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 987, Wed Nov 21 21:14:44 2001 UTC revision 1192, Wed May 15 14:02:06 2002 UTC
# Line 8  Line 8 
8  The form of an entry should be:  The form of an entry should be:
9    
10  Name:  Name:
11  Date:  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Lal George
17    Date: 2002/05/15 09:20:10 EDT
18    Tag: george-20020515-pseudo-op-decls
19    Description:
20    
21            All pseudo-ops emitted before the first segment declaration
22    such as TEXT, DATA, and BSS directives are assumed to be global
23    declarations and are emitted first in the assembly file. This is
24    useful in a number of situations where one has pseudo-ops that are not
25    specific to any segment, and also works around the constraint that one
26    cannot have client pseudo-ops in the TEXT segment.
27    
28            Because no segment is associated with these declarations it is
29    an error to allocate any space or objects before the first segment
30    directive and an exception will be raised. However, we cannot make
31    this check for client pseudo-ops.
32    
33            These top level declarations are a field in the CFG graph_info.
34    In theory you can continue to add to this field after the CFG has been
35    built -- provided you know what you are doing;-)
36    
37    ----------------------------------------------------------------------
38    Name: Matthias Blume
39    Date: 2002/05/13 16:40:00 EDT
40    Tag: blume-20020513-pp-etc
41    Description:
42    
43    A few minor bugfixes:
44    
45      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
46        (Bogus printouts for redefined bindings still occur.  Compiler
47        bug should no longer occur now.  We need to redo the prettyprinter
48        from scratch.)
49    
50      - CM pathname printer now also adds escape sequences for ( and )
51    
52      - commend and docu fixes for ml-nlffi
53    
54    ----------------------------------------------------------------------
55    Name: Matthias Blume
56    Date: 2002/05/10 16:40:00 EDT
57    Tag: blume-20020510-erg-textio
58    Description:
59    
60    Applied the following bugfix provided by Emden Gansner:
61    
62        Output is corrupted when outputSubstr is used rather than output.
63    
64        The problem occurs when a substring
65    
66            ss = (s, dataStart, dataLen)
67    
68        where dataStart > 0, fills a stream buffer with avail bytes left.
69        avail bytes of s, starting at index dataStart, are copied into the
70        buffer, the buffer is flushed, and then the remaining dataLen-avail
71        bytes of ss are copied into the beginning of the buffer. Instead of
72        starting this copy at index dataStart+avail in s, the current code
73        starts the copy at index avail.
74    
75        Fix:
76        In text-io-fn.sml, change line 695 from
77             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
78        to
79             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
80    
81    ----------------------------------------------------------------------
82    Name: Matthias Blume
83    Date: 2002/04/12 13:55:00 EDT
84    Tag: blume-20020412-assyntax
85    Description:
86    
87    1. Grabbed newer assyntax.h from the XFree86 project.
88    2. Fiddled with how to compile X86.prim.asm without warnings.
89    3. (Very) Minor cleanup in CM.
90    
91    ----------------------------------------------------------------------
92    Name: Matthias Blume
93    Date: 2002/04/01 (no joke!) 17:07:00 EST
94    Tag: blume-20020401-x86div
95    Description:
96    
97    Added full support for div/mod/rem/quot on the x86, using the machine
98    instruction's two results (without clumsily recomputing the remainder)
99    directly where appropriate.
100    
101    Some more extensive power-of-two support was added to the x86 instruction
102    selector (avoiding expensive divs, mods, and muls where they can be
103    replaced with cheaper shifts and masks).  However, this sort of thing
104    ought to be done earlier, e.g., within the CPS optimizer so that
105    all architectures benefit from it.
106    
107    The compiler compiles to a fixed point, but changes might be somewhat
108    fragile nevertheless.  Please, report any strange things that you might
109    see wrt. div/mod/quot/rem...
110    
111    ----------------------------------------------------------------------
112    Name: Matthias Blume
113    Date: 2002/03/29 17:22:00
114    Tag: blume-20020329-div
115    Description:
116    
117    Fixed my broken div/mod logic.  Unfortunately, this means that the
118    inline code for div/mod now has one more comparison than before.
119    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
120    The problem was with quotient = 0, because that alone does not tell
121    us which way the rounding went.  One then has to look at whether
122    remainder and divisor have the same sign...  :(
123    
124    Anyway, I replaced the bootfiles with fresh ones...
125    
126    ----------------------------------------------------------------------
127    Name: Matthias Blume
128    Date: 2002/03/29 14:10:00 EST
129    Tag: blume-20020329-inlprims
130    Description:
131    
132    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
133    
134    Primops have changed. This means that the bin/boot-file formats have
135    changed as well.
136    
137    To make sure that there is no confusion, I made a new version.
138    
139    
140    CHANGES:
141    
142    * removed REMT from mltree (remainder should never overflow).
143    
144    * added primops to deal with divisions of all flavors to the frontend
145    
146    * handled these primops all the way through so they map to their respective
147      MLRISC support
148    
149    * used these primops in the implementation of Int, Int32, Word, Word32
150    
151    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
152    
153    * parameterized INLMIN, INLMAX, and INLABS by a numkind
154    
155    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
156      floating point
157    
158    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
159      and Real (but Real.abs maps to a separate floating-point-only primop)
160    
161    
162    TODO items:
163    
164    * Hacked Alpha32 instruction selection, disabling the selection of REMx
165      instructions because the machine instruction encoder cannot handle
166      them.  (Hppa, PPC, and Sparc instruction selection did not handle
167      REM in the first place, and REM is supported by the x86 machine coder.)
168    
169    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
170      selection phase.  (The two can be streamlined because the hardware
171      delivers both quotient and remainder at the same time anyway.)
172    
173    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
174      (Currently the behavior is inconsistent both across architectures and
175      wrt. the draft Basis spec.)
176    
177    * Word8 should eventually be handled natively, too.
178    
179    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
180      as if there currently is no execution path that could trigger it in
181      SML/NJ.  (The assumptions underlying functions arith and promotable do not
182      hold for things like multiplication and division.)
183    
184    ----------------------------------------------------------------------
185    Name: Matthias Blume
186    Date: 2002/03/27 16:27:00 EST
187    Tag: blume-20020327-mlrisc-divisions
188    Description:
189    
190    Added support for all four division operations (ML's div, mod, quot,
191    and rem) to MLRISC.  In the course of doing so, I also rationalized
192    the naming (no more annoying switch-around of DIV and QUOT), by
193    parameterizing the operation by div_rounding_mode (which can be either
194    DIV_TO_ZERO or DIV_TO_NEGINF).
195    
196    The generic MLTreeGen functor takes care of compiling all four
197    operations down to only round-to-zero div.
198    
199    Missing pieces:
200    
201      * Doing something smarter than relying on MLTreeGen on architectures
202        like, e.g., the x86 where hardware division delivers both quotient and
203        remainder at the same time.  With this, the implementation of the
204        round-to-neginf operations could be further streamlined.
205    
206      * Remove inlining support for div/mod/rem from the frontend and replace it
207        with primops that get carried through to the backend.  Do this for all
208        int and word types.
