API in MLRISC.
920    
921         a) There is now a callback to support saving/restoring of
922            dedicated but caller-save registers around the actual call
923            instruction.
924         b) One can optionally specify a comment-annotation for the
925            call instruction.
926    
927    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
928       (For example, the comment annotation shows the C prototype of
929        the function being called.)
930    
931    ----------------------------------------------------------------------
932    Name: Matthias Blume
933    Date: 2001/06/01 13:30:00 EDT
934    Tag: blume-20010601-nlffi-cleanup
935    Description:
936    
937    This is mostly a cleanup of MLFFI stuff:
938    
939       - some signature files have been put into a more exposed place
940       - the ugly 'f type parameter is gone (simplifies types tremendously!)
941       - ml-nlffigen changed accordingly
942       - tutorial updated
943    
944    Other changes:
945    
946       - author's affiliation in CM manual(s) updated
947       - some more recognized keywords added to Allen's sml.sty
948    
949    ----------------------------------------------------------------------
950    Name: Matthias Blume
951    Date: 2001/05/25 15:30:00 EDT
952    Tag: blume-20010525-iptr
953    Description:
954    
955      - put the official 110.33-README (as it appears on the ftp server) under
956        CVS
957      - fixed a small bug related to incomplete pointer types in
958        ml-nlffigen
959      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
960        (it does not need the 'f type parameter)
961    
962    ----------------------------------------------------------------------
963    Name: Matthias Blume
964    Date: 2001/05/23 14:30:00 EDT
965    Tag: Release_110_33
966    Description:
967    
968    New version number (110.33).  New bootfiles.
969    
970    ----------------------------------------------------------------------
971    Name: Matthias Blume
972    Date: 2001/05/22 18:06:00 EDT
973    Tag: blume-20010522-targets
974    Description:
975    
976    Made install.sh use file config/targets.customized if it exists, falling
977    back to config/targets if it doesn't.  This way one can have a customized
978    version of the targets file without touching the "real thing", thus
979    eliminating the constant fear of accidentally checking something bogus
980    back into the CVS repository...  (File config/targets.customized must
981    not be added to the repository!)
982    
983    ----------------------------------------------------------------------
984    Name: Matthias Blume
985    Date: 2001/05/22 16:30:00 EDT
986    Tag: blume-20010522-minitut
987    Description:
988    
989    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
990       struct returns.
991    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
992       incomplete, preliminary documentation for NLFFI.
993    
994    ----------------------------------------------------------------------
995    Name: Matthias Blume
996    Date: 2001/05/14 11:30:00 EDT
997    Tag: blume-20010514-script
998    Description:
999    
1000    Some bugs in install script fixed.
1001    
1002    In addition to that I also made a slight change to the NLFFI API:
1003    Functors generated by ml-nlffigen now take the dynamic library as a
1004    straight functor argument, not as a suspended one.  (The original
1005    functor code used to force the suspension right away anyway, so there
1006    was nothing gained by this complication of the interface.)
1007    
1008    ----------------------------------------------------------------------
1009    Name: Matthias Blume
1010    Date: 2001/05/11 14:35:00 EDT
1011    Tag: blume-20010511-ml-nlffi
1012    Description:
1013    
1014    I finally took the plunge and added my new FFI code to the main
1015    repository.  For x86-linux it is now ready for prime-time.
1016    
1017    There are two new subdirectories of "src":
1018    
1019      - ml-nlffi-lib:
1020           The utility library for programs using the FFI interface.
1021           Here is the implementation of $/c.cm and its associated low-level
1022           partners $/c-int.cm and $/memory.cm.
1023      - ml-nlffigen:
1024           A stand-alone program for generating ML glue code from C source
1025           code.
1026    
1027    Building ml-nlffigen requires $/ckit-lib.cm.
1028    
1029    The config/install.sh script has been updates to do the Right Thing
1030    (hopefully).
