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/NOTES/HISTORY
ViewVC logotype

Diff of /sml/trunk/NOTES/HISTORY

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 583, Thu Mar 23 21:52:30 2000 UTC revision 686, Tue Jul 18 08:58:50 2000 UTC
# Line 12  Line 12 
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14  ----------------------------------------------------------------------  ----------------------------------------------------------------------
15    Name: Matthias Blume
16    Date: 2000/07/18 18:00:00 JST
17    Tag: blume-20000718-Version_110_29
18    Description:
19    
20    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
21    
22    2. Updated config/version to 110.29
23    
24    3. Updated config/srcarchiveurl
25    
26    3. New boot files!
27       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
28    
29    ----------------------------------------------------------------------
30    Name: Matthias Blume
31    Date: 2000/07/11 13:58:00 JST
32    Tag: blume-20000711-doctypo
33    Description:
34    
35    Fixed a few typos in CM manual.
36    
37    ----------------------------------------------------------------------
38    Name: Allen Leung
39    Date: 2000/06/15 00:38:00
40    Tag: leunga-20000704-sparc-x86
41    
42    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
43    2. fix to x86 RET bug [found by Dan Grossman]
44    3. sparc assembly bug fix for ticc instructions [found by Fermin]
45    
46       Affects c-- and moby only
47    
48    ----------------------------------------------------------------------
49    Name: Matthias Blume
50    Date: 2000/07/04 15:26:00
51    Tag: blume-20000704-trigger
52    Description:
53    
54    1. Improvements to CM manual.
55    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
56       of getting a back-trace.  The function, when called, raises an
57       internal exception which explicitly carries the full back-trace history,
58       so it is unaffected by any intervening handle-raise pairs ("trivial"
59       or not).  The interactive loop will print that history once it arrives
60       at top level.
61       Short of having all exceptions implicitly carry the full history, the
62       recommended way of using this facility is:
63         - compile your program with instrumentation "on"
64         - run it, when it raises an exception, look at the history
65         - if the history is "cut off" because of some handler, go and modify
66           your program so that it explicitly calls BTrace.trigger
67         - recompile (still instrumented), and rerun; look at the full history
68    
69    ----------------------------------------------------------------------
70    Name: Matthias Blume
71    Date: 2000/07/03 15:36:00 JST
72    Tag: blume-20000702-manual
73    Description:
74    
75    Small corrections and updates to CM manual.
76    
77    ----------------------------------------------------------------------
78    Name: Matthias Blume
79    Date: 2000/06/29 16:04:00 JST
80    Tag: blume-20000629-yacctool
81    Description:
82    
83    Changes:
84    
85    1. Class "mlyacc" now takes separate arguments to pass options to
86       generated .sml- and .sig-files independently.
87    2. Corresponding CM manual updates.
88    3. BTrace module now also reports call sites.  (However, for loop clusters
89       it only shows from where the cluster was entered.)  There are associated
90       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
91    
92    ----------------------------------------------------------------------
93    Name: Matthias Blume
94    Date: 2000/06/27 16:51:00 JST
95    Tag: blume-20000627-noweb
96    Description:
97    
98    Changes:
99    
100     1. Implemented "subdir" and "witness" options for noweb tool.
101        This caused some slight internal changes in CM's tool implementation.
102     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
103        some remaining issues from earlier path anchor changes.
104     3. Updated CM manual accordingly.
105    
106     4. Changed implementation of back-tracing so that I now consider it
107        ready for prime-time.
108    
109        In particular, you don't have to explicitly trigger the back-trace
110        anymore.  Instead, if you are running BTrace-instrumented code and
111        there is an uncaught exception (regardless of whether or not it was
112        raised in instrumented code), the top-level evalloop will print
113        the back-trace.
114    
115        Features:
116    
117          - Instrumented and uninstrumented code work together seemlessly.
118            (Of course, uninstrumented code is never mentioned in actual
119             back-traces.)
120    
121          - Asymptotic time- and space-complexity of instrumented code is
122            equal to that of uninstrumented code.  (This means that
123            tail-recursion is preserved by the instrumentation phase.)
124    
125          - Modules whose code has been instrumented in different sessions
126            work together without problem.
127    
128          - There is no penalty whatsoever on uninstrumented code.
129    
130          - There is no penalty on "raise" expressions, even in
131            instrumented code.
132    
133        A potential bug (or perhaps it is a feature, too):
134    
135          A back-trace reaches no further than the outermost instrumented
136          non-trivial "raise".  Here, a "trivial" raise is one that is the
137          sole RHS of a "handle" rule.  Thus, back-traces reach trough
138    
139               <exp> handle e => raise e
140    
141          and even
142    
143               <exp> handle Foo => raise Bar
144    
145          and, of course, through
146    
147               <exp> handle Foo => ...
148    
149         if the exception was not Foo.
150    
151         Back-traces always reach right through any un-instrumented code
152         including any of its "handle" expressions, trivial or not.
153    
154       To try this out, do the following:
155    
156         - Erase all existing binfiles for your program.
157           (You may keep binfiles for those modules where you think you
158            definitely don't need back-tracing.)
159         - Turn on back-trace instrumentation:
160              SMLofNJ.Internals.BTrace.mode (SOME true);
161         - Recompile your program.  (I.e., run "CM.make" or "use".)
162         - You may now turn instrumentation off again (if you want):
163              SMLofNJ.Internals.BTrace.mode (SOME false);
164         - Run your program as usual.  If it raises an exception that
165           reaches the interactive toplevel, then a back-trace will
166           automatically be printed.  After that, the toplevel loop
167           will print the exception history as usual.
168    
169    ----------------------------------------------------------------------
170    Name: Matthias Blume
171    Date: 2000/06/26 09:56:46 JST
172    Tag: blume-20000626-setup
173    Description:
174    
175    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
176          ML code before and after compiling a file (e.g., to set compiler
177          flags)
178    
179    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
180              - associated changes to btrace.sml (BTrace instrumentation pass)
181              - cleaner implementation of btimp.sml (BTrace tracing and report
182                module)
183    
184    CM manual: * new path encoding documented
185               * description of setup-parameter to "sml" added
186    
187    The biggest user-visible change to back-tracing is that it is no
188    longer necessary to compile all traced modules within the same
189    session.  (This was a real limitation.)
190    
191    ----------------------------------------------------------------------
192    Name: Matthias Blume
193    Date: 2000/06/24 12:40:00 JST
194    Tag: blume-20000624-startup
195    Description:
196    
197    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
198    bit too often -- to put it mildly. :)
199    
200    ----------------------------------------------------------------------
201    Name: Matthias Blume
202    Date: 2000/06/23 18:20:00 JST
203    Tag: blume-20000623-btrace
204    Description:
205    
206    This updates adds a backtrace facility to aid programmers in debugging
207    their programs.  This involves the following changes:
208    
209    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
210       keeping track of the current call stack.  When programs are compiled
211       in a special mode, the compiler will insert calls to these hooks
212       into the user program.
213       "Hook" means that it is possible for different implementations of
214       back-tracing to register themselves (at different times).
215    
216    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
217       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
218       It can be turned on using this call:
219         SMLofNJ.Internals.BTrace.mode (SOME true);
220       Turning it off again:
221         SMLofNJ.Internals.BTrace.mode (SOME false);
222       Querying the current status:
223         SMLofNJ.Internals.BTrace.mode NONE;
224       Annotated programs are about twice as big as normal ones, and they
225       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
226       where all hooks do nothing).  The slowdown with a plugin that is
227       actually useful (such as the one supplied by default) is even greater,
228       but in the case of the default plugin it is still only an constant
229       factor (amortized).
