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 1155, Wed Mar 20 20:52:51 2002 UTC revision 1249, Tue Jun 25 21:23:40 2002 UTC
# Line 14  Line 14 
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16  Name: Matthias Blume  Name: Matthias Blume
17    Date: 2002/06/25 17:23:00 EDT
18    Tag: blume-20020625-fpmax
19    Description:
20    
21    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
22    case.  (The sense of the isNaN test was reversed -- which made min and
23    max always return their first argument.)
24    
25    ----------------------------------------------------------------------
26    Name: Matthias Blume
27    Date: 2002/06/11
28    Tag: blume-20020611-unixpath
29    Description:
30    
31    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
32    
33    ----------------------------------------------------------------------
34    Name: Matthias Blume
35    Date: 2002/06/10 16:35:00 EDT
36    Tag: blume-20020610-ieeereal
37    Description:
38    
39    I back-ported my implementation of IEEEReal.fromString from the
40    idlbasis-devel branch so that we can test it.
41    
42    Another small change is that ppDec tries to give more information
43    than just "<sig>" in the case of functors.  However, this code is
44    broken in some mysterious way if the functor's body's signature
45    has not been declared by ascription but gets inferred from the
46    implementation.  This needs fixing...
47    
48    ----------------------------------------------------------------------
49    Name: Matthias Blume
50    Date: 2002/05/31
51    Tag: blume-20020531-btrace-mode
52    Description:
53    
54    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
55    the wayside when I switched over to using Controls everywhere.)
56    
57    ----------------------------------------------------------------------
58    Name: Lal George
59    Date: 2002/05/23 12:21:40 EDT
60    Tag: george-20020523-visual-labels
61    Description:
62    
63    Labels are now displayed in the graphical output to make
64    the fall-through and target blocks obvious.
65    
66    ----------------------------------------------------------------------
67    Name: Matthias Blume
68    Date: 2002/05/22 11:03:00 EDT
69    Tag: blume-20020522-shrink
70    Description:
71    
72    John tweaked yesterday's fix for 1131 to handle an out-of-memory
73    situation that comes up when allocating huge arrays.
74    
75    ----------------------------------------------------------------------
76    Name: Matthias Blume
77    Date: 2002/05/21 16:00:00 EDT
78    Tag: Release_110_40
79    Description:
80    
81    New working release (110.40).  New bootfiles.
82    
83    [Also: John Reppy fixed GC bug 1131.]
84    
85    ----------------------------------------------------------------------
86    Name: Matthias Blume
87    Date: 2002/05/21 12:35:00 EDT
88    Tag: blume-20020521-cmdoc
89    Description:
90    
91    CM documentation update.
92    
93    ----------------------------------------------------------------------
94    Name: Matthias Blume
95    Date: 2002/05/21 10:55:00 EDT
96    Tag: blume-20020521-misc
97    Description:
98    
99    - John tweaked runtime to be silent on heap export (except when
100      GC messages are on).
101    - I added a few more things (cross-compiling versions of CMB) to
102      config/preloads (as suggestions).
103    
104    ----------------------------------------------------------------------
105    Name: Matthias Blume
106    Date: 2002/05/20 22:25:00 EDT
107    Tag: blume-20020520-controls
108    Description:
109    
110    - Added ControlUtil structure to control-lib.cm.
111    - Use it throughout.
112    - Used Controls facility to define MLRISC controls (as opposed to
113      registering MLRISC control ref cells with Controls after the fact)
114    - Fixed messed-up controls priorities.
115    
116    * Removed again all the stuff from config/preloads that one wouldn't
117      be able to preload at the time the initial heap image is built.
118      (Many libraries, e.g., CML, do not exist yet at this time.  The
119       only libraries that can be preloaded via config/preloads are those
120       that come bundled with the bootfiles.)
121    
122    ----------------------------------------------------------------------
123    Name: Matthias Blume
124    Date: 2002/05/20 10:59:00 EDT
125    Tag: blume-20020520-preloads
126    Description:
127    
128    Added a lot of commented-out suggestions for things to be included
129    in config/preloads.