209    
210    ----------------------------------------------------------------------
211    Name: Matthias Blume
212    Date: 2002/03/25 17:25:00 EST
213    Tag: blume-20020325-divmod
214    Description:
215    
216    I improved (hopefully without breaking them) the implementation of Int.div,
217    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
218    advantage of the following observations:
219    
220      Let  q = x quot y      r = x rem y
221           d = x div  y      m = x mod y
222    
223    where "quot" is the round-to-zero version of integer division that
224    hardware usually provides.  Then we have:
225    
226         r = x - q * y        where neither the * nor the - will overflow
227         d = if q >= 0 orelse x = q * y then q else q - 1
228                              where neither the * nor the - will overflow
229         m = if q >= 0 orelse r = 0 then r else r + y
230                              where the + will not overflow
231    
232    This results in substantial simplification of the generated code.
233    The following table shows the number of CFG nodes and edges generated
234    for
235            fun f (x, y) = x OPER y
236            (* with OPER \in div, mod, quot, rem *)
237    
238    
239        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
240        --------------------------------------------------------
241         div |         24 |         39 |         12 |         16
242         mod |         41 |         71 |         12 |         16
243        quot |          8 |         10 |          8 |         10
244         rem |         10 |         14 |          8 |         10
245    
246    
247    ----------------------------------------------------------------------
248    Name: Matthias Blume
249    Date: 2002/03/25 22:06:00 EST
250    Tag: blume-20020325-cprotobug
251    Description:
252    
253    Fixed a bug in cproto (c prototype decoder).
254    
255    ----------------------------------------------------------------------
256    Name: Matthias Blume
257    Date: 2002/03/25 16:00:00 EST
258    Tag: blume-20020325-raw-primops
259    Description:
260    
261    I did some cleanup to Allen's new primop code and
262    replaced yesterday's bootfiles with new ones.
263    (But they are stored in the same place.)
264    
265    ----------------------------------------------------------------------
266    Name: Matthias Blume
267    Date: 2002/03/24 22:40:00 EST
268    Tag: blume-20020324-bootfiles
269    Description:
270    
271    Made the bootfiles that Allen asked for.
272    
273    ----------------------------------------------------------------------
274    Name: Allen Leung
275    Date: 2002/03/23 15:50:00 EST
276    Tag: leunga-20020323-flint-cps-rcc-primops
277    Description:
278    
279      1. Changes to FLINT primops:
280    
281        (* make a call to a C-function;
282         * The primop carries C function prototype information and specifies
283         * which of its (ML-) arguments are floating point. C prototype
284         * information is for use by the backend, ML information is for
285         * use by the CPS converter. *)
286      | RAW_CCALL of { c_proto: CTypes.c_proto,
287                       ml_args: ccall_type list,
288                       ml_res_opt: ccall_type option,
289                       reentrant : bool
290                     } option
291       (* Allocate uninitialized storage on the heap.
292        * The record is meant to hold short-lived C objects, i.e., they
293        * are not ML pointers.  With the tag, the representation is
294        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
295        *)
296      | RAW_RECORD of {tag:bool,sz:int}
297      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
298    
299      2.  These CPS primops are now overloaded:
300    
301           rawload of {kind:numkind}
302           rawstore of {kind:numkind}
303    
304          The one argument form is:
305    
306             rawload {kind} address
307    
308          The two argument form is:
309    
310             rawload {kind} [ml object, byte-offset]
311    
312      3. RAW_CCALL/RCC now takes two extra arguments:
313    
314         a. The first is whether the C call is reentrant, i.e., whether
315            ML state should be saved and restored.
316         b. The second argument is a string argument specifying the name of
317            library and the C function.
318    
319         These things are currently not handled in the code generator, yet.
320    
321      4. In CProto,
322    
323         An encoding type of "bool" means "ml object" and is mapped into
324         C prototype of PTR.  Note that "bool" is different than "string",
325         even though "string" is also mapped into PTR, because "bool"
326         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
327    
328      5. Pickler/unpicker
329    
330         Changed to handle RAW_RECORD and newest RAW_CCALL
331    
332      6. MLRiscGen,
333    
334         1. Changed to handle the new rawload/rawstore/rawrecord operators.
335         2. Code for handling C Calls has been moved to a new module CPSCCalls,
336            in the file CodeGen/cpscompile/cps-c-calls.sml
337    
338      7. Added the conditional move operator
339    
340             condmove of branch
341    
342         to cps.  Generation of this is still buggy so it is currently
343         disabled.
344    
345    ----------------------------------------------------------------------
346    Name: Lal George
347    Date: 2002/03/22 14:18:25 EST
348    Tag: george-20020322-cps-branch-prob
349    Description:
350    
351    Implemented the Ball-Larus branch prediction-heuristics, and
352    incorporated graphical viewers for control flow graphs.
353    
354    Ball-Larus Heuristics:
355    ---------------------
356    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
357    
358    By design it uses the Dempster-Shafer theory for combining
359    probabilities.  For example, in the function:
360    
361        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
362    
363    the ball-larus heuristics predicts that the n=0 is unlikely
364    (OH-heuristic), and the 'then' branch is unlikely because of the
365    RH-heuristic -- giving the 'then' branch an even lower combined
366    probability using the Dempster-Shafer theory.
367    
368    Finally, John Reppy's loop analysis in MLRISC, further lowers the
369    probability of the 'then' branch because of the loop in the else
370    branch.
371    
372    
373    Graphical Viewing:
374    ------------------
375    I merely plugged in Allen's graphical viewers into the compiler. The
376    additional code is not much. At the top level, saying:
377    
378            Control.MLRISC.getFlag "cfg-graphical-view" := true;
379    
380    will display the graphical view of the control flow graph just before
381    back-patching.  daVinci must be in your path for this to work. If
382    daVinci is not available, then the default viewer can be changed
383    using:
384    
385            Control.MLRISC.getString "viewer"
386    
387    which can be set to "dot" or "vcg" for the corresponding viewers. Of
388    course, these viewers must be in your path.
389    
390    The above will display the compilation unit at the level of clusters,
391    many of which are small, boring, and un-interesting. Also setting:
392    
393            Control.MLRISC.getInt "cfg-graphical-view_size"
394    
395    will display clusters that are larger than the value set by the above.
396    
397    
398    ----------------------------------------------------------------------
399    Name: Matthias Blume
400    Date: 2002/03/21 22:20:00 EST
401    Tag: blume-20020321-kmp-bugfix
402    Description:
403    
404    Changed the interface to the KMP routine in PreString and fixed
405    a minor bug in one place where it was used.
406    
407    ----------------------------------------------------------------------
408    Name: Allen Leung
409    Date: 2002/03/21 20:30:00 EST
410    Tag: leunga-20020321-cfg
411    Description:
412    
413      Fixed a potential problem in cfg edge splitting.
414    
415    ----------------------------------------------------------------------
416    Name: Allen Leung
417    Date: 2002/03/21 17:15:00 EST
418    Tag: leunga-20020321-x86-fp-cfg
419    Description:
420    
421      1. Recoded the buggy parts of x86-fp.
422    
423         a. All the block reordering code has been removed.
424            We now depend on the block placement phases to do this work.
425    
426         b. Critical edge splitting code has been simplified and moved into the
427            CFG modules, as where they belong.
428    
429         Both of these were quite buggy and complex.  The code is now much, much
430         simpler.
431    
432      2. X86 backend.
433    
434         a. Added instructions for 64-bit support.  Instruction selection for
435            64-bit has not been committed, however, since that
436            requires changes to MLTREE which haven't been approved by
437            Lal and John.
438    
439         b. Added support for FUCOMI and FUCOMIP when generating code for
440            PentiumPro and above.  We only generate these instructions in
441            the fast-fp mode.
442    
443         c. Added cases for JP and JNP in X86FreqProps.
444    
445      3. CFG
446    
447         CFG now has a bunch of methods for edge splitting and merging.
448    
449      4. Machine description.
450    
451         John's simplification of MLTREE_BASIS.fcond broke a few machine
452         description things:
453    
454         rtl-build.{sig,sml} and hppa.mdl fixed.
455    
456         NOTE: the machine description stuff in the repository is still broken.
457               Again, I can't put my fixes in because that involves
458               changes to MLTREE.
459    
460    ----------------------------------------------------------------------
461    Name: Matthias Blume
462    Date: 2002/03/20 15:55:00 EST
463    Tag: blume-20020320-kmp
464    Description:
465    
466    Implemented Knuth-Morris-Pratt string matching in PreString and used
467    it for String.isSubstring, Substring.isSubstring, and
468    Substring.position.
469    
470    (Might need some stress-testing.  Simple examples worked fine.)
471    
472    ----------------------------------------------------------------------
473    Name: Matthias Blume
474    Date: 2002/03/19 16:37:00 EST
475    Tag: blume-20020319-witnesses
476    Description:
477    
478    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
479    
480    This implements a generic mechanism for changing constness qualifiers
481    anywhere within big C types without resorting to outright "casts".