230    
231    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
232       with a sub-structure BTrace for controlling back-tracing.  In particular,
233       the above-mentioned function "mode" controls whether the annotation
234       phase is invoked by the compiler.  Another important function is
235       "trigger": when called it aborts the current execution and causes
236       the top-level loop to print a full back-trace.
237    
238    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
239       for back-tracing.  It keeps track of the dynamic call stack and in
240       addition to that it keeps a partial history at each "level" of that
241       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
242       d and b (at separate times, dynamically), then the report will show:
243    
244       GOTO   d
245             /c
246       GOTO  \b
247       CALL   a
248    
249       This shows that there was an initial non-tail call of a, then a
250       tail-call to b or c, looping behavior in a cluster of functions that
251       consist of b and c, and then a goto from that cluster (i.e., either from
252       b or from c) to d.
253    
254       Note that (depending on the user program) the amount of information
255       that the back-trace module has to keep track of at each level is bounded
256       by a constant.  Thus, the whole implementation has the same asymptotical
257       complexity as the original program (both in space and in time).
258    
259    5. compiler/TopLevel/interact/evalloop.sml has been modified to
260       handle the special exception SMLofNJ.Internals.BTrace.BTrace
261       which is raised by the "trigger" function mentioned above.
262    
263    Notes on usage:
264    
265    - Annotated code works well together with unannotated code:
266    Unannotated calls simply do not show up at all in the backtrace.
267    
268    - It is not a good idea to let modules that were annotated during
269    different sessions run at the same time.  This is because the compiler
270    chooses small integers to identify individual functions, and there
271    will be clashes if different modules were compiled in separate sessions.
272    (Nothing will crash, and you will even be told about the clashes, but
273    back-trace information will in general not be useful.)
274    
275    - Back-tracing can be confused by callcc and capture.
276    
277    - The only way of getting a back-trace right now is to explicitly
278    invoke the "trigger" function from your user program.  Eventually, we
279    should make every exception carry back-trace information (if
280    available).  But since this creates more overhead at "raise"-time
281    (similar to the current exnHistory overhead), I have not yet
282    implemented this.  (The implementation will be rather easy.)  With
283    exceptions carrying back-trace information, this facility will be even
284    more useful because users don't need to modify their programs...
285    
286    - While it is possible to compile the compiler with back-trace
287    annotations turned on (I did it to get some confidence in
288    correctness), you must make absolutely sure that core.sml and
289    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
290    be compiled with annotation because there is no core access yet, but
291    if you compile btimp.sml with annotation, then the system will go into
292    an infinite recursion and crash.)
293    Since CM currently does not know about BTrace, the only way to turn
294    annotations on and off for different modules of the compiler is to
295    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
296    this is awkward and clumsy.
297    
298    Sample sessions:
299    
300    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
301    - SMLofNJ.Internals.BTrace.mode (SOME true);
302    [autoloading]
303    [autoloading done]
304    val it = false : bool
305    - structure X = struct
306    -     fun main n = let
307    -         fun a (x, 0) = d x
308    -           | a (x, n) = b (x, n - 1)
309    -         and b (x, n) = c (x, n)
310    -         and c (x, n) = a (x, n)
311    -         and d x = e (x, 3)
312    -         and e (x, 0) = f x
313    -           | e (x, n) = e (x, n - 1)
314    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
315    -           | f n = n * g (n - 1)
316    -         and g n = a (n, 3)
317    -     in
318    -         f n
319    -     end
320    - end;
321    structure X : sig val main : int -> int end
322    - X.main 3;
323    *** BACK-TRACE ***
324    GOTO   stdIn:4.2-13.20: X.main[2].f
325    GOTO-( stdIn:4.2-13.20: X.main[2].e
326    GOTO   stdIn:4.2-13.20: X.main[2].d
327         / stdIn:4.2-13.20: X.main[2].a
328         | stdIn:4.2-13.20: X.main[2].b
329    GOTO-\ stdIn:4.2-13.20: X.main[2].c
330    CALL   stdIn:4.2-13.20: X.main[2].g
331    GOTO   stdIn:4.2-13.20: X.main[2].f
332    GOTO-( stdIn:4.2-13.20: X.main[2].e
333    GOTO   stdIn:4.2-13.20: X.main[2].d
334         / stdIn:4.2-13.20: X.main[2].a
335         | stdIn:4.2-13.20: X.main[2].b
336    GOTO-\ stdIn:4.2-13.20: X.main[2].c
337    CALL   stdIn:4.2-13.20: X.main[2].g
338    GOTO   stdIn:4.2-13.20: X.main[2].f
339    GOTO-( stdIn:4.2-13.20: X.main[2].e
340    GOTO   stdIn:4.2-13.20: X.main[2].d
341         / stdIn:4.2-13.20: X.main[2].a
342         | stdIn:4.2-13.20: X.main[2].b
343    GOTO-\ stdIn:4.2-13.20: X.main[2].c
344    CALL   stdIn:4.2-13.20: X.main[2].g
345    GOTO   stdIn:4.2-13.20: X.main[2].f
346    CALL   stdIn:2.15-17.4: X.main[2]
347    -
348    
349    (Note that because of a FLINt bug the above code currently does not
350    compile without BTrace turned on.)
351    
352    Here is another example, using my modified Tiger compiler:
353    
354    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
355    - SMLofNJ.Internals.BTrace.mode (SOME true);
356    [autoloading]
357    [autoloading done]
358    val it = false : bool
359    - CM.make "sources.cm";
360    [autoloading]
361    ...
362    [autoloading done]
363    [scanning sources.cm]
364    [parsing (sources.cm):parse.sml]
365    [creating directory CM/SKEL ...]
366    [parsing (sources.cm):tiger.lex.sml]
367    ...
368    [wrote CM/sparc-unix/semant.sml]
369    [compiling (sources.cm):main.sml]
370    [wrote CM/sparc-unix/main.sml]
371    [New bindings added.]
372    val it = true : bool
373    - Main.compile ("../testcases/merge.tig", "foo.out");
374    *** BACK-TRACE ***
375    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
376    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
377    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
378    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
379    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
380    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
381    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
382         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
383    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
384    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
385    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
386    CALL   main.sml:1.18-118.4: Main.compile[2]
387    -
388    
389    ----------------------------------------------------------------------
390    Name: Matthias Blumen
391    Date: 2000/06/21 18:00:00 JST
392    Tag: blume-20000621-manual
393    Description:
394    
395    CM manual update: Path environments documented.
396    
397    ----------------------------------------------------------------------
398    Name: Matthias Blume
399    Date: 2000/06/19 13:40:00
400    Tag: blume-20000619-manual
401    Description:
402    
403    CM manual and system/README update.  This only covers the fact that
404    there are no more implicit anchors.  (Path environments and the "bind"
405    option to "cm" have yet to be documented.)
406    
407    ----------------------------------------------------------------------
408    Name: Matthias Blume
409    Date: 2000/06/19 11:05:00 JST
410    Tag: blume-20000619-chdir-bugfix
411    Description:
412    
413    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
414    
415    ----------------------------------------------------------------------
416    Name: Matthias Blume
417    Date: 2000/06/18 22:00:10 JST
418    Tag: blume-20000618-implicit-anchors-really-gone
419    Description:
420    
421    I updates the previous HISTORY entry where I forgot to mention that
422    implicit anchors are no longer with us.