130    
131    ----------------------------------------------------------------------
132    Name: Allen Leung
133    Date: 2002/05/18 14:20:00 EDT
134    Tag: leunga-20020518-mdl
135    Description:
136    
137    o Made the mdl tool stuff compile and run again.
138    o I've disabled all the stuff that depends on RTL specifications; they
139      are all badly broken anyway.
140    
141    ----------------------------------------------------------------------
142    Name: Matthias Blume
143    Date: 2002/05/17 16:49:00 EDT
144    Tag: blume-20020517-controls
145    Description:
146    
147    0. John Reppy made several modifications to the SML/NJ library.
148       In particular, there is a shiny new controls-lib.cm.
149    
150    1. Pushed new controls interface through compiler so that everything
151       compiles again.
152    
153    2. Added FormatComb and FORMAT_COMB to the CML version of the
154       SML/NJ library (so that CML compiles again).
155    
156    3. Modified init scripts because XXX_DEFAULT environment variables
157       are no longer with us.  (Boot-time initialization is now done
158       using the same environment variables that are also used for
159       startup-time initialization of controls.)
160    
161    ----------------------------------------------------------------------
162    Name: Lal George
163    Date: 2002/05/15 09:20:10 EDT
164    Tag: george-20020515-pseudo-op-decls
165    Description:
166    
167            All pseudo-ops emitted before the first segment declaration
168    such as TEXT, DATA, and BSS directives are assumed to be global
169    declarations and are emitted first in the assembly file. This is
170    useful in a number of situations where one has pseudo-ops that are not
171    specific to any segment, and also works around the constraint that one
172    cannot have client pseudo-ops in the TEXT segment.
173    
174            Because no segment is associated with these declarations it is
175    an error to allocate any space or objects before the first segment
176    directive and an exception will be raised. However, we cannot make
177    this check for client pseudo-ops.
178    
179            These top level declarations are a field in the CFG graph_info.
180    In theory you can continue to add to this field after the CFG has been
181    built -- provided you know what you are doing;-)
182    
183    ----------------------------------------------------------------------
184    Name: Matthias Blume
185    Date: 2002/05/13 16:40:00 EDT
186    Tag: blume-20020513-pp-etc
187    Description:
188    
189    A few minor bugfixes:
190    
191      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
192        (Bogus printouts for redefined bindings still occur.  Compiler
193        bug should no longer occur now.  We need to redo the prettyprinter
194        from scratch.)
195    
196      - CM pathname printer now also adds escape sequences for ( and )
197    
198      - commend and docu fixes for ml-nlffi
199    
200    ----------------------------------------------------------------------
201    Name: Matthias Blume
202    Date: 2002/05/10 16:40:00 EDT
203    Tag: blume-20020510-erg-textio
204    Description:
205    
206    Applied the following bugfix provided by Emden Gansner:
207    
208        Output is corrupted when outputSubstr is used rather than output.
209    
210        The problem occurs when a substring
211    
212            ss = (s, dataStart, dataLen)
213    
214        where dataStart > 0, fills a stream buffer with avail bytes left.
215        avail bytes of s, starting at index dataStart, are copied into the
216        buffer, the buffer is flushed, and then the remaining dataLen-avail
217        bytes of ss are copied into the beginning of the buffer. Instead of
218        starting this copy at index dataStart+avail in s, the current code
219        starts the copy at index avail.
220    
221        Fix:
222        In text-io-fn.sml, change line 695 from
223             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
224        to
225             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
226    
227    ----------------------------------------------------------------------
228    Name: Matthias Blume
229    Date: 2002/04/12 13:55:00 EDT
230    Tag: blume-20020412-assyntax
231    Description:
232    
233    1. Grabbed newer assyntax.h from the XFree86 project.
234    2. Fiddled with how to compile X86.prim.asm without warnings.
235    3. (Very) Minor cleanup in CM.
236    
237    ----------------------------------------------------------------------
238    Name: Matthias Blume
239    Date: 2002/04/01 (no joke!) 17:07:00 EST
240    Tag: blume-20020401-x86div
241    Description:
242    
243    Added full support for div/mod/rem/quot on the x86, using the machine
244    instruction's two results (without clumsily recomputing the remainder)
245    directly where appropriate.