482    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
483    modify the constness at the outermost level.)
484    The implementation of "convert" is based on the idea of "witness"
485    values -- values that are not used by the operation but whose types
486    "testify" to their applicability.  On the implementation side, "convert"
487    is simply a projection (returning its second curried argument).  With
488    cross-module inlining, it should not result in any machine code being
489    generated.
490    
491    ----------------------------------------------------------------------
492    Name: Matthias Blume
493    Date: 2002/03/15 16:40:00 EST
494    Tag: blume-20020315-basis
495    Description:
496    
497    Provided (preliminary?) implementations for
498    
499      {String,Substring}.{concatWith,isSuffix,isSubstring}
500    
501    and
502    
503      Substring.full
504    
505    Those are in the Basis spec but they were missing in SML/NJ.
506    
507    ----------------------------------------------------------------------
508    Name: Matthias Blume
509    Date: 2002/03/14 21:30:00 EST
510    Tag: blume-20020314-controls
511    Description:
512    
513    Controls:
514    ---------
515    
516    1. Factored out the recently-added Controls : CONTROLS stuff and put
517       it into its own library $/controls-lib.cm.  The source tree for
518       this is under src/smlnj-lib/Controls.
519    
520    2. Changed the names of types and functions in this interface, so they
521       make a bit more "sense":
522    
523          module -> registry
524          'a registry -> 'a group
525    
526    3. The interface now deals in ref cells only.  The getter/setter interface
527       is (mostly) gone.
528    
529    4. Added a function that lets one register an already-existing ref cell.
530    
531    5. Made the corresponding modifications to the rest of the code so that
532       everything compiles again.
533    
534    6. Changed the implementation of Controls.MLRISC back to something closer
535       to the original.  In particular, this module (and therefore MLRISC)
536       does not depend on Controls.  There now is some link-time code in
537       int-sys.sml that registers the MLRISC controls with the Controls
538       module.
539    
540    CM:
541    ---
542    
543      * One can now specify the lambda-split aggressiveness in init.cmi.
544    
545    ----------------------------------------------------------------------
546    Name: Allen Leung
547    Date: 2002/03/13 17:30:00 EST
548    Tag: leunga-20020313-x86-fp-unary
549    Description:
550    
551    Bug fix for:
552    
553    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
554    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
555    > - fun f(x,(y,z)) = Real.~ y;
556    > [autoloading]
557    > [autoloading done]
558    >       fchsl   (%eax), 184(%esp)
559    > Error: MLRisc bug: X86MCEmitter.emitInstr
560    >
561    > uncaught exception Error
562    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
563    
564    The problem was that the code generator did not generate any fp registers
565    in this case, and the ra didn't know that it needed to run the X86FP phase to
566    translate the pseudo fp instruction.   This only happened with unary fp
567    operators in certain situations.
568    
569    ----------------------------------------------------------------------
570    Name: Matthias Blume
571    Date: 2002/03/13 14:00:00 EST
572    Tag: blume-20020313-overload-etc
573    Description:
574    
575    1. Added _overload as a synonym for overload for backward compatibility.
576       (Control.overloadKW must be true for either version to be accepted.)
577    
578    2. Fixed bug in install script that caused more things to be installed
579       than what was requested in config/targets.
580    
581    3. Made CM aware of the (_)overload construct so that autoloading
582       works.
583    
584    ----------------------------------------------------------------------
585    Name: Matthias Blume
586    Date: 2002/03/12 22:03:00 EST
587    Tag: blume-20020312-url
588    Description:
589    
590    Forgot to update BOOT and srcarchiveurl.
591    
592    ----------------------------------------------------------------------
593    Name: Matthias Blume
594    Date: 2002/03/12 17:30:00 EST
595    Tag: blume-20020312-version110392
596    Description:
597    
598    Yet another version number bump (because of small changes to the
599    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
600    
601    Changes:
602    
603      The new pid generation scheme described a few weeks ago was overly
604      complicated.  I implemented a new mechanism that is simpler and
605      provides a bit more "stability":  Once CM has seen a compilation
606      unit, it keeps its identity constant (as long as you do not delete
607      those crucial CM/GUID/* files).  This means that when you change
608      an interface, compile, then go back to the old interface, and
609      compile again, you arrive at the original pid.
610    
611      There now also is a mechanism that instructs CM to use the plain
612      environment hash as a module's pid (effectively making its GUID
613      the empty string).  For this, "noguid" must be specified as an
614      option to the .sml file in question within its .cm file.
615      This is most useful for code that is being generated by tools such
616      as ml-nlffigen (because during development programmers tend to
617      erase the tool's entire output directory tree including CM's cached
618      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
619      revert to the old, broken behavior of SML/NJ, but in specific cases
620      where there is no danger of interface confusion, its use is ok
621      (I think).
622    
623      ml-nlffigen by default generates "noguid" annotations.  They can be
624      turned off by specifying -guid in its command line.
625    
626    ----------------------------------------------------------------------
627    Name: Lal George
628    Date: 2002/03/12 12 14:42:36 EST
629    Tag: george-20020312-frequency-computation
630    Description:
631    
632    Integrated jump chaining and static block frequency into the
633    compiler. More details and numbers later.
634    
635    ----------------------------------------------------------------------
636    Name: Lal George
637    Date: 2002/03/11 11 22:38:53 EST
638    Tag: george-20020311-jump-chain-elim
639    Description:
640    
641    Tested the jump chain elimination on all architectures (except the
642    hppa).  This is on by default right now and is profitable for the
643    alpha and x86, however, it may not be profitable for the sparc and ppc
644    when compiling the compiler.
645    
646    The gc test will typically jump to a label at the end of the cluster,
647    where there is another jump to an external cluster containing the actual
648    code to invoke gc. This is to allow factoring of common gc invocation
649    sequences. That is to say, we generate:
650    
651            f:
652               testgc
653               ja   L1      % jump if above to L1
654    
655            L1:
656               jmp L2
657    
658    
659    After jump chain elimination the 'ja L1' instructions is converted to
660    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
661    up being implemented in their long form (if L2 is far away) using:
662    
663            jbe     L3      % jump if below or equal to L3
664            jmp     L2
665         L3:
666            ...
667    
668    
669    For large compilation units L2  may be far away.
670    
671    
672    ----------------------------------------------------------------------
673    Name: Matthias Blume
674    Date: 2002/03/11 13:30:00 EST
675    Tag: blume-20020311-mltreeeval
676    Description:
677    
678    A functor parameter was missing.
679    
680    ----------------------------------------------------------------------
681    Name: Allen Leung
682    Date: 2002/03/11 10:30:00 EST
683    Tag: leunga-20020311-runtime-string0
684    Description:
685    
686       The representation of the empty string now points to a
687    legal null terminated C string instead of unit.  It is now possible
688    to convert an ML string into C string with InlineT.CharVector.getData.
689    This compiles into one single machine instruction.
690    
691    ----------------------------------------------------------------------
692    Name: Allen Leung
693    Date: 2002/03/10 23:55:00 EST
694    Tag: leunga-20020310-x86-call
695    Description:
696    
697       Added machine generation for CALL instruction (relative displacement mode)
698    
699    ----------------------------------------------------------------------
700    Name: Matthias Blume
701    Date: 2002/03/08 16:05:00
702    Tag: blume-20020308-entrypoints
703    Description:
704    
705    Version number bumped to 110.39.1.  NEW BOOTFILES!
706    
707    Entrypoints: non-zero offset into a code object where execution should begin.
708    
709    - Added the notion of an entrypoint to CodeObj.
710    - Added reading/writing of entrypoint info to Binfile.
711    - Made runtime system bootloader aware of entrypoints.
712    - Use the address of the label of the first function given to mlriscGen
713      as the entrypoint.  This address is currently always 0, but it will
714      not be 0 once we turn on block placement.