423    
424    The current update also gets rid of the (now useless) controller
425    CM.Control.implicit_anchors.
426    
427    ----------------------------------------------------------------------
428    Name: Matthias Blume
429    Date: 2000/06/16 17:30:00 JST
430    Tag: blume-20000616-anchorenv
431    Description:
432    
433    This patch implements the long anticipated (just kidding :) "anchor
434    environment" mechanism.  In the course of doing this, I also
435    re-implemented CM's internal "SrcPath" module from scratch.  The new
436    one should be more robust in certain boundary cases.  In any case, it
437    is a lot cleaner than its predecessor (IMHO).
438    
439    This time, although there is yet another boot file format change, I
440    kept the unpickler backward-compatible.  As a result, no new bootfiles
441    are necessary and bootstrapping is straightforward.  (You cannot read
442    new bootfiles into an old system, but the other way around is no
443    problem.)
444    
445    Visible changes:
446    
447    ** 0. Implicit path anchors (without the leading $-symbol) are no
448    longer recognized at all. This means that such path names are not
449    illegal either.  For example, the name basis.cm simply refers to a
450    local file called "basis.cm" (i.e, the name is an ordinary path
451    relative to .cm-files directory).  Or, to put it differently, only
452    names that start with $ are anchored paths.
453    
454    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
455    vanished.
456    
457    John (Reppy) had critizised this as soon as I originally proposed and
458    implemented it, but at that time I did not really deeply believe
459    him. :) Now I came full-circle because I need the $<singlearc> syntax
460    in another place where it cannot be seen as an abbreviation for
461    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
462    seems to mean (i.e., it "expands" into the corresponding anchor
463    value).
464    
465    However, when paths are used as members in CM description files, it
466    continues to be true that there must be at least another arc after the
467    anchor.  This is now enforced separately during semantic analysis
468    (i.e., from a lexical/syntactical point of view, the notation is ok.)
469    
470    ** 2. The "cm" class now accepts an option "bind".  The option's value
471    is a sub-option list of precisely two items -- one labeled "anchor"
472    and the other one labeled "value".  As you might expect, "anchor" is
473    used to specify an anchor name to be bound, and "value" specifies what
474    the anchor is being bound to.
475    
476    The value must be a directory name and can be given in either standard
477    syntax (including the possibility that it is itself an anchored path)
478    or native syntax.
479    
480    Examples:
481    
482       foo.cm (bind:(anchor:bar value:$mystuff/bar))
483       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
484    
485    and so on.
486    
487    The meaning of this is that the .cm-file will be processed with an
488    augmented anchor environment where the given anchor(s) is/are bound to
489    the given values(s).
490    
491    The rationale for having this feature is this: Suppose you are trying
492    to use two different (already stable) libraries a.cm and b.cm (that
493    you perhaps didn't write yourself).  Further, suppose each of these
494    two libraries internally uses its own auxiliary library $aux/lib.cm.
495    Normally you would now have a problem because the anchor "lib" can not
496    be bound to more than one value globally.  Therefore, the project that
497    uses both a.cm and b.cm must locally redirect the anchor to some other
498    place:
499    
500       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
501       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
502    
503    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
504    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
505    
506    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
507    you when you do so at the time of CM.stabilize).  Therefore, you can
508    also use an anchored path as the value:
509    
510      a.cm (bind:(anchor:lib value:$a-lib))
511      b.cm (bind:(anchor:lib value:$b-lib))
512    
513    Now you can globally configure (using the usual CM.Anchor.anchor or
514    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
515    itself is always locally bound, setting it globally is no longer
516    meaningful or necessary (but it does not hurt either).  In fact, "lib"
517    can still be used as a global anchor for separate purposes.  As a
518    matter of fact, one can locally define "lib" in terms of a global
519    "lib":
520    
521      a.cm (bind:(anchor:lib value:$lib/a))
522      b.cm (bind:(anchor:lib value:$lib/b))
523    
524    ** 3: The encoding of path names has changed.  This affects the way
525    path names are shown in CM's progress report and also the internal
526    protocol encoding used for parallel make.
527    
528    The encoding now uses one or more ':'-separated segments.  Each
529    segments corresponds to a file that has been specified relative to the
530    file given by its preceding segment.  The first segment is either
531    relative to the CWD, absolute, or anchored.  Each segment itself is
532    basically a Unix pathname; all segments but the first are relative.
533    
534    Example:
535    
536       $foo/bar/baz.cm:a/b/c.sml
537    
538    This path denotes the file bar/a/b/c.sml relative to the directory
539    denoted by anchor "foo".  Notice that the encoding also includes
540    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
541    paths are resolved relative to the .cm-files directory, so baz.cm must
542    be ignored to get the "real" pathname.
543    
544    To make this fact more obvious, CM puts the names of such "virtual
545    arcs" into parentheses when they appear in progress reports. (No
546    parentheses will appear in the internal protocol encoding.)  Thus,
547    what you really see is:
548    
549      $foo/bar/(baz.cm):a/b/c.sml
550    
551    I find this notation to be much more informative than before.
552    
553    Another new feature of the encoding is that special characters
554    including parentheses, colons, (back)slashes, and white space are
555    written as \ddd (where ddd is the decimal encoding of the character).
556    
557    *** The CM manual still needs to be updated.
558    
559    ----------------------------------------------------------------------
560    Name: Allen Leung
561    Date: 2000/06/15 00:38:00
562    Tag: leunga-20000615-x86-peephole
563    
564    x86 Peephole fix by Fermin.  Affects c-- and moby only.
565    
566    ----------------------------------------------------------------------
567    Name: Matthias Blume
568    Date: 2000/06/12 11:40:00
569    Tag: blume-20000612-parmakefix
570    Description:
571    
572    More cleanup after changing the file naming scheme: This time I
573    repaired the parallel make mechanism for CMB.make which I broke earlier.
574    
575    ----------------------------------------------------------------------
576    Name: Allen Leung
577    Date: 2000/06/09 01:25:00
578    Tag: leunga-20000609-various
579    
580    None of these things should affect normal SML/NJ operations
581    
582    1. Peephole improvements provided by Fermin (c--)
583    2. New annotation DEFUSE for adding extra dependence (moby)
584    3. New X86 LOCK instructions (moby)
585    4. New machine description language for reservation tables (scheduling)
586    5. Fixes to various optimization/analysis modules (branch chaining, dominator
587       trees etc.)
588    6. I've changed the CM files so that they can work with versions
589       110.0.6, 110.25 and 110.28
590    
591    ----------------------------------------------------------------------
592    Name: Matthias Blume
593    Date: 2000/06/09 12:40:00
594    Tag: blume-20000609-log
595    Description:
596    
597    - Removed all(?) remaining RCS Log entries from sources.
598    
599    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
600      anchored paths).
601    
602    ----------------------------------------------------------------------
603    Name: Matthias Blume
604    Date: 2000/06/07 17:00:00 JST
605    Tag: blume-20000607-no-implicit-anchors
606    Description:
607    
608    1. This update changes the default setting for
609    CM.Control.implicit_anchors from true to false.  This means that
610    implicit anchors are no longer permitted by default.  I also tried to
611    make sure that nothing else still relies on implicit anchors.
612    (This is the next step on the schedule towards a CM that does not even
613    have the notion of implicit anchors anymore.)
614    
615    2. More CM manual updates.
616    
617    3. I managed to track down and fix the pickling bug I mentioned last
618    time.  Because of the previously existing workaround, this entails no
619    immediate practical changes.
620    
621    ----------------------------------------------------------------------
622    Name: Matthias Blume
623    Date: 2000/06/06 11:15:00 JST
624    Tag: blume-20000606-lazierpickle
625    Description:
626    
627    !!!! NEW BOOT FILES !!!!
628    
629    * The main purpose of this update is to make library pickles lazier in
630    order to reduce the initial space penalty for autoloading a library.