246    
247    Some more extensive power-of-two support was added to the x86 instruction
248    selector (avoiding expensive divs, mods, and muls where they can be
249    replaced with cheaper shifts and masks).  However, this sort of thing
250    ought to be done earlier, e.g., within the CPS optimizer so that
251    all architectures benefit from it.
252    
253    The compiler compiles to a fixed point, but changes might be somewhat
254    fragile nevertheless.  Please, report any strange things that you might
255    see wrt. div/mod/quot/rem...
256    
257    ----------------------------------------------------------------------
258    Name: Matthias Blume
259    Date: 2002/03/29 17:22:00
260    Tag: blume-20020329-div
261    Description:
262    
263    Fixed my broken div/mod logic.  Unfortunately, this means that the
264    inline code for div/mod now has one more comparison than before.
265    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
266    The problem was with quotient = 0, because that alone does not tell
267    us which way the rounding went.  One then has to look at whether
268    remainder and divisor have the same sign...  :(
269    
270    Anyway, I replaced the bootfiles with fresh ones...
271    
272    ----------------------------------------------------------------------
273    Name: Matthias Blume
274    Date: 2002/03/29 14:10:00 EST
275    Tag: blume-20020329-inlprims
276    Description:
277    
278    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
279    
280    Primops have changed. This means that the bin/boot-file formats have
281    changed as well.
282    
283    To make sure that there is no confusion, I made a new version.
284    
285    
286    CHANGES:
287    
288    * removed REMT from mltree (remainder should never overflow).
289    
290    * added primops to deal with divisions of all flavors to the frontend
291    
292    * handled these primops all the way through so they map to their respective
293      MLRISC support
294    
295    * used these primops in the implementation of Int, Int32, Word, Word32
296    
297    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
298    
299    * parameterized INLMIN, INLMAX, and INLABS by a numkind
300    
301    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
302      floating point
303    
304    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
305      and Real (but Real.abs maps to a separate floating-point-only primop)
306    
307    
308    TODO items:
309    
310    * Hacked Alpha32 instruction selection, disabling the selection of REMx
311      instructions because the machine instruction encoder cannot handle
312      them.  (Hppa, PPC, and Sparc instruction selection did not handle
313      REM in the first place, and REM is supported by the x86 machine coder.)
314    
315    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
316      selection phase.  (The two can be streamlined because the hardware
317      delivers both quotient and remainder at the same time anyway.)
318    
319    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
320      (Currently the behavior is inconsistent both across architectures and
321      wrt. the draft Basis spec.)
322    
323    * Word8 should eventually be handled natively, too.
324    
325    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
326      as if there currently is no execution path that could trigger it in
327      SML/NJ.  (The assumptions underlying functions arith and promotable do not
328      hold for things like multiplication and division.)
329    
330    ----------------------------------------------------------------------
331    Name: Matthias Blume
332    Date: 2002/03/27 16:27:00 EST
333    Tag: blume-20020327-mlrisc-divisions
334    Description:
335    
336    Added support for all four division operations (ML's div, mod, quot,
337    and rem) to MLRISC.  In the course of doing so, I also rationalized
338    the naming (no more annoying switch-around of DIV and QUOT), by
339    parameterizing the operation by div_rounding_mode (which can be either
340    DIV_TO_ZERO or DIV_TO_NEGINF).
341    
342    The generic MLTreeGen functor takes care of compiling all four
343    operations down to only round-to-zero div.
344    
345    Missing pieces:
346    
347      * Doing something smarter than relying on MLTreeGen on architectures
348        like, e.g., the x86 where hardware division delivers both quotient and
349        remainder at the same time.  With this, the implementation of the
350        round-to-neginf operations could be further streamlined.
351    
352      * Remove inlining support for div/mod/rem from the frontend and replace it
353        with primops that get carried through to the backend.  Do this for all
354        int and word types.