715    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
716      with entry points) from mlriscGen.
717    
718    ----------------------------------------------------------------------
719    Name: Allen Leung
720    Date: 2002/03/07 20:45:00 EST
721    Tag: leunga-20020307-x86-cmov
722    Description:
723    
724       Bug fixes for CMOVcc on x86.
725    
726       1. Added machine code generation for CMOVcc
727       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
728       3. CMOVcc cannot have an immediate operand as argument.
729    
730    ----------------------------------------------------------------------
731    Name: Matthias Blume
732    Date: 2002/03/07 16:15:00 EST
733    Tag: blume-20020307-controls
734    Description:
735    
736    This is a very large but mostly boring patch which makes (almost)
737    every tuneable compiler knob (i.e., pretty much everything under
738    Control.* plus a few other things) configurable via both the command
739    line and environment variables in the style CM did its configuration
740    until now.
741    
742    Try starting sml with '-h' (or, if you are brave, '-H')
743    
744    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
745    implements the underlying generic mechanism.
746    
747    The interface to some of the existing such facilities has changed somewhat.
748    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
749    (The getFoo interface is still there for backward-compatibility, but its
750    use is deprecated.)
751    
752    The ml-build script passes -Cxxx=yyy command-line arguments through so
753    that one can now twiddle the compiler settings when using this "batch"
754    compiler.
755    
756    TODO items:
757    
758    We should go through and throw out all controls that are no longer
759    connected to anything.  Moreover, we should go through and provide
760    meaningful (and correct!) documentation strings for those controls
761    that still are connected.
762    
763    Currently, multiple calls to Controls.new are accepted (only the first
764    has any effect).  Eventually we should make sure that every control
765    is being made (via Controls.new) exactly once.  Future access can then
766    be done using Controls.acc.
767    
768    Finally, it would probably be a good idea to use the getter-setter
769    interface to controls rather than ref cells.  For the time being, both
770    styles are provided by the Controls module, but getter-setter pairs are
771    better if thread-safety is of any concern because they can be wrapped.
772    
773    *****************************************
774    
775    One bug fix: The function blockPlacement in three of the MLRISC
776    backpatch files used to be hard-wired to one of two possibilities at
777    link time (according to the value of the placementFlag).  But (I
778    think) it should rather sense the flag every time.
779    
780    *****************************************
781    
782    Other assorted changes (by other people who did not supply a HISTORY entry):
783    
784    1. the cross-module inliner now works much better (Monnier)
785    2. representation of weights, frequencies, and probabilities in MLRISC
786       changed in preparation of using those for weighted block placement
787       (Reppy, George)
788    
789    ----------------------------------------------------------------------
790    Name: Lal George
791    Date: 2002/03/07 14:44:24 EST 2002
792    Tag: george-20020307-weighted-block-placement
793    
794    Tested the weighted block placement optimization on all architectures
795    (except the hppa) using AMPL to generate the block and edge frequencies.
796    Changes were required in the machine properties to correctly
797    categorize trap instructions. There is an MLRISC flag
798    "weighted-block-placement" that can be used to enable weighted block
799    placement, but this will be ineffective without block/edge
800    frequencies (coming soon).
801    
802    
803    ----------------------------------------------------------------------
804    Name: Lal George
805    Date: 2002/03/05 17:24:48 EST
806    Tag: george-20020305-linkage-cluster
807    
808    In order to support the block placement optimization, a new cluster
809    is generated as the very first cluster (called the linkage cluster).
810    It contains a single jump to the 'real' entry point for the compilation
811    unit. Block placement has no effect on the linkage cluster itself, but
812    all the other clusters  have full freedom in the manner in which they
813    reorder blocks or functions.
814    
815    On the x86 the typical linkage code that is generated is:
816       ----------------------
817            .align 2
818       L0:
819            addl    $L1-L0, 72(%esp)
820            jmp     L1
821    
822    
823            .align  2
824       L1:
825       ----------------------
826    
827    72(%esp) is the memory location for the stdlink register. This
828    must contain the address of the CPS function being called. In the
829    above example, it contains the address of  L0; before
830    calling L1 (the real entry point for the compilation unit), it
831    must contain the address for L1, and hence
832    
833            addl $L1-L0, 72(%esp)
834    
835    I have tested this on all architectures except the hppa.The increase
836    in code size is of course negligible
837    
838    ----------------------------------------------------------------------
839    Name: Allen Leung
840    Date: 2002/03/03 13:20:00 EST
841    Tag: leunga-20020303-mlrisc-tools
842    
843      Added #[ ... ] expressions to mlrisc tools
844    
845    ----------------------------------------------------------------------
846    Name: Matthias Blume
847    Date: 2002/02/27 12:29:00 EST
848    Tag: blume-20020227-cdebug
849    Description:
850    
851    - made types in structure C and C_Debug to be equal
852    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
853    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
854    
855    ----------------------------------------------------------------------
856    Name: Matthias Blume
857    Date: 2002/02/26 12:00:00 EST
858    Tag: blume-20020226-ffi
859    Description:
860    
861    1. Fixed a minor bug in CM's "noweb" tool:
862       If numbering is turned off, then truly don't number (i.e., do not
863       supply the -L option to noweb).  The previous behavior was to supply
864       -L'' -- which caused noweb to use the "default" line numbering scheme.
865       Thanks to Chris Richards for pointing this out (and supplying the fix).
866    
867    2. Once again, I reworked some aspects of the FFI:
868    
869       A. The incomplete/complete type business:
870    
871       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
872         gone!
873       - ML types representing an incomplete type are now *equal* to
874         ML types representing their corresponding complete types (just like
875         in C).  This is still safe because ml-nlffigen will not generate
876         RTTI for incomplete types, nor will it generate functions that
877         require access to such RTTI.   But when ML code generated from both
878         incomplete and complete versions of the C type meet, the ML types
879         are trivially interoperable.
880    
881         NOTE:  These changes restore the full generality of the translation
882         (which was previously lost when I eliminated functorization)!
883    
884       B. Enum types:
885    
886       - Structure C now has a type constructor "enum" that is similar to
887         how the "su" constructor works.  However, "enum" is not a phantom
888         type because each "T enum" has values (and is isomorphic to
889         MLRep.Signed.int).
890       - There are generic access operations for enum objects (using
891         MLRep.Signed.int).
892       - ml-nlffigen will generate a structure E_foo for each "enum foo".
893         * The structure contains the definition of type "mlrep" (the ML-side
894         representation type of the enum).  Normally, mlrep is the same
895         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
896         then mlrep will be defined as a datatype -- thus facilitating
897         pattern matching on mlrep values.
898         ("-ec" will be suppressed if there are duplicate values in an
899          enumeration.)
900         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
901         will be generated for each C enum constant xxx.
902         * Conversion functions m2i and i2m convert between mlrep and
903         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
904         * Coversion functions c and ml convert between mlrep and "tag enum".
905         * Access functions (get/set) fetch and store mlrep values.
906       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
907         enumerations are merged into one single enumeration represented by
908         structure E_'.
909    
910    ----------------------------------------------------------------------
911    Name: Allen Leung
912    Date: 2002/02/25 04:45:00 EST
913    Tag: leunga-20020225-cps-spill
914    
915    This is a new implementation of the CPS spill phase.
916    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
917    In case of problems, replace it with the old file spill.sml
918    
919    The current compiler runs into some serious performance problems when
920    constructing a large record.  This can happen when we try to compile a
921    structure with many items.  Even a very simple structure like the following
922    makes the compiler slow down.
923    
924        structure Foo = struct
925           val x_1 = 0w1 : Word32.int
926           val x_2 = 0w2 : Word32.int
927           val x_3 = 0w3 : Word32.int
928           ...