631    As a result, it is now possible to have $smlnj/compiler.cm
632    pre-registered.  This should take care of the many complaints or
633    inquiries about missing structure Compiler.  This required changes to
634    CM's internal data structures and small tweaks to some algorithms.
635    
636    As a neat additional effect, it is no longer necessary (for the sake
637    of lean heap image files) to distinguish between a "minimal" CM and a
638    "full" CM.  Now, there is only one CM (i.e., the "full" version:
639    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
640    interactive top level. ($smlnj/cm/minimal.cm is gone.)
641    
642    To make the life of compiler-hackers easier, "makeml" now also
643    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
644    words, after you bootstrap a new sml for the first time, you will not
645    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
646    around you will still have to do it, though.)
647    
648    * A second change consists of major updates to the CM manual.  There
649    are now several appendices with summary information and also a full
650    specification of the CM description file syntax.
651    
652    * In directory src/system I added the script "allcross".  This script
653    invokes sml and cross-compiles the compiler for all supported
654    architectures.  (Useful when providing a new set of boot files.)
655    
656    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
657    added a small tweak to pickle-util.sml to work around this problem,
658    but it is not a proper fix yet.  I will investigate further.  (The
659    effect of the bug was an inflation of library pickle size.)
660    
661    * Version number increased to 110.28.1 (to avoid compatibility problems).
662    
663    ----------------------------------------------------------------------
664    Name: Allen Leung
665    Date: 2000/05/25 17:28 EDT
666    Tag: leunga-20000525-ra
667    Description:
668    
669      Fixed a bug in freezing phase of the register allocator.
670    
671    ----------------------------------------------------------------------
672    Name: Allen Leung
673    Date: 2000/05/15 22:53 EDT
674    Tag: leunga-20000515-alpha-x86-ra
675    Description:
676    
677      1. Alpha
678    
679          Slight cleanup.  Removed the instruction SGNXL
680    
681      2. X86
682    
683          Added the following instructions to the instruction set:
684    
685            ROLx, RORx,
686            BTx, BTSx, BTLx, BTRx,
687            XCHGx, and variants with the LOCK prefix
688    
689      3. Register Allocation
690    
691          The module ra-rewrite-with-renaming has been improved.
692    
693      These have no effect on SML/NJ.
694    
695    ----------------------------------------------------------------------
696    Name: Matthias Blume
697    Date: 2000/05/15 16:20:00 JST
698    Tag: blume-20000515-lightrebuild
699    Description:
700    
701    1. I added an alternative to "-rebuild" to "makeml".  The difference is
702       that prior to calling CMB.make' the CM-variable "LIGHT" will be
703       defined.  In effect, the command will not build any cross-compiler
704       backends and therefore finish more quickly.
705    
706       The "fixpt" script also takes a "-light" switch to be able to use
707       this new facility while compiling for a fixpoint.
708    
709    2. I replaced all mentions of anchored paths in group owner specifications
710       with simple relative paths (usually starting with "..").
711       The rationale is that a library's internal workings should not be
712       compromised by the lack of some anchor.  (An anchor is necessary
713       for someone who wants to refer to the library by an anchored path,
714       but it should not be necessary to build the same library in the first
715       place.)
716    
717    3. I changed the way CM's tool mechanism determines the shell command
718       string used for things like ml-yacc etc. so that it does not break
719       when CM.Control.implicit_anchors is turned off.
720    
721    ----------------------------------------------------------------------
722    Name: Matthias Blume
723    Date: 2000/05/12 18:20:00 JST
724    Tag: blume-20000512-ml-build
725    Description:
726    
727    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
728    getting installed properly (by config/install.sh).
729    
730    ----------------------------------------------------------------------
731    Name: Matthias Blume
732    Date: 2000/05/12 17:30:00 JST
733    Tag: blume-20000512-anchors
734    Description:
735    
736    !!! NEW BOOT FILES !!!
737    
738    This change is in preparation of fading out support for "implicitly
739    anchored path names".  I went through all sources and used the
740    explicit (and relatively new) $-notation.  See system/README and the
741    CM manual for more info on this.
742    
743    I also modified the anchoring scheme for some things such as "smlnj",
744    "MLRISC", "cm", etc. to take advantage of the fact that explicit
745    anchors are more expressive: anchor name and first arc do not have to
746    coincide.  This entails the following user-visible change:
747    
748    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
749    particular, when you fire up sml with a command-line argument, say,
750    e.g.:
751    
752       sml '$smlnj/cmb.cm'
753    
754    At the ML toplevel prompt:
755    
756       CM.autoload "$smlnj/cmb.cm";
757    
758    There is also a new controller in CM.Control that can be used to turn
759    off all remaining support for implicit anchors by saying:
760    
761        CM.autoload "$smlnj/
762        #set CM.Control.implicit_anchors false;
763    
764    This causes CM to reject implicitly anchored paths.  This is (for the
765    time being) less permissive than the "final" version where there will
766    be no more such implicit anchors and relative paths will be just that:
767    relative.
768    
769    The next step (version after next version?) will be to make the
770    default for CM.Control.implicit_anchors false.  After the dust has
771    settled, I can then produce the "final" version of this...
772    
773    Note: Since bootstrapping is a bit tricky, I provided new boot files.
774    
775    ----------------------------------------------------------------------
776    Name: Matthias Blume
777    Date: 2000/05/11 16:30:00 JST
778    Tag: blume-20000511-sources
779    Description:
780    
781    The main change is that I added function CM.sources as a generalized
782    version of the earlier CM.makedepend.  This entails the following
783    additional changes:
784    
785      - CM.makedepend has been dropped.
786    
787      - CM manual has been updated.
788    
789      - TOOLS signature and API have been changed.
790    
791    ----------------------------------------------------------------------
792    Name: Allen Leung
793    Date: 2000/05/10 21:17 EDT
794    Tag: leunga-20000510-moby-c--ssa
795    Description:
796    
797      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
798    None of these affect SML/NJ.
799    
800    1. Register Allocation
801    
802        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
803           This module tries to remove local (i.e. basic block level) redundancies
804           during spilling.
805    
806        b. A new framework for performing region based register allocation.
807           Not yet entirely functional.
808    
809    2. X86
810    
811       a. DefUse for POP was missing the stack pointer [found by Lal]
812       b. Reload for CALL was incorrect in X86Spill [found by John]
813       c. Various fixes in X86Spill so that it can be used correctly for
814          the new spilling module.
815    
816    3. SSA/IR
817    
818       a. New module ir/dj-dataflow.sml implements elimination based
819          data flow analysis.
820    
821    4. MLRiscGen
822    
823       a. Fix for gc type annotation
824    
825    5. MDGen
826    
827       Various fixes for machine description -> ml code translation.  For ssa
828       only.
829    
830    ----------------------------------------------------------------------
831    Name: Allen Leung
832    Date: 2000/05/08 22:17 EDT
833    Tag: leunga-20000508-labexp
834    Description:
835    
836      Fermin has found a few assembly problems with constant expressions
837      generated in LabelExp.  Mostly, the problems involve extra parentheses,
838      which choke on dumb assemblers.  This is his fix.
839    
840    ----------------------------------------------------------------------
841    Name: Dave MacQueen
842    Date: 2000/04/09 14:00 EDT
843    Tag: dbm-20000502-Version_110_28
844    Description:
845    
846    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
847    
848    2. Updated config/version to 110.28
849    
850    3. Updated config/srcarchiveurl
851    
852    3. New boot files!