355    
356    ----------------------------------------------------------------------
357    Name: Matthias Blume
358    Date: 2002/03/25 17:25:00 EST
359    Tag: blume-20020325-divmod
360    Description:
361    
362    I improved (hopefully without breaking them) the implementation of Int.div,
363    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
364    advantage of the following observations:
365    
366      Let  q = x quot y      r = x rem y
367           d = x div  y      m = x mod y
368    
369    where "quot" is the round-to-zero version of integer division that
370    hardware usually provides.  Then we have:
371    
372         r = x - q * y        where neither the * nor the - will overflow
373         d = if q >= 0 orelse x = q * y then q else q - 1
374                              where neither the * nor the - will overflow
375         m = if q >= 0 orelse r = 0 then r else r + y
376                              where the + will not overflow
377    
378    This results in substantial simplification of the generated code.
379    The following table shows the number of CFG nodes and edges generated
380    for
381            fun f (x, y) = x OPER y
382            (* with OPER \in div, mod, quot, rem *)
383    
384    
385        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
386        --------------------------------------------------------
387         div |         24 |         39 |         12 |         16
388         mod |         41 |         71 |         12 |         16
389        quot |          8 |         10 |          8 |         10
390         rem |         10 |         14 |          8 |         10
391    
392    
393    ----------------------------------------------------------------------
394    Name: Matthias Blume
395    Date: 2002/03/25 22:06:00 EST
396    Tag: blume-20020325-cprotobug
397    Description:
398    
399    Fixed a bug in cproto (c prototype decoder).
400    
401    ----------------------------------------------------------------------
402    Name: Matthias Blume
403    Date: 2002/03/25 16:00:00 EST
404    Tag: blume-20020325-raw-primops
405    Description:
406    
407    I did some cleanup to Allen's new primop code and
408    replaced yesterday's bootfiles with new ones.
409    (But they are stored in the same place.)
410    
411    ----------------------------------------------------------------------
412    Name: Matthias Blume
413    Date: 2002/03/24 22:40:00 EST
414    Tag: blume-20020324-bootfiles
415    Description:
416    
417    Made the bootfiles that Allen asked for.
418    
419    ----------------------------------------------------------------------
420    Name: Allen Leung
421    Date: 2002/03/23 15:50:00 EST
422    Tag: leunga-20020323-flint-cps-rcc-primops
423    Description:
424    
425      1. Changes to FLINT primops:
426    
427        (* make a call to a C-function;
428         * The primop carries C function prototype information and specifies
429         * which of its (ML-) arguments are floating point. C prototype
430         * information is for use by the backend, ML information is for
431         * use by the CPS converter. *)
432      | RAW_CCALL of { c_proto: CTypes.c_proto,
433                       ml_args: ccall_type list,
434                       ml_res_opt: ccall_type option,
435                       reentrant : bool
436                     } option
437       (* Allocate uninitialized storage on the heap.
438        * The record is meant to hold short-lived C objects, i.e., they
439        * are not ML pointers.  With the tag, the representation is
440        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
441        *)
442      | RAW_RECORD of {tag:bool,sz:int}
443      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
444    
445      2.  These CPS primops are now overloaded:
446    
447           rawload of {kind:numkind}
448           rawstore of {kind:numkind}
449    
450          The one argument form is:
451    
452             rawload {kind} address
453    
454          The two argument form is:
455    
456             rawload {kind} [ml object, byte-offset]
457    
458      3. RAW_CCALL/RCC now takes two extra arguments:
459    
460         a. The first is whether the C call is reentrant, i.e., whether
461            ML state should be saved and restored.
462         b. The second argument is a string argument specifying the name of
463            library and the C function.
464    
465         These things are currently not handled in the code generator, yet.
466    
467      4. In CProto,
468    
469         An encoding type of "bool" means "ml object" and is mapped into
470         C prototype of PTR.  Note that "bool" is different than "string",
471         even though "string" is also mapped into PTR, because "bool"
472         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
473    
474      5. Pickler/unpicker
475    
476         Changed to handle RAW_RECORD and newest RAW_CCALL
477    
478      6. MLRiscGen,
479    
480         1. Changed to handle the new rawload/rawstore/rawrecord operators.
481         2. Code for handling C Calls has been moved to a new module CPSCCalls,
482            in the file CodeGen/cpscompile/cps-c-calls.sml
483    
484      7. Added the conditional move operator
485    
486             condmove of branch
487    
488         to cps.  Generation of this is still buggy so it is currently
489         disabled.