929           val x_N = 0wN : Word32.int
930        end
931    
932    The following table shows the compile time, from N=1000 to N=4000,
933    with the old compiler:
934    
935    N
936    1000   CPS 100 spill                           0.04u  0.00s  0.00g
937           MLRISC ra                               0.06u  0.00s  0.05g
938              (spills = 0 reloads = 0)
939           TOTAL                                   0.63u  0.07s  0.21g
940    
941    1100   CPS 100 spill                           8.25u  0.32s  0.64g
942           MLRISC ra                               5.68u  0.59s  3.93g
943              (spills = 0 reloads = 0)
944           TOTAL                                   14.71u  0.99s  4.81g
945    
946    1500   CPS 100 spill                           58.55u  2.34s  1.74g
947           MLRISC ra                               5.54u  0.65s  3.91g
948              (spills = 543 reloads = 1082)
949           TOTAL                                   65.40u  3.13s  6.00g
950    
951    2000   CPS 100 spill                           126.69u  4.84s  3.08g
952           MLRISC ra                               0.80u  0.10s  0.55g
953              (spills = 42 reloads = 84)
954           TOTAL                                   129.42u  5.10s  4.13g
955    
956    3000   CPS 100 spill                           675.59u  19.03s  11.64g
957           MLRISC ra                               2.69u  0.27s  1.38g
958              (spills = 62 reloads = 124)
959           TOTAL                                   682.48u  19.61s  13.99g
960    
961    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
962           MLRISC ra                               4.96u  0.27s  2.72g
963              (spills = 85 reloads = 170)
964           TOTAL                                   2375.26u  57.21s  48.00g
965    
966    As you can see the old cps spill module suffers from some serious
967    performance problem.  But since I cannot decipher the old code fully,
968    instead of patching the problems up, I'm reimplementing it
969    with a different algorithm.  The new code is more modular,
970    smaller when compiled, and substantially faster
971    (O(n log n) time and O(n) space).  Timing of the new spill module:
972    
973    4000  CPS 100 spill                           0.02u  0.00s  0.00g
974          MLRISC ra                               0.25u  0.02s  0.15g
975             (spills=1 reloads=3)
976          TOTAL                                   7.74u  0.34s  1.62g
977    
978    Implementation details:
979    
980    As far as I can tell, the purpose of the CPS spill module is to make sure the
981    number of live variables at any program point (the bandwidth)
982    does not exceed a certain limit, which is determined by the
983    size of the spill area.
984    
985    When the bandwidth is too large, we decrease the register pressure by
986    packing live variables into spill records.  How we achieve this is
987    completely different than what we did in the old code.
988    
989    First, there is something about the MLRiscGen code generator
990    that we should be aware of:
991    
992    o MLRiscGen performs code motion!
993    
994       In particular, it will move floating point computations and
995       address computations involving only the heap pointer to
996       their use sites (if there is only a single use).
997       What this means is that if we have a CPS record construction
998       statement
999    
1000           RECORD(k,vl,w,e)
1001    
1002       we should never count the new record address w as live if w
1003       has only one use (which is often the case).
1004    
1005       We should do something similar to floating point, but the transformation
1006       there is much more complex, so I won't deal with that.
1007    
1008    Secondly, there are now two new cps primops at our disposal:
1009    
1010     1. rawrecord of record_kind option
1011        This pure operator allocates some uninitialized storage from the heap.
1012        There are two forms:
1013    
1014         rawrecord NONE [INT n]  allocates a tagless record of length n
1015         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
1016                                     and initializes the tag.
1017    
1018     2. rawupdate of cty
1019          rawupdate cty (v,i,x)
1020          Assigns to x to the ith component of record v.
1021          The storelist is not updated.
1022    
1023    We use these new primops for both spilling and increment record construction.
1024    
1025     1. Spilling.
1026    
1027        This is implemented with a linear scan algorithm (but generalized
1028        to trees).  The algorithm will create a single spill record at the
1029        beginning of the cps function and use rawupdate to spill to it,
1030        and SELECT or SELp to reload from it.  So both spills and reloads
1031        are fine-grain operations.  In contrast, in the old algorithm
1032        "spills" have to be bundled together in records.
1033    
1034        Ideally, we should sink the spill record construction to where
1035        it is needed.  We can even split the spill record into multiple ones
1036        at the places where they are needed.  But CPS is not a good
1037        representation for global code motion, so I'll keep it simple and
1038        am not attempting this.
1039    
1040     2. Incremental record construction (aka record splitting).
1041    
1042        Long records with many component values which are simulatenously live
1043        (recall that single use record addresses are not considered to
1044         be live) are constructed with rawrecord and rawupdate.
1045        We allocate space on the heap with rawrecord first, then gradually
1046        fill it in with rawupdate.  This is the technique suggested to me
1047        by Matthias.
1048    
1049        Some restrictions on when this is applicable:
1050        1. It is not a VECTOR record.  The code generator currently does not handle
1051           this case. VECTOR record uses double indirection like arrays.
1052        2. All the record component values are defined in the same "basic block"
1053           as the record constructor.  This is to prevent speculative
1054           record construction.
1055    
1056    ----------------------------------------------------------------------
1057    Name: Allen Leung
1058    Date: 2002/02/22 01:02:00 EST
1059    Tag: leunga-20020222-mlrisc-tools
1060    
1061    Minor bug fixes in the parser and rewriter
1062    
1063    ----------------------------------------------------------------------
1064    Name: Allen Leung
1065    Date: 2002/02/21 20:20:00 EST
1066    Tag: leunga-20020221-peephole
1067    
1068    Regenerated the peephole files.  Some contained typos in the specification
1069    and some didn't compile because of pretty printing bugs in the old version
1070    of 'nowhere'.
1071    
1072    ----------------------------------------------------------------------
1073    Name: Allen Leung
1074    Date: 2002/02/19 20:20:00 EST
1075    Tag: leunga-20020219-mlrisc-tools
1076    Description:
1077    
1078       Minor bug fixes to the mlrisc-tools library:
1079    
1080       1.  Fixed up parsing colon suffixed keywords
1081       2.  Added the ability to shut the error messages up
1082       3.  Reimplemented the pretty printer and fixed up/improved
1083           the pretty printing of handle and -> types.
1084       4.  Fixed up generation of literal symbols in the nowhere tool.
1085       5.  Added some SML keywords to to sml.sty
1086    
1087    ----------------------------------------------------------------------
1088    Name: Matthias Blume
1089    Date: 2002/02/19 16:20:00 EST
1090    Tag: blume-20020219-cmffi
1091    Description:
1092    
1093    A wild mix of changes, some minor, some major:
1094    
1095    * All C FFI-related libraries are now anchored under $c:
1096        $/c.cm      --> $c/c.cm
1097        $/c-int.cm  --> $c/internals/c-int.cm
1098        $/memory.cm --> $c/memory/memory.cm
1099    
1100    * "make" tool (in CM) now treats its argument pathname slightly
1101      differently:
1102        1. If the native expansion is an absolute name, then before invoking
1103           the "make" command on it, CM will apply OS.Path.mkRelative
1104           (with relativeTo = OS.FileSys.getDir()) to it.
1105        2. The argument will be passed through to subsequent phases of CM
1106           processing without "going native".  In particular, if the argument
1107           was an anchored path, then "make" will not lose track of that anchor.
1108    
1109    * Compiler backends now "know" their respective C calling conventions
1110      instead of having to be told about it by ml-nlffigen.  This relieves
1111      ml-nlffigen from one of its burdens.
1112    
1113    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
1114    
1115    * Export C_DEBUG and C_Debug from $c/c.cm.
1116    
1117    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
1118      subtyping relationship between incomplete pointers and their complete
1119      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
1120      with the convention of instantiating 'o with ('t, 'c) obj whenever
1121      the pointer target type is complete.  In the incomplete case, 'o
1122      will be instantiated with some "'c iobj" -- a type obtained by
1123      using one of the functors PointerToIncompleteType or PointerToCompleteType.
1124    
1125      Operations that work on both incomplete and complete pointer types are
1126      typed as taking an 'o ptr while operations that require the target to
1127      be known are typed as taking some ('t, 'c) obj ptr.