1031    
1032    Notice that the source tree for the C-Kit will not be put under "src"
1033    but directly under the installation root directory.  (This is the
1034    structure that currently exists on the CVS server when you check out
1035    module "sml".)  Fortunately, config/install.sh knows about this oddity.
1036    
1037    Bugs: No documentation yet.
1038    
1039    ----------------------------------------------------------------------
1040    Name: Matthias Blume
1041    Date: 2001/05/09 16:35:00 EDT
1042    Tag: blume-20010509-cpscontract
1043    Description:
1044    
1045    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
1046    wrapper/unwrapper elimination did not decrement usage counts and some
1047    dead variables got overlooked by the dead-up logic.)
1048    
1049    ----------------------------------------------------------------------
1050    Name: Lal George
1051    Date: 2001/05/08  17:26:09 EDT
1052    Tag: george-20010508-omit-frameptr
1053    Description:
1054    
1055    Changes to implement the omit-frame-pointer optimization to support
1056    raw C calls. For now, there is only support on the Intel x86, but
1057    other architectures will follow as more experience is gained with this.
1058    
1059    
1060    ----------------------------------------------------------------------
1061    Name: Matthias Blume
1062    Date: 2001/05/07 14:40:00 EDT
1063    Tag: blume-20010507-proxies
1064    Description:
1065    
1066    I made into "proxy libraries" all libraries that qualify for such a
1067    change.  (A qualifying library is a library that has another library or
1068    groups as its sole member and repeats that member's export list
1069    verbatim.  A proxy library avoids this repetition by omitting its export
1070    list, effectively inheriting the list that its (only) member exports.
1071    See the CM manual for more explanation.)
1072    The main effect is that explicit export lists for these libraries
1073    do not have to be kepts in sync, making maintenance a bit easier.
1074    
1075    I also added copyright notices to many .cm-files.
1076    
1077    Last but not least, I made a new set of bootfiles.
1078    
1079    ----------------------------------------------------------------------
1080    Name: Matthias Blume
1081    Date: 2001/05/04 17:00:00 EDT
1082    Tag: blume-20010504-cm-lsplit
1083    Description:
1084    
1085    0. John merged pending changes to $/smlnj-lib.cm
1086    
1087    1. Allen's previous change accidentally backed out of one of Lal's
1088       earlier changes.  I undid this mistake (re-introducing Lal's change).
1089    
1090    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
1091       within the compiler where applicable.  There is some code simplification
1092       because of that.
1093    
1094    3. The "split" phase (in FLINT) is now part of the default list of phases.
1095       Compiler.Control.LambdaSplitting.* can be used to globally control the
1096       lambda-splitting (cross-module-inlining) engine.  In addition to that,
1097       it can now also be controlled on a per-source basis: CM has been taught
1098       a new tool parameter applicable to ML source files.
1099    
1100       - To turn lambda-splitting off completely:
1101            local open Compiler.Control.LambdaSplitting in
1102                val _ = set Off
1103            end
1104       - To make "no lambda-splitting" the global default (but allow per-source
1105         overriding); this is the initial setting:
1106            local open Compiler.Control.LambdaSplitting in
1107                val _ = set (Default NONE)
1108            end
1109       - To make "lambda-splitting with aggressiveness a" the global default
1110         (and allow per-source overriding):
1111            local open Compiler.Control.LambdaSplitting in
1112                val _ = set (Default (SOME a))
1113            end
1114    
1115       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
1116         write (in the respective .cm-file):
1117            a.sml (lambdasplitting:off)
1118       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
1119            a.sml (lambdasplitting:on)
1120       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
1121         <a> is a decimal non-negative integer):
1122            a.sml (lambdasplitting:<a>)
1123       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
1124            a.sml (lambdasplitting:infinity)
1125       - To use the global default for a.sml:
1126            a.sml (lambdasplitting:default)
1127         or simply
1128            a.sml
1129    
1130    ----------------------------------------------------------------------
1131    Name: Allen Leung
1132    Date: 2001/05/04 01:57:00 EDT
1133    Tag: leunga-20010504-sync
1134    Description:
1135    
1136      MLRISC features.
1137    
1138      1. Fix to CMPXCHG instructions.