490    
491    ----------------------------------------------------------------------
492    Name: Lal George
493    Date: 2002/03/22 14:18:25 EST
494    Tag: george-20020322-cps-branch-prob
495    Description:
496    
497    Implemented the Ball-Larus branch prediction-heuristics, and
498    incorporated graphical viewers for control flow graphs.
499    
500    Ball-Larus Heuristics:
501    ---------------------
502    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
503    
504    By design it uses the Dempster-Shafer theory for combining
505    probabilities.  For example, in the function:
506    
507        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
508    
509    the ball-larus heuristics predicts that the n=0 is unlikely
510    (OH-heuristic), and the 'then' branch is unlikely because of the
511    RH-heuristic -- giving the 'then' branch an even lower combined
512    probability using the Dempster-Shafer theory.
513    
514    Finally, John Reppy's loop analysis in MLRISC, further lowers the
515    probability of the 'then' branch because of the loop in the else
516    branch.
517    
518    
519    Graphical Viewing:
520    ------------------
521    I merely plugged in Allen's graphical viewers into the compiler. The
522    additional code is not much. At the top level, saying:
523    
524            Control.MLRISC.getFlag "cfg-graphical-view" := true;
525    
526    will display the graphical view of the control flow graph just before
527    back-patching.  daVinci must be in your path for this to work. If
528    daVinci is not available, then the default viewer can be changed
529    using:
530    
531            Control.MLRISC.getString "viewer"
532    
533    which can be set to "dot" or "vcg" for the corresponding viewers. Of
534    course, these viewers must be in your path.
535    
536    The above will display the compilation unit at the level of clusters,
537    many of which are small, boring, and un-interesting. Also setting:
538    
539            Control.MLRISC.getInt "cfg-graphical-view_size"
540    
541    will display clusters that are larger than the value set by the above.
542    
543    
544    ----------------------------------------------------------------------
545    Name: Matthias Blume
546    Date: 2002/03/21 22:20:00 EST
547    Tag: blume-20020321-kmp-bugfix
548    Description:
549    
550    Changed the interface to the KMP routine in PreString and fixed
551    a minor bug in one place where it was used.
552    
553    ----------------------------------------------------------------------
554    Name: Allen Leung
555    Date: 2002/03/21 20:30:00 EST
556    Tag: leunga-20020321-cfg
557    Description:
558    
559      Fixed a potential problem in cfg edge splitting.
560    
561    ----------------------------------------------------------------------
562    Name: Allen Leung
563    Date: 2002/03/21 17:15:00 EST
564    Tag: leunga-20020321-x86-fp-cfg
565    Description:
566    
567      1. Recoded the buggy parts of x86-fp.
568    
569         a. All the block reordering code has been removed.
570            We now depend on the block placement phases to do this work.
571    
572         b. Critical edge splitting code has been simplified and moved into the
573            CFG modules, as where they belong.
574    
575         Both of these were quite buggy and complex.  The code is now much, much
576         simpler.
577    
578      2. X86 backend.
579    
580         a. Added instructions for 64-bit support.  Instruction selection for
581            64-bit has not been committed, however, since that
582            requires changes to MLTREE which haven't been approved by
583            Lal and John.
584    
585         b. Added support for FUCOMI and FUCOMIP when generating code for
586            PentiumPro and above.  We only generate these instructions in
587            the fast-fp mode.
588    
589         c. Added cases for JP and JNP in X86FreqProps.
590    
591      3. CFG
592    
593         CFG now has a bunch of methods for edge splitting and merging.
594    
595      4. Machine description.
596    
597         John's simplification of MLTREE_BASIS.fcond broke a few machine
598         description things:
599    
600         rtl-build.{sig,sml} and hppa.mdl fixed.
601    
602         NOTE: the machine description stuff in the repository is still broken.
603               Again, I can't put my fixes in because that involves
604               changes to MLTREE.
605    
606    ----------------------------------------------------------------------
607    Name: Matthias Blume
608  Date: 2002/03/20 15:55:00 EST  Date: 2002/03/20 15:55:00 EST
609  Tag: blume-20020320-kmp  Tag: blume-20020320-kmp
610  Description:  Description:

Legend:
Removed from v.1155  
changed lines
  Added in v.1249

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