1128    
1129      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
1130      where void is an eqtype without any values.  This makes it possible
1131      to work on voidptr values using functions meant to operate on light
1132      incomplete pointers.
1133    
1134    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
1135      been vastly simplified.
1136    
1137    ----------------------------------------------------------------------
1138    Name: Matthias Blume
1139    Date: 2002/02/19 10:48:00 EST
1140    Tag: blume-20020219-pqfix
1141    Description:
1142    
1143    Applied Chris Okasaki's bug fix for priority queues.
1144    
1145    ----------------------------------------------------------------------
1146    Name: Matthias Blume
1147    Date: 2002/02/15 17:05:00
1148    Tag: Release_110_39
1149    Description:
1150    
1151    Last-minute retagging is becoming a tradition... :-(
1152    
1153    This is the working release 110.39.
1154    
1155    ----------------------------------------------------------------------
1156    Name: Matthias Blume
1157    Date: 2002/02/15 16:00:00 EST
1158    Tag: Release_110_39-orig
1159    Description:
1160    
1161    Working release 110.39.  New bootfiles.
1162    
1163    (Update: There was a small bug in the installer so it wouldn't work
1164    with all shells.  So I retagged. -Matthias)
1165    
1166    ----------------------------------------------------------------------
1167    Name: Matthias Blume
1168    Date: 2002/02/15 14:17:00 EST
1169    Tag: blume-20020215-showbindings
1170    Description:
1171    
1172    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
1173    the latter can be useful for exploring what bindings are available at
1174    the interactive prompt.  (The first function returns only the list
1175    of symbols that are really bound, the second prints those but also the
1176    ones that CM's autoloading mechanism knows about.)
1177    
1178    ----------------------------------------------------------------------
1179    Name: Matthias Blume
1180    Date: 2002/02/15 12:08:00 EST
1181    Tag: blume-20020215-iptrs
1182    Description:
1183    
1184    Two improvements to ml-nlffigen:
1185    
1186      1. Write files only if they do not exist or if their current contents
1187         do not coincide with what's being written.  (That is, avoid messing
1188         with the time stamps unless absolutely necessary.)
1189    
1190      2. Implement a "repository" mechanism for generated files related
1191         to "incomplete pointer types".   See the README file for details.
1192    
1193    ----------------------------------------------------------------------
1194    Name: Matthias Blume
1195    Date: 2002/02/14 11:50:00 EST
1196    Tag: blume-20020214-quote
1197    Description:
1198    
1199    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
1200    because of the new and improved tag generation scheme.  (Thanks to Allen
1201    Leung for pointing it out.)
1202    
1203    ----------------------------------------------------------------------
1204    Name: Lal George
1205    Date: 2002/02/14 09:55:27 EST 2002
1206    Tag: george-20020214-isabelle-bug
1207    Description:
1208    
1209    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
1210    of Isabelle on the x86.
1211    
1212    From Allen:
1213    -----------
1214     I've found the problem:
1215    
1216         in ra-core.sml, I use the counter "blocked" to keep track of the
1217         true number of elements in the freeze queue.  When the counter goes
1218         to zero, I skip examining the queue.  But I've messed up the
1219         bookkeeping in combine():
1220    
1221             else ();
1222             case !ucol of
1223               PSEUDO => (if !cntv > 0 then
1224                     (if !cntu > 0 then blocked := !blocked - 1 else ();
1225                                        ^^^^^^^^^^^^^^^^^^^^^^^
1226                      moveu := mergeMoveList(!movev, !moveu)
1227                     )
1228                  else ();
1229    
1230         combine() is called to coalesce two nodes u and v.
1231         I think I was thinking that if the move counts of u and v are both
1232         greater than zero then after they are coalesced then one node is
1233         removed from the freeze queue.  Apparently I was thinking that
1234         both u and v are of low degree, but that's clearly not necessarily true.
1235    
1236    
1237    02/12/2002:
1238        Here's the patch.  HOL now compiles.
1239    
1240        I don't know how this impact on performance (compile
1241        time or runtime).  This bug caused the RA (especially on the x86)
1242        to go thru the potential spill phase when there are still nodes on the
1243        freeze queue.
1244    
1245    
1246    
1247    
1248    ----------------------------------------------------------------------
1249    Name: Matthias Blume
1250    Date: 2002/02/13 22:40:00 EST
1251    Tag: blume-20020213-fptr-rtti
1252    Description:
1253    
1254    Fixed a bug in ml-nlffigen that was introduced with one of the previous
1255    updates.
1256    
1257    ----------------------------------------------------------------------
1258    Name: Matthias Blume
1259    Date: 2002/02/13 16:41:00 EST
1260    Tag: blume-20020213-cmlpq
1261    Description:
1262    
1263    Added new priority queue export symbols (which have just been added to
1264    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
1265    not compile and the installer would choke.)
1266    
1267    ----------------------------------------------------------------------
1268    Name: Matthias Blume
1269    Date: 2002/02/13 16:15:00 EST
1270    Tag: blume-20020213-various
1271    Description:
1272    
1273    1. More tweaks to ml-nlffigen:
1274    
1275       - better internal datastructures (resulting in slight speedup)
1276       - "-match" option requires exact match
1277       - "localized" gensym counters (untagged structs/unions nested within
1278         other structs/unions or within typedefs get a fresh counter; their
1279         tag will be prefixed by a concatenation of their parents' tags)
1280       - bug fixes (related to calculation of transitive closure of types
1281         to be included in the output)
1282    
1283    2. Minor Basis updates:
1284    
1285       - added implementations for List.collate and Option.app
1286    
1287    ----------------------------------------------------------------------
1288    Name: Matthias Blume
1289    Date: 2002/02/11 15:55:00 EST
1290    Tag: blume-20020211-gensym
1291    Description:
1292    
1293    Added a "-gensym" option to command line of ml-nlffigen.  This can be
1294    used to specify a "stem" -- a string that is inserted in all "gensym'd"
1295    names (ML structure names that correspond to unnamed C structs, unions,
1296    and enums), so that separate runs of ml-nlffigen do not clash.
1297    
1298    ----------------------------------------------------------------------
1299    Name: Matthias Blume
1300    Date: 2002/02/11 12:05:00 EST
1301    Tag: blume-20020211-gensml
1302    Description:
1303    
1304    A quick fix for a problem with GenSML (in the pgraph-util library):
1305    Make generation of toplevel "local" optional.  (Strictly speaking,
1306    signature definitions within "local" are not legal SML.)
1307    
1308    Other than that: updates to INSTALL and cm/TODO.
1309    
1310    ----------------------------------------------------------------------
1311    Name: Matthias Blume
1312    Date: 2002/02/08 15:00:00 EST
1313    Tag: blume-20020208-uniquepid
1314    Description:
1315    
1316    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
1317    
1318    1. The installer (config/install.sh) has gotten smarter:
1319    
1320         - Configuration options are a bit easier to specify now
1321           (in config/targets).
1322         - Bug in recognizing .tar.bz2 files fixed.
1323         - Installer automatically resolves dependencies between
1324           configuration options (e.g., if you ask for eXene, you will
1325           also get cml -- regardless whether you asked for it or not).
1326         - Installer can run in "quieter mode" by setting the environment
1327           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
1328           "completely silent", though.
1329         - Build HashCons library as part of smlnj-lib.
1330    
1331    2. A new scheme for assigning persistent identifiers to compilation
1332       units (and, by extension, to types etc.) has been put into place.
1333       This fixes a long-standing bug where types and even dynamic values
1334       can get internally confused, thereby compromising type safety
1335       (abstraction) and dynamic correctness.  See
1336    
1337         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
1338    
1339       for an example of how things could go wrong until now.
1340    
1341       The downside of the new scheme is that pids are not quite as
1342       persistent as they used to be: CM will generate a fresh pid
1343       for every compilation unit that it thinks it sees for the first
1344       time.  That means that if you compile starting from a clean, fresh
1345       source tree at two different times, you end up with different
1346       binaries.