853       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
854    
855    ----------------------------------------------------------------------
856    Name: Matthias Blume
857    Date: 2000/05/01 19:05:00 JST
858    Tag: blume-20000501-noweb
859    Description:
860    
861    A new noweb tool has been added.  The existing system is entirely
862    unaffected by this, but some CM users have asked for renewed noweb
863    support.  Everything is documented in the CM manual.
864    
865    New (plugin) libraries:
866    
867       noweb-tool.cm
868       nw-ext.cm
869    
870    ----------------------------------------------------------------------
871    Name: Dave MacQueen
872    Date: 2000/04/30 12:40PM EDT
873    Tag: dbm-20000430-bug_fixes
874    Description:
875    
876    1. Fix for bug 1498
877       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
878       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
879         added toRealArray function
880       smlnj/src/compiler/MiscUtil/print/ppobj.sml
881         added check for tag Obj.RealArray to array printing case in ppObj
882    
883    2. Fix for bug 1510
884       smlnj/src/compiler/Semant/types/typesutil.sml
885         fixed definition of dummyargs (used by equalTycon) so that
886         dummy args are distinct types
887    
888    ----------------------------------------------------------------------
889    Name: Matthias Blume
890    Date: 2000/04/30 01:00:00 JST
891    Tag: blume-20000430-versions
892    Description:
893    
894    1. CM version numbering added.  This is an implementation of Lal's
895       proposal for adding version numbers and version checking to .cm
896       files.  Lal said that his proposal was just that -- a proposal.
897       For the time being I went ahead and implemented it so that people
898       can comment on it.  Everything is completely backward-compatible
899       (except for the stable library format, i.e., new bootfiles!).
900    
901       As usual, see the CM manual for details.
902    
903    2. An alternative syntax for anchored paths has been implemented.
904       Dave has recently voiced the same concerns that I had when I did
905       this, so there should be some support.  My take is that eventually
906       I will let support for the current syntax (where anchors are
907       "implicit") fade out in favor of the new, explicit syntax.
908       In order to be backward-compatible, both old and new syntax are
909       currently supported.
910    
911       Again, see the CM manual for details.
912    
913    3. Parallel make is trying to be slightly smarter:  When the master
914       process finds a "bottleneck", i.e., when there is only one
915       compilation unit that can be compiled and everybody else is
916       waiting on it, then it will simply compile it directly instead
917       of clumsily telling one of the slaves to do it.
918    
919    4. Support for "unsharing" added.  This is necessary in order to be
920       able to have two different versions of the same library running
921       at the same time (e.g., for trying out a new MLRISC while still
922       having the old MLRISC linked into the current compiler, etc.)
923       See the CM manual.
924    
925    5. Simple "makedepend" functionality added for generating Makefile
926       dependency information.  (This is rather crude at the moment.
927       Expect some changes here in the future.)
928    
929    6. ".fun" added as a recognized suffix for ML files. Also documented
930       explicitly in the manual that the fallback behavior (unknown suffix
931       -> ML file) is not an official feature!
932    
933    7. Small changes to the pickler for stable libraries.
934    
935    8. Several internal changes to CM (for cleanup/improvement).
936    
937    
938    !!!! NEW BINFILES !!!!
939    
940    ----------------------------------------------------------------------
941    Name: Matthias Blume
942    Date: 2000/04/28 17:30:00 JST
943    Tag: blume-20000428-pathconfig
944    Description:
945    
946    1. I changed config/install.sh to remove duplicate entries from the
947       lib/pathconfig file at the end.  Moreover, the final version of
948       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
949       in src/system/installml.
950    
951    2. The config/install.sh script now consistently uses relative
952       pathnames in lib/pathconfig whenever the anchor is in the lib
953       directory.  (So far this was true for the libraries that come
954       pre-compiled and bundled as part of the bootfiles but not for
955       libraries that are compiled by the script itself.)
956    
957    ----------------------------------------------------------------------
958    Name: Matthias Blume
959    Date: 2000/04/26 13:10:00 JST
960    Tag: blume-20000426-fun_suffix
961    Description:
962    
963    Added ".fun" as a recognized file name suffix (for ML code).
964    
965    ----------------------------------------------------------------------
966    Name: Allen Leung
967    Date: 2000/04/25 17:00:00 EST
968    Tag: leunga-20000425-alpha-ra
969    Description:
970    
971    1. Alpha
972    
973        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
974    in C--.
975    
976    2. RA
977    
978       Added a flag "ra-dump-size" to print out the size of the flowgraph
979       and the interference graph.
980    
981    ----------------------------------------------------------------------
982    Name: Dave MacQueen
983    Date: 2000/04/25/
984    Tag: dbm-20000425-mlyacc_doc_examples
985    Description:
986      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
987      Updated all three examples in src/ml-yacc/examples to run
988      under 110.* using CM.make.
989    
990    ----------------------------------------------------------------------
991    Name: Allen Leung
992    Date: 2000/04/20 23:04:00 EST
993    Tag: leunga-20000420-ssa-c---stuff
994    Description:
995    
996      This update synchronizes my repository with Yale's.  Most of these
997    changes, however, do not affect SML/NJ at all (the RA is an exception).
998    
999    1. Register Allocator
1000    
1001       a. An improvement in the interference graph construction:
1002          Given a copy
1003    
1004                s <- t
1005    
1006          no interference edge between s and t is added for this definition of s.
1007    
1008       b. I've added two new spill heuristic modules that Fermin and I developed
1009          (in the new library RA.cm). These are unused in SML/NJ but maybe
1010          useful for others (Moby?)
1011    
1012    2. X86
1013    
1014       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1015    
1016    3. Alpha
1017    
1018       a. Added the BSR instruction and code generation that goes with it [C--]
1019       b. Other fixes too numerous to recount provided by Fermin [C--]
1020    
1021    4. Regmaps
1022    
1023       a. The regmaps are not initialized with the identity physical bindings
1024          at creation time.  This is unneeded.
1025    
1026    5. MLRISC Optimizations
1027    
1028       a. The DJ-Graph module can now compute the iterated dominance frontiers
1029          intersects with liveness incrementally in linear time! Woohoo!
1030          This is now used in my new SSA construction algorithm.
1031    
1032       b. THe branch reorganization module is now smarter about linear chains of
1033          basic blocks.
1034    
1035    
1036    ----------------------------------------------------------------------
1037    Name: Matthias Blume
1038    Date: 2000/04/12 13:52:00 JST
1039    Tag: blume_main_v110p27_1
1040    Description:
1041    
1042    Changed install.sh script to handle archive files without version number
1043    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1044    name of the boot file archive.
1045    
1046    ----------------------------------------------------------------------
1047    Name: Dave MacQueen
1048    Date: 2000/04/09 14:00 EDT
1049    Tag: dbm-20000410-Version_110_27
1050    Description:
1051    
1052    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1053    
1054    2. Updated src/config/version to 110.27
1055    
1056    3. New boot files!
1057    
1058    ----------------------------------------------------------------------
1059    Name: Allen Leung
1060    Date: 2000/04/09 19:09:00 EST
1061    Tag: leunga-20000409-misc
1062    Description:
1063    
1064    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1065    
1066    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1067    
1068    ----------------------------------------------------------------------
1069    Name: Stefan
1070    Date: 2000/04/07 10:00:00 EDT
1071    Tag: monnier-20000406-branch-handling
1072    Description:
1073    
1074    Improved handling of branches (mostly those generated from
1075    polymorphic equality), removed switchoff and changed the
1076    default optimization settings (more cpsopt and less flintopt).