1139      2. Changed RA interface to allow annotations in callbacks.
1140      3. Added a new method to the stream interface to allow annotations updates.
1141    
1142    ----------------------------------------------------------------------
1143    Name: Matthias Blume
1144    Date: 2001/05/01 11:45:00 EDT
1145    Tag: blume-20010501-pcedittmp
1146    Description:
1147    
1148    Changed install.sh to use the current working directory instead of
1149    /usr/tmp for a temporary file (pcedittmp).  The previous choice
1150    of /usr/tmp caused trouble with MacOS X because of file premission
1151    problems.
1152    
1153    ----------------------------------------------------------------------
1154    Name: Matthias Blume
1155    Date: 2001/04/20 11:10:00 EDT
1156    Tag: blume-20010420-inMLflag
1157    Description:
1158    
1159     - added vp_limitPtrMask to vproc-state.h
1160       (for use by the raw-C-calls mechanism to implement proper interrupt
1161        handling)
1162     - made the ML compiler aware of various data-structure offsets so it
1163       can generate code for accessing the vp_inML flag and vp_limitPtrMask
1164     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
1165       raw C-calls
1166    
1167    ----------------------------------------------------------------------
1168    Name: Lal George
1169    Date: 2001/04/20 09:15:28 EDT
1170    Tag: george-20010420-macosX
1171    Description:
1172    
1173     - Changes to port to Mac OS X; Darwin.
1174    
1175     - In the process I found that sqrt was broken on the PPC, because the
1176       fsqrt instruction is not implemented.
1177    
1178    ----------------------------------------------------------------------
1179    Name: Matthias Blume
1180    Date: 2001/04/18 12:45:00 EDT
1181    Tag: blume-20010418-ccalls
1182    Description:
1183    
1184     - fixed two off-by-4 errors in the x86-specific c-calls implementation
1185       (this bug prevented structure arguments containing pointers from being
1186        passed correctly)
1187     - changed the raw-C-call code in mlriscGen.sml in such a way that
1188       structure arguments are represented as a pointer to the beginning
1189       of the structure (instead of having a series of synthesized arguments,
1190       one for each structure member)
1191    
1192     - made makeml script's verbosity level configurable via environment
1193       variable (MAKEML_VERBOSITY)
1194    
1195     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
1196       in rawmem-x86.sml; we are now using the real thing
1197    
1198    ----------------------------------------------------------------------
1199    Name: Matthias Blume
1200    Date: 2001/03/22 16:25:00 EST
1201    Tag: blume-20010322-bootfiles
1202    Description:
1203    
1204    Created a new set of bootfiles (for your automatic installation convenience).
1205    
1206    ----------------------------------------------------------------------
1207    Name: Matthias Blume
1208    Date: 2001/03/22 15:10:00 EST
1209    Tag: blume-20010322-rawmem-parcm
1210    Description:
1211    
1212    1. All "raw memory access" primitives for the new FFI are implemented now
1213       (at least on the x86).
1214    2. Some further cleanup of CM's parallel make mechanism.
1215    
1216    ----------------------------------------------------------------------
1217    Name: Matthias Blume
1218  Date: 2001/03/19 17:53:00 EST  Date: 2001/03/19 17:53:00 EST
1219  Tag: blume-20010319-parallel  Tag: blume-20010319-parallel
1220  Description:  Description:
# Line 2024  Line 3225 
3225       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3226       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3227       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
3228       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3229       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3230       Dave's hands. :)       Dave's hands. :)
3231    

Legend:
Removed from v.801  
changed lines
  Added in v.987

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