1347    
1348       Cutoff recompilation, however, has not been compromised because
1349       CM keeps pid information in special caches between runs.
1350    
1351    ----------------------------------------------------------------------
1352    Name: Lal George
1353    Date: 2002/02/07 15:34:13 EST 2002
1354    Tag: <none>
1355    Description:
1356    
1357    Compilers that generate assembly code may produce  global labels
1358    whose value is resolved at link time. The various peephole optimization
1359    modules did not take this in account.
1360    
1361    TODO. The Labels.addrOf function should really return an option
1362    type so that clients are forced to deal with this issue, rather
1363    than an exception being raised.
1364    
1365    ----------------------------------------------------------------------
1366    Name: Lal George
1367    Date: 2002/02/06 13:55:02 EST
1368    Tag: george-20020206-ra-breakup
1369    Description:
1370    
1371    1. A bug fix from Allen.
1372    
1373        A typo causes extra fstp %st(0)'s to be generated at compensation
1374        edges, which might cause stack underflow traps at runtime.  This
1375        occurs in fft where there are extraneous fstps right before the 'into'
1376        trap instruction (in this case they are harmless since none of the
1377        integers overflow.)
1378    
1379    2. Pulled out various utility modules that were embedded in the modules
1380       of the register allocator. I need these modules for other purposes, but
1381       they are not complete enough to put into a library (just yet).
1382    ----------------------------------------------------------------------
1383    Name: Matthias Blume
1384    Date: 2002/01/31 16:05:00 EST
1385    Tag: blume-20020131-sparc-ccalls
1386    Description:
1387    
1388    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
1389       of extra stack space by mistake.  Fixed.
1390    
1391    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
1392    
1393    ----------------------------------------------------------------------
1394    Name: Allen Leung
1395    Date: 2002/01/30
1396    Tag: leunga-20020130-nowhere-bug-fix
1397    Description:
1398    
1399       MLRISC bug fixes:
1400       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
1401       2. MachineInt.fromString was negating its value.
1402    
1403    ----------------------------------------------------------------------
1404    Name: Matthias Blume
1405    Date: 2002/01/29
1406    Tag: blume-20020129-INSTALL
1407    Description:
1408    
1409    - Added somewhat detailed installation instructions (file INSTALL).
1410    - Fixed curl-detection bug in config/install.sh.
1411    - It is now possible to select the URL getter using the URLGETTER
1412      environment variable:
1413    
1414          not set / "unknown"      --> automatic detection (script tries wget,
1415                                       curl, and lynx)
1416          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
1417                                       how to properly invoke them)
1418          other                    --> use $URLGETTER directly, it must take
1419                                       precisely two command-line arguments
1420                                       (source URL and destination file name)
1421    
1422    ----------------------------------------------------------------------
1423    Name: Matthias Blume
1424    Date: 2002/01/28
1425    Tag: blume-20020128-sparc-ccalls
1426    Description:
1427    
1428    - Fixed problem with calculation of "used" registers in sparc-c-calls.
1429    - Make use of the allocParam argument in sparc-c-calls.
1430    
1431    ----------------------------------------------------------------------
1432    Name: Matthias Blume
1433    Date: 2002/01/28
1434    Tag: blume-20020128-allocParam
1435    Description:
1436    
1437    John Reppy:  Changes c-calls API to accept client-callback for
1438    allocating extra stack space.
1439    me: Corresponding changes to mlriscGen (using a dummy argument that
1440        does not change the current behavior).
1441    
1442    ----------------------------------------------------------------------
1443    Name: Matthias Blume
1444    Date: 2002/01/28 12:00:00
1445    Tag: Release_110_38
1446    Description:
1447    
1448    This time for real!!!
1449    
1450    ----------------------------------------------------------------------
1451    Name: Matthias Blume
1452    Date: 2002/01/28 10:56:00 EST
1453    Tag: blume-20020128-retraction
1454    Description:
1455    
1456    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
1457       with blume-Release_110_38-retracted.)
1458    
1459    1. Fixed a problem with incorrect rounding modes in real64.sml.
1460       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
1461    
1462    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
1463       and enums fixed.  The naming of corresponding ML identifiers should
1464       now be consistent again.
1465    
1466    ----------------------------------------------------------------------
1467    Name: Allen Leung
1468    Date: 2002/01/27
1469    Tag: leunga-20020127-nowhere
1470    Description:
1471    
1472       Added a target called nowhere in the configuration scripts.
1473       Enabling this will build the MLRISC 'nowhere' tool (for translating
1474       programs with where-clauses into legal SML code) during installation.
1475    
1476    ----------------------------------------------------------------------
1477    Name: Matthias Blume
1478    Date: 2002/01/25 21:27:00 EST
1479    Tag: blume-Release_110_38-retracted
1480    Description:
1481    
1482    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
1483    
1484    README will be added later.
1485    
1486    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
1487    (Release_110_38) removed.  Reason: Last-minute bug fixes.
1488    
1489    ----------------------------------------------------------------------
1490    Name: Matthias Blume
1491    Date: 2002/01/25
1492    Tag: blume-20020125-ffi
1493    Description:
1494    
1495    A large number of tweaks and improvements to ml-nlffi-lib and
1496    ml-nlffigen:
1497    
1498       - ML represenation types have been streamlined
1499       - getter and setter functions work with concrete values, not abstract
1500         ones where possible
1501       - ml-nlffigen command line more flexible (see README file there)
1502       - some bugs have been fixed (hopefully)
1503    
1504    ----------------------------------------------------------------------
1505    Name: Lal George
1506    Date: 2002/01/24
1507    Tag: george-20020124-risc-ra-interface
1508    Description:
1509    
1510       There is a dramatic simplification in the interface to the
1511       register allocator for RISC architectures as a result of making
1512       parallel copy instructions explicit.
1513    
1514    ----------------------------------------------------------------------
1515    Name: Matthias Blume
1516    Date: 2002/01/22
1517    Tag: blume-20020122-x86-ccalls
1518    Description:
1519    
1520    Bug fix for c-calls on x86 (having to do with how char- and
1521    short-arguments are being handled).
1522    
1523    ----------------------------------------------------------------------
1524    Name: Matthias Blume
1525    Date: 2002/01/21
1526    Tag: blume-20020121-ff
1527    Description:
1528    
1529    Another day of fiddling with the FFI...
1530    
1531    1. Bug fix/workaround:  CKIT does not complain about negative array
1532       dimensions, so ml-nlffigen has to guard itself against this possibility.
1533       (Otherwise a negative dimension would send it into an infinite loop.)
1534    
1535    2. Some of the abstract types (light objects, light pointers, most "base"
1536       types) in structure C are now eqtypes.
1537    
1538    3. Added constructors and test functions for NULL function pointers.
1539    
1540    ----------------------------------------------------------------------
1541    Name: Matthias Blume
1542    Date: 2002/01/18
1543    Tag: blume-20020118-ready-for-new-release
1544    Description:
1545    
1546    Made config/srcarchiveurl point to a new place.  (Will provide boot
1547    files shortly.)
1548    
1549    Maybe we christen this to be 110.38?
1550    
1551    ----------------------------------------------------------------------
1552    Name: Matthias Blume
1553    Date: 2002/01/18
1554    Tag: blume-20020118-more-ffifiddle
1555    Description:
1556    
1557    Today's FFI fiddling:
1558    
1559      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
1560        C.Set.* that use concrete (MLRep.*) arguments and results instead
1561        of abstract ones.
1562    
1563      - Provided word-style bit operations etc. for "int" representation
1564        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
1565        and Long.
1566    
1567    ----------------------------------------------------------------------
1568    Name: Matthias Blume
1569    Date: 2002/01/18
1570    Tag: blume-20020118-use-x86-fp
1571    Description:
1572    
1573    Now that x86-fast-fp seems to be working, I turned it back on again
1574    by default.  (Seems to work fine now, even with the FFI.)
1575    
1576    Other than that, I added some documentation about the FFI to
1577    src/ml-nlffigen/README and updated the FFI test examples in
1578    src/ml-nlffi-lib/Tests/*.