1077    
1078    ----------------------------------------------------------------------
1079    Name: Allen Leung
1080    Date: 2000/04/06 01:30:00 EST
1081    Tag: leunga-20000406-peephole-x86-SSA-2
1082    Description:
1083    
1084       Forgot a few files.
1085    
1086    ----------------------------------------------------------------------
1087    Name: Allen Leung
1088    Date: 2000/04/06 00:36:00 EST
1089    Tag: leunga-20000406-peephole-x86-SSA
1090    Description:
1091    
1092    1.  New Peephole code
1093    
1094    2.  Minor improvement to X86 instruction selection
1095    
1096    3.  Various fixes to SSA and machine description -> code translator
1097    
1098    ----------------------------------------------------------------------
1099    Name: Matthias Blume
1100    Date: 2000/04/05 12:30:00 JST
1101    Tag: blume_main_v110p26p2_3
1102    Description:
1103    
1104    This update just merges three minor cosmetic updates to CM's sources
1105    to get ready for the 110.27 code freeze on Friday.  No functionality
1106    has changed.
1107    
1108    ----------------------------------------------------------------------
1109    Name: Allen Leung
1110    Date: 2000/04/04 19:39:00 EST
1111    Tag: leunga-20000404-x86-asm
1112    Description:
1113    
1114    1.  Fixed a problem in X86 assembly.
1115    
1116        Things like
1117    
1118           jmp %eax
1119           jmp (%eax)
1120    
1121        should be output as
1122    
1123           jmp *%eax
1124           jmp *(%eax)
1125    
1126    2.  Assembly output
1127    
1128          Added a new flag
1129    
1130              "asm-indent-copies" (default to false)
1131    
1132          When this flag is on, parallel copies will be indented an extra level.
1133    
1134    ----------------------------------------------------------------------
1135    Name: Allen Leung
1136    Date: 2000/04/04 03:18:00 EST
1137    Tag: leunga-20000404-C--Moby
1138    Description:
1139    
1140        All of these fixes are related to C--, Moby, and my own optimization
1141        stuff; so they shouldn't affect SML/NJ.
1142    
1143    1.  X86
1144    
1145        Various fixes related floating point, and extensions.
1146    
1147    2.  Alpha
1148    
1149        Some extra patterns related to loads with signed/zero extension
1150        provided by Fermin.
1151    
1152    3.  Assembly
1153    
1154        When generating assembly, resolve the value of client defined constants,
1155        instead of generating symbolic values.  This is controlled by the
1156        new flag "asm-resolve-constants", which is default to true.
1157    
1158    4.  Machine Descriptions
1159    
1160        a. The precedence parser was slightly broken when parsing infixr symbols.
1161        b. The type generalizing code had the bound variables reversed, resulting
1162           in a problem during arity raising.
1163        c. Various fixes in machine descriptions.
1164    
1165    ----------------------------------------------------------------------
1166    Name: Matthias Blume
1167    Date: 2000/04/03 16:05:00 JST
1168    Tag: blume_main_v110p26p2_2
1169    Description:
1170    
1171    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
1172    now done via the ordinary static environment that is context to each
1173    compilation unit.
1174    
1175    To this end, I arranged that instead of "structure Core" as "structure
1176    _Core" is bound in the pervasive environment.  Core access is done via
1177    _Core (which can never be accidentally rebound because _Core is not a
1178    legal surface-syntax symbol).
1179    
1180    The current solution is much cleaner because the core environment is
1181    now simply part of the pervasive environment which is part of every
1182    compilation unit's context anyway.  In particular, this eliminates all
1183    special-case handling that was necessary until now in order to deal
1184    with dynamic and symbolic parts of the core environment.
1185    
1186    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1187    compilation manager's bootstrap compiler (actually: in the "init group"
1188    handling).  See the comments in src/system/smlnj/init/init.cmi for
1189    more details.
1190    
1191    I also tried to track down all mentions of "Core" (as string argument
1192    to Symbol.strSymbol) in the compiler and replaced them with a
1193    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
1194    name appears in one place only.
1195    
1196    Binfile and bootfile format have not changed, but the switchover from
1197    the old "init.cmi" to the new one is a bit tricky, so I supplied new
1198    bootfiles anyway.
1199    
1200    ----------------------------------------------------------------------
1201    Name: Allen Leung
1202    Date: 2000/04/02 21:17:00 EST
1203    Tag: leunga-20000402-mltree
1204    Description:
1205    
1206       1. Renamed the constructor CALL in MLTREE by popular demand.
1207       2. Added a bunch of files from my repository.  These are currently
1208          used by other non-SMLNJ backends.
1209    
1210    ----------------------------------------------------------------------
1211    Name: Allen Leung
1212    Date: 2000/03/31 21:15:00 EST
1213    Tag: leunga-20000331-aliasing
1214    Description:
1215    
1216    This update contains a rewritten (and hopefully more correct) module
1217    for extracting aliasing information from CPS.
1218    
1219       To turn on this feature:
1220    
1221            Compiler.Control.CG.memDisambiguate := true
1222    
1223       To pretty print the region information with assembly
1224    
1225           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
1226    
1227       To control how many levels of aliasing information are printed, use:
1228    
1229           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
1230    
1231       The default of n is 3.
1232    
1233    ----------------------------------------------------------------------
1234    Name: David MacQueen
1235    Date: 2000/03/31 11:15:00 EST
1236    Tag: dbm-20000331-runtime_fix
1237    Description:
1238    
1239    This update contains:
1240    
1241    1. runtime/c-lib/c-libraries.c
1242       includes added in revision 1.2 caused compilation errors on hppa-hpux
1243    
1244    2. fix for bug 1556
1245       system/Basis/Implementation/NJ/internal-signals.sml
1246    
1247    ----------------------------------------------------------------------
1248    Name: Matthias Blume
1249    Date: 2000/03/31 18:00:00 JST
1250    Tag: blume_main_v110p26p2_1
1251    Description:
1252    
1253    This update contains:
1254    
1255    1. A small change to CM's handling of stable libraries:
1256       CM now maintains one "global" modmap that is used for all stable
1257       libraries.  The use of such a global modmap maximizes sharing and
1258       minimizes the need for re-traversing parts of environments during
1259       modmap construction.  (However, this has minor impact since modmap
1260       construction seems to account for just one percent or less of total
1261       compile time.)
1262    
1263    2. I added a "genmap" phase to the statistics.  This is where I got the
1264       "one percent" number (see above).
1265    
1266    3. CM's new tool parameter mechanism just became _even_ better. :)
1267       - The parser understands named parameters and recursive options.
1268       - The "make" and "shell" tools use these new features.
1269         (This makes it a lot easier to cascade these tools.)
1270       - There is a small syntax change: named parameters use a
1271    
1272           <name> : ( <option> ... )            or
1273           <name> : <string>
1274    
1275         syntax.  Previously, named parameters were implemented in an
1276         ad-hoc fashion by each tool individually (by parsing strings)
1277         and had the form
1278    
1279           <name>=<string>
1280    
1281       See the CM manual for a full description of these issues.
1282    
1283    ----------------------------------------------------------------------
1284    Name: Matthias Blume
1285    Date: 2000/03/30 18:00:00 JST
1286    Tag: blume_main_v110p26p2_0
1287    Description:
1288    
1289    !!!!! WARNING !!!!!!
1290    !!  New binfiles  !!
1291    !!!!!!!!!!!!!!!!!!!!
1292    
1293    This update contains:
1294    
1295    1. Moderate changes to CM:
1296    
1297       - Changes to CM's tools mechanism.  In particular, it is now possible
1298       to have tools that accept additional "command line" parameters
1299       (specified in the .cm file at each instance where the tool's class is
1300       used).
1301    
1302       This was done to accommodate the new "make" and "shell" tools which
1303       facilitate fairly seamless hookup to portions of code managed using
1304       Makefiles or Shell scripts.
1305    
1306       There are no classes "shared" or "private" anymore.  Instead, the
1307       sharing annotation is now a parameter to the "sml" class.
1308    
1309       There is a bit of generic machinery for implementing one's own
1310       tools that accept command-line parameters.  However, I am not yet fully
1311       satisfied with that part, so expect changes here in the future.
1312    
1313       All existing tools are described in the CM manual.
1314    
1315       - Slightly better error handling.  (CM now suppresses many followup
1316       error messages that tended to be more annoying than helpful.)
1317    
1318    2. Major changes to the compiler's static environment data structures.
1319    
1320       - no CMStaticEnv anymore.
1321            - no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
1322              but it is called Environment), no conversions between different
1323              kinds of static environments
1324    
1325       - There is still a notion of a "modmap", but such modmaps are generated
1326         on demand at the time when they are needed.  This sounds slow, but I
1327         sped up the code that generates modmaps enough for this not to lead to
1328         a slowdown of the compiler (at least I didn't detect any).
1329    
1330       - To facilitate rapid modmap generation, static environments now
1331         contain an (optional) "modtree" structure.  Modtree annotations are
1332         constructed by the unpickler during unpickling.  (This means that
1333         the elaborator does not have to worry about modtrees at all.)
1334         Modtrees have the advantage that they are compositional in the same
1335         way as the environment data structure itself is compositional.
1336         As a result, modtrees never hang on to parts of an environment that
1337         has already been rendered "stale" by filtering or rebinding.
1338    
1339       - I went through many, many trials and errors before arriving at the
1340         current solution.  (The initial idea of "linkpaths" did not work.)
1341         But the result of all this is that I have touched a lot of files that
1342         depend on the "modules" and "types" data structures (most of the
1343         elaborator). There were a lot of changes during my "linkpath" trials
1344         that could have been reverted to their original state but weren't.
1345         Please, don't be too harsh on me for messing with this code a bit more
1346         than what was strictly necessary...  (I _did_ resist the tempation
1347         of doing any "global reformatting" to avoid an untimely death at
1348         Dave's hands. :)
1349    
1350       - One positive aspect of the previous point:  At least I made sure that
1351         all files that I touched now compile without warnings (other than
1352         "polyEqual").
1353    
1354       - compiler now tends to run "leaner" (i.e., ties up less memory in
1355         redundant modmaps)
1356    
1357    ----------------------------------------------------------------------
1358    Name: Allen Leung
1359    Date: 2000/03/29 18:00:00
1360    Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
1361    Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
1362    Description:
1363    
1364       This update contains *MAJOR* changes to the way code is generated from CPS
1365    in the module mlriscGen, and in various backend modules.
1366    
1367    CHANGES
1368    =======
1369    
1370    1. MLRiscGen: forward propagation fix.
1371    
1372       There was a bug in forward propagation introduced at about the same time
1373       as the MLRISC x86 backend, which prohibits coalescing to be
1374       performed effectively in loops.
1375    
1376       Effect: speed up of loops in RISC architectures.
1377               By itself, this actually slowed down certain benchmarks on the x86.
1378    
1379    2. MLRiscGen:  forward propagating addresses from consing.
1380    
1381       I've changed the way consing code is generated.  Basically I separated
1382       out the initialization part:
1383    
1384            store tag,   offset(allocptr)
1385            store elem1, offset+4(allocptr)
1386            store elem2, offset+8(allocptr)
1387            ...
1388            store elemn, offset+4n(allocptr)
1389    
1390       and the address computation part:
1391    
1392            celladdr <- offset+4+alloctpr
1393    
1394       and move the address computation part
1395    
1396       Effect:  register pressure is generally lower as a result.  This
1397                makes compilation of certain expressions much faster, such as
1398                long lists with non-trivial elements.
1399    
1400                 [(0,0), (0,0), .... (0,0)]
1401    
1402    3. MLRiscGen: base pointer elimination.
1403    
1404        As part of the linkage mechanism, we generate the sequence:
1405    
1406         L:  ...  <- start of the code fragment
1407    
1408         L1:
1409             base pointer <- linkreg - L1 + L
1410    
1411         The base pointer was then used for computing relocatable addresses
1412       in the code fragment.  Frequently (such as in lots of continuations)
1413       this is not needed.  We now eliminate this sequence whenever possible.
1414    
1415         For compile time efficiency, I'm using a very stupid local heuristic.
1416       But in general, this should be done as a control flow analysis.
1417    
1418       Effect:  Smaller code size.  Speed up of most programs.
1419    
1420    4. Hppa back end
1421    
1422         Long jumps in span dependence resolution used to depend on the existence
1423      of the base pointer.
1424    
1425         A jump to a long label L was expanded into the following sequence:
1426    
1427          LDIL %hi(L-8192), %r29
1428          LDO  %lo(L-8192)(%r29), %r29
1429          ADD  %r29, baseptr, %r29
1430          BV,n %r0(%r29)
1431    
1432         In the presence of change (3) above, this will not work.  I've changed
1433       it so that the following sequence of instructions are generated, which
1434       doesn't mention the base pointer at all:
1435    
1436             BL,n  L', %r29           /* branch and link, L' + 4 -> %r29 */
1437        L':  ADDIL L-(L'+4), %r29     /* Compute address of L */
1438             BV,n  %r0(%r29)          /* Jump */
1439    
1440    5. Alpha back end
1441    
1442          New alpha instructions LDB/LDW have been added, as per Fermin's
1443       suggestions.   This is unrelated to all other changes.
1444    
1445    6. X86 back end
1446    
1447         I've changed andl to testl in the floating point test sequence
1448         whenever appropriate.  The Intel optimization guide states that
1449         testl is preferable to andl.
1450    
1451    7. RA (x86 only)
1452    
1453         I've improved the spill propagation algorithm, using an approximation
1454       of maximal weighted independent sets.   This seems to be necessary to
1455       alleviate the negative effect in light of the slow down in (1).
1456    
1457         I'll write down the algorithm one of these days.
1458    
1459    8. MLRiscGen: frequencies
1460    
1461         I've added an annotation that states that all call gc blocks have zero
1462       execution frequencies.  This improves register allocation on the x86.
1463    
1464    BENCHMARKS
1465    ==========
1466    
1467       I've only perform the comparison on 110.25.
1468    
1469       The platforms are:
1470    
1471        HPPA  A four processor HP machine (E9000) with 5G of memory.
1472        X86   A 300Hhz Pentium II with 128M of memory, and
1473        SPARC An Ultra sparc 2 with 512M of memory.
1474    
1475       I used the following parameters for the SML benchmarks:
1476    
1477                 @SMLalloc
1478         HPPA    256k
1479         SPARC   512k
1480         X86     256k
1481    
1482    COMPILATION TIME
1483    ----------------
1484       Here are the numbers comparing the compilation times of the compilers.
1485       I've only compared 110.25 compiling the new sources versus
1486       a fixpoint version of the new compiler compiling the same.
1487    
1488                     110.25                                  New
1489               Total  Time in RA  Spill+Reload   Total  Time In RA Spill+Reload
1490         HPPA   627s    116s        2684+3584     599s    95s       1003+1879
1491         SPARC  892s    173s        2891+3870     708s    116s      1004+1880
1492         X86    999s    315s       94006+130691   987s    296s    108877+141957
1493    
1494                   110.25         New
1495                Code Size      Code Size
1496         HPPA   8596736         8561421
1497         SPARC  8974299         8785143
1498         X86    9029180         8716783
1499    
1500       So in summary, things are at least as good as before.   Dramatic
1501       reduction in compilation is obtained on the Sparc; I can't explain it,
1502       but it is reproducible.  Perhaps someone should try to reproduce this
1503       on their own machines.