1579    
1580    ----------------------------------------------------------------------
1581    Name: Allen Leung
1582    Date: 2002/01/17
1583    Tag: leunga-20020117-x86-fast-fp-call
1584    Description:
1585    
1586       1. Fixed a problem with handling return fp values when x86's fast fp
1587          mode is turned on.
1588    
1589       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
1590          of %f32.
1591    
1592       3. Added a constructor INT32lit to the ast of MLRISC tools.
1593    
1594    ----------------------------------------------------------------------
1595    Name: Matthias Blume
1596    Date: 2002/01/16
1597    Tag: blume-20020116-ffifiddle
1598    Description:
1599    
1600    More fiddling with the FFI interface:
1601    
1602     - Make constness 'c instead of rw wherever possible.  This eliminates
1603       the need for certain explicit coercions.  (However, due to ML's
1604       value polymorphism, there will still be many cases where explicit
1605       coercions are necessary.  Phantom types are not the whole answer
1606       to modeling a subtyping relationship in ML.)
1607    
1608     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
1609    
1610     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
1611       clumsy workaround for varargs problem.)
1612    
1613    ----------------------------------------------------------------------
1614    Name: Lal George
1615    Date: 2002/01/15
1616    Tag: <none>
1617    Description:
1618    
1619    1. Since COPY instructions are no longer native to the architecture,
1620       a generic functor can be used to implement the expandCopies function.
1621    
1622    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
1623       TEXT segment.
1624    
1625    ----------------------------------------------------------------------
1626    Name: Matthias Blume
1627    Date: 2002/01/15
1628    Tag: blume-20020115-ffiupdates
1629    Description:
1630    
1631    1. Fix for bug resulting in single-precision float values being returned
1632       incorrectly from FFI calls.
1633    
1634    2. Small modifications to C FFI API:
1635    
1636        - memory-allocation routines return straight objects (no options)
1637          and raise an exception in out-of-memory situations
1638        - unsafe extensions to cast between function pointers and pointers
1639          from/to ints
1640        - added structure C_Debug as an alternative to structure C where
1641          pointer-dereferencing (|*| and |*!) always check for null-pointers
1642        - added open_lib' to DynLinkage;  open_lib' works like open_lib
1643          but also takes a (possibly empty) list of existing library handles
1644          that the current library depends on
1645    
1646    ----------------------------------------------------------------------
1647    Name: Matthias Blume
1648    Date: 2002/01/10
1649    Tag: blume-20020110-newffigen
1650    Description:
1651    
1652    1. Updates to portable graph code.
1653    
1654    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
1655       more scalable now so that even huge interfaces such as the one
1656       for GTK compile in finite time and space. :-)
1657       See src/ml-nlffigen/README for details on what's new.
1658    
1659    ----------------------------------------------------------------------
1660    Name: Lal George
1661    Date: 2001/01/09 14:31:35 EST 2002
1662    Tag: george-20011206-rm-native-copy
1663    Description:
1664    
1665            Removed the native COPY and FCOPY instructions
1666            from all the architectures and replaced it with the
1667            explicit COPY instruction from the previous commit.
1668    
1669            It is now possible to simplify many of the optimizations
1670            modules that manipulate copies. This has not been
1671            done in this change.
1672    
1673    ----------------------------------------------------------------------
1674    Name: Lal George
1675    Date: 2001/12/06 16:50:13 EST 2001
1676    Tag: george-20011206-mlrisc-instruction
1677    Description:
1678    
1679    Changed the representation of instructions from being fully abstract
1680    to being partially concrete. That is to say:
1681    
1682      from
1683            type instruction
1684    
1685      to
1686            type instr                              (* machine instruction *)
1687    
1688            datatype instruction =
1689                LIVE of {regs: C.cellset, spilled: C.cellset}
1690              | KILL of {regs: C.cellset, spilled: C.cellset}
1691              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
1692              | ANNOTATION of {i: instruction, a: Annotations.annotation}
1693              | INSTR of instr
1694    
1695    This makes the handling of certain special instructions that appear on
1696    all architectures easier and uniform.
1697    
1698    LIVE and KILL say that a list of registers are live or killed at the
1699    program point where they appear. No spill code is generated when an
1700    element of the 'regs' field is spilled, but the register is moved to
1701    the 'spilled' (which is present, more for debugging than anything else).
1702    
1703    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
1704    We used to generate:
1705    
1706            DEFFREG f1
1707            f1 := f2 + f3
1708            trapb
1709    
1710    but now generate:
1711    
1712            f1 := f2 + f3
1713            trapb
1714            LIVE {regs=[f1,f2,f3], spilled=[]}
1715    
1716    Furthermore, the DEFFREG (hack) required that all floating point instruction
1717    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
1718    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
1719    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
1720    this architecture).
1721    
1722    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
1723    all the architectures. This will result in further simplification of the
1724    register allocator that must be aware of them for coalescing purposes, and
1725    will also simplify certain aspects of the machine description that provides
1726    callbacks related to parallel copies.
1727    
1728    ANNOTATION should be obvious, and now INSTR represents the honest to God
1729    machine instruction set!
1730    
1731    The <arch>/instructions/<arch>Instr.sml files define certain utility
1732    functions for making porting easier -- essentially converting upper case
1733    to lower case. All machine instructions (of type instr) are in upper case,
1734    and the lower case form generates an MLRISC instruction. For example on
1735    the alpha we have:
1736    
1737      datatype instr =
1738         LDA of {r:cell, b:cell, d:operand}
1739       | ...
1740    
1741      val lda : {r:cell, b:cell, d:operand} -> instruction
1742        ...
1743    
1744    where lda is just (INSTR o LDA), etc.
1745    
1746    ----------------------------------------------------------------------
1747    Name: Matthias Blume
1748    Date: 2001/11/22 21:40:00 EST
1749    Tag: Release_110_37
1750    Description:
1751    
1752    Release 110.37.  This time for real.
1753    
1754    ----------------------------------------------------------------------
1755    Name: Matthias Blume
1756    Date: 2001/11/21 16:35:00 EST
1757    Tag: blume-20011121-foot-in-mouth
1758    Description:
1759    
1760    Removed the "Release_110_37" tag because of a serious bug.
1761    This will be re-tagged once the bug is fixed.
1762    
1763    ----------------------------------------------------------------------
1764  Name: Matthias Blume  Name: Matthias Blume
1765  Date: 2001/11/21 16:14:00 EST  Date: 2001/11/21 16:14:00 EST
1766  Tag: blume-20011121-forgottenfile  Tag: blume-20011121-forgottenfile
# Line 24  Line 1772 
1772  ----------------------------------------------------------------------  ----------------------------------------------------------------------
1773  Name: Matthias Blume  Name: Matthias Blume
1774  Date: 2001/11/21 16:10:00 EST  Date: 2001/11/21 16:10:00 EST
1775  Tag: Release_110_37  Tag: blume-20011121-invalid_110_37
1776  Description:  Description:
1777    
1778    Note: I removed the original tag "Release_110_37" from this commit
1779          because we found a serious bug in all non-x86 backends.
1780          - Matthias
1781    
1782  1. Modifications to the SML/NJ code generator and to the runtime system  1. Modifications to the SML/NJ code generator and to the runtime system
1783     so that code object name strings are directly inserted into code     so that code object name strings are directly inserted into code
1784     objects at code generation time.  The only business the runtime system     objects at code generation time.  The only business the runtime system
# Line 248  Line 2000 
2000    
2001  ----------------------------------------------------------------------  ----------------------------------------------------------------------
2002  Name: Matthias Blume  Name: Matthias Blume
 >>>>>>> 1.169  
2003  Date: 2001/09/18 15:35:00 EDT  Date: 2001/09/18 15:35:00 EDT
2004  Tag: blume-20010918-readme11036  Tag: blume-20010918-readme11036
2005  Description:  Description:

Legend:
Removed from v.987  
changed lines
  Added in v.1192

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