1504    
1505    SML BENCHMARKS
1506    --------------
1507    
1508        On the average, all benchmarks perform at least as well as before.
1509    
1510          HPPA         Compilation Time     Spill+Reload      Run Time
1511                     110.25  New            110.25    New   110.25  New
1512    
1513          barnesHut  3.158  3.015  4.75%    1+1       0+0   2.980  2.922   2.00%
1514              boyer  6.152  5.708  7.77%    0+0       0+0   0.218  0.213   2.34%
1515       count-graphs  1.168  1.120  4.32%    0+0       0+0  22.705 23.073  -1.60%
1516                fft  0.877  0.792 10.74%    1+3       1+3   0.602  0.587   2.56%
1517        knuthBendix  3.180  2.857 11.32%    0+0       0+0   0.675  0.662   2.02%
1518             lexgen  6.190  5.290 17.01%    0+0       0+0   0.913  0.788  15.86%
1519               life  0.803  0.703 14.22%   25+25      0+0   0.153  0.140   9.52%
1520              logic  2.048  2.007  2.08%    6+6       1+1   4.133  4.008   3.12%
1521         mandelbrot  0.077  0.080 -4.17%    0+0       0+0   0.765  0.712   7.49%
1522             mlyacc 22.932 20.937  9.53%  154+181    32+57  0.468  0.430   8.91%
1523            nucleic  5.183  5.060  2.44%    2+2       0+0   0.125  0.120   4.17%
1524      ratio-regions  3.357  3.142  6.84%    0+0       0+0  116.225 113.173 2.70%
1525                ray  1.283  1.290 -0.52%    0+0       0+0   2.887  2.855   1.11%
1526             simple  6.307  6.032  4.56%   28+30      5+7   3.705  3.658   1.28%
1527                tsp  0.888  0.862  3.09%    0+0       0+0   7.040  6.893   2.13%
1528               vliw 24.378 23.455  3.94%  106+127    25+45  2.758  2.707   1.91%
1529      --------------------------------------------------------------------------
1530       Average                     6.12%                                   4.09%
1531    
1532          SPARC        Compilation Time     Spill+Reload      Run Time
1533                     110.25  New            110.25    New   110.25  New
1534    
1535          barnesHut  3.778  3.592  5.20%    2+2       0+0   3.648  3.453    5.65%
1536              boyer  6.632  6.110  8.54%    0+0       0+0   0.258  0.242    6.90%
1537       count-graphs  1.435  1.325  8.30%    0+0       0+0  33.672 34.737   -3.07%
1538                fft  0.980  0.940  4.26%    3+9       2+6   0.838  0.827    1.41%
1539        knuthBendix  3.590  3.138 14.39%    0+0       0+0   0.962  0.967   -0.52%
1540             lexgen  6.593  6.072  8.59%    1+1       0+0   1.077  1.078   -0.15%
1541               life  0.972  0.868 11.90%   26+26      0+0   0.143  0.140    2.38%
1542              logic  2.525  2.387  5.80%    7+7       1+1   5.625  5.158    9.05%
1543         mandelbrot  0.090  0.093 -3.57%    0+0       0+0   0.855  0.728   17.39%
1544             mlyacc 26.732 23.827 12.19%  162+189    32+57  0.550  0.560   -1.79%
1545            nucleic  6.233  6.197  0.59%    3+3       0+0   0.163  0.173   -5.77%
1546      ratio-regions  3.780  3.507  7.79%    0+0       0+0 133.993 131.035   2.26%
1547                ray  1.595  1.550  2.90%    1+1       0+0   3.440  3.418    0.63%
1548             simple  6.972  6.487  7.48%   29+32      5+7   3.523  3.525   -0.05%
1549                tsp  1.115  1.063  4.86%    0+0       0+0   7.393  7.265    1.77%
1550               vliw 27.765 24.818 11.87%  110+135    25+45  2.265  2.135    6.09%
1551      ----------------------------------------------------------------------------
1552       Average                     6.94%                                    2.64%
1553    
1554          X86          Compilation Time     Spill+Reload      Run Time
1555                     110.25  New            110.25    New   110.25  New
1556    
1557          barnesHut  5.530  5.420  2.03%  593+893   597+915   3.532  3.440   2.66%
1558              boyer  8.768  7.747 13.19%  493+199   301+289   0.327  0.297  10.11%
1559       count-graphs  2.040  2.010  1.49%  298+394   315+457  26.578 28.660  -7.26%
1560                fft  1.327  1.302  1.92%  112+209   115+210   1.055  0.962   9.71%
1561        knuthBendix  5.218  5.475 -4.69%  451+598   510+650   0.928  0.932  -0.36%
1562             lexgen  9.970  9.623  3.60% 1014+841  1157+885   0.947  0.928   1.97%
1563               life  1.183  1.183  0.00%  162+182   145+148   0.127  0.103  22.58%
1564              logic  3.285  3.512 -6.45%  514+684   591+836   5.682  5.577   1.88%
1565         mandelbrot  0.147  0.143  2.33%   38+41     33+54    0.703  0.690   1.93%
1566             mlyacc 35.457 32.763  8.22% 3496+4564 3611+4860  0.552  0.550   0.30%
1567            nucleic  7.100  6.888  3.07%  239+168   201+158   0.175  0.173   0.96%
1568      ratio-regions  6.388  6.843 -6.65% 1182+257   981+300  120.142 120.345 -0.17%
1569                ray  2.332  2.338 -0.29%  346+398   402+494   3.593  3.540   1.51%
1570             simple  9.912  9.903  0.08% 1475+941  1579+1168  3.057  3.178  -3.83%
1571                tsp  1.623  1.532  5.98%  266+200   250+211   8.045  7.878   2.12%
1572               vliw 33.947 35.470 -4.29% 2629+2774 2877+3171  2.072  1.890   9.61%
1573      ----------------------------------------------------------------------------
1574       Average                     1.22%                                     3.36%
1575    
1576    ----------------------------------------------------------------------
1577  Name: Allen Leung  Name: Allen Leung
1578  Date: 2000/03/23 16:25:00  Date: 2000/03/23 16:25:00
1579  Tag: leunga-20000323-fix_x86_alpha  Tag: leunga-20000323-fix_x86_alpha
# Line 66  Line 1628 
1628    
1629  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
1630  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
1631  specifiel the layout of the tool command line.  See the CM manual for  specifies the layout of the tool command line.  See the CM manual for
1632  explanation.  explanation.
1633    
1634  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
1635  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
1636  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
1637  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 113  Line 1675 
1675  rebuild.  Having sets of the form <base><k>.{bin,boot}.<arch>-unix for  rebuild.  Having sets of the form <base><k>.{bin,boot}.<arch>-unix for
1676  <k>=1,2,... is normally not a good idea when invoking fixpt.  However,  <k>=1,2,... is normally not a good idea when invoking fixpt.  However,
1677  they might be the result of an earlier partial run of fixpt (which  they might be the result of an earlier partial run of fixpt (which
1678  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
1679  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
1680  and, thus, saves a lot of time.  and, thus, saves a lot of time.
1681    
# Line 163  Line 1725 
1725    it from that remote directory.    it from that remote directory.
1726    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
1727    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
1728    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
1729    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
1730    
1731    For CVS users, this mechanism is not relevant for source archives, but    For CVS users, this mechanism is not relevant for source archives, but

Legend:
Removed from v.583  
changed lines
  Added in v.686

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