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 583, Thu Mar 23 21:52:30 2000 UTC revision 694, Thu Jul 27 16:00:25 2000 UTC
# Line 12  Line 12 
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14  ----------------------------------------------------------------------  ----------------------------------------------------------------------
15    Name: Lal George
16    Date: 2000/07/27 11:53:14 EDT
17    
18    Tag: lal-20000727-linux-ppc
19    Description:
20    
21     Made changes to support Linux PPC.
22     p.s. I have confirmation that the 110.29 boot files work fine.
23    
24    ----------------------------------------------------------------------
25    Name: Matthias Blume
26    Date: 2000/07/27 17:40:00 JST
27    Tag: blume-20000727-scripts
28    Description:
29    
30    !!!! WARNING !!!!
31    You must recompile the runtime system!
32    !!!! WARNING !!!!
33    
34    This is basically another round of script-enhancements:
35    
36    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
37       and undefine CM preprocessor symbols.
38    
39    2. ml-build avoids generating a new heap image if it finds that the
40       existing one is still ok.  (The condition is that no ML file had to
41       be recompiled and all ML files are found to be older that the heap
42       file.)
43    
44       To make this work smoothly, I also hacked the runtime system as
45       well as SMLofNJ.SysInfo to get access to the heap image suffix
46       (.sparc-solaris, ...) that is currently being used.
47    
48       Moreover, the signature of CM.mk_standalone has changed.  See the
49       CM manual.
50    
51    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
52       CM manual for details.)
53    
54    4. More CM manual updates:
55        - all of the above has been documented.
56        - there is now a section describing the (CM-related) command line
57          arguments that are accepted by the "sml" command
58    
59    ----------------------------------------------------------------------
60    Name: Matthias Blume
61    Date: 2000/07/25 16:20:00 JST
62    Tag: blume-20000725-makedepend
63    Description:
64    
65    Added a script called ml-makedepend.  This can be used in makefiles
66    for Unix' make in a way very similar to the "makedepend" command for
67    C.
68    
69    The script internally uses function CM.sources.
70    
71    Synopsis:
72    
73        ml-makedepend [-f makefile] cmfile targetname
74    
75    The default for the makefile is "makefile" (or "Makefile" should
76    "makefile" not exist).
77    
78    ml-makedepend adds a cmfile/targetname-specific section to this
79    makefile (after removing the previous version of this section).  The
80    section contains a single dependency specification with targetname on
81    the LHS (targetname is an arbitrary name), and a list of files derived
82    from the cmfile on the RHS.  Some of the files on the RHS are
83    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
84    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
85    path names.  The makefile writer is responsible for making sure that
86    these variables have correct at the time "make" is invoked.
87    
88    ----------------------------------------------------------------------
89    Name: Matthias Blume
90    Date: 2000/07/22 23:30:00 JST
91    Tag: blume-20000722-urlupdate
92    Description:
93    
94    Changed BOOT and config/srcarchiveurl to point to BL server:
95    
96        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
97    
98    ----------------------------------------------------------------------
99    Name: Matthias Blume
100    Date: 2000/07/18 18:00:00 JST
101    Tag: blume-20000718-Version_110_29
102    Description:
103    
104    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
105    
106    2. Updated config/version to 110.29
107    
108    3. Updated config/srcarchiveurl
109    
110    3. New boot files!
111       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
112    
113    ----------------------------------------------------------------------
114    Name: Matthias Blume
115    Date: 2000/07/11 13:58:00 JST
116    Tag: blume-20000711-doctypo
117    Description:
118    
119    Fixed a few typos in CM manual.
120    
121    ----------------------------------------------------------------------
122    Name: Allen Leung
123    Date: 2000/06/15 00:38:00
124    Tag: leunga-20000704-sparc-x86
125    
126    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
127    2. fix to x86 RET bug [found by Dan Grossman]
128    3. sparc assembly bug fix for ticc instructions [found by Fermin]
129    
130       Affects c-- and moby only
131    
132    ----------------------------------------------------------------------
133    Name: Matthias Blume
134    Date: 2000/07/04 15:26:00
135    Tag: blume-20000704-trigger
136    Description:
137    
138    1. Improvements to CM manual.
139    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
140       of getting a back-trace.  The function, when called, raises an
141       internal exception which explicitly carries the full back-trace history,
142       so it is unaffected by any intervening handle-raise pairs ("trivial"
143       or not).  The interactive loop will print that history once it arrives
144       at top level.
145       Short of having all exceptions implicitly carry the full history, the
146       recommended way of using this facility is:
147         - compile your program with instrumentation "on"
148         - run it, when it raises an exception, look at the history
149         - if the history is "cut off" because of some handler, go and modify
150           your program so that it explicitly calls BTrace.trigger
151         - recompile (still instrumented), and rerun; look at the full history
152    
153    ----------------------------------------------------------------------
154    Name: Matthias Blume
155    Date: 2000/07/03 15:36:00 JST
156    Tag: blume-20000702-manual
157    Description:
158    
159    Small corrections and updates to CM manual.
160    
161    ----------------------------------------------------------------------
162    Name: Matthias Blume
163    Date: 2000/06/29 16:04:00 JST
164    Tag: blume-20000629-yacctool
165    Description:
166    
167    Changes:
168    
169    1. Class "mlyacc" now takes separate arguments to pass options to
170       generated .sml- and .sig-files independently.
171    2. Corresponding CM manual updates.
172    3. BTrace module now also reports call sites.  (However, for loop clusters
173       it only shows from where the cluster was entered.)  There are associated
174       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
175    
176    ----------------------------------------------------------------------
177    Name: Matthias Blume
178    Date: 2000/06/27 16:51:00 JST
179    Tag: blume-20000627-noweb
180    Description:
181    
182    Changes:
183    
184     1. Implemented "subdir" and "witness" options for noweb tool.
185        This caused some slight internal changes in CM's tool implementation.
186     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
187        some remaining issues from earlier path anchor changes.
188     3. Updated CM manual accordingly.
189    
190     4. Changed implementation of back-tracing so that I now consider it
191        ready for prime-time.
192    
193        In particular, you don't have to explicitly trigger the back-trace
194        anymore.  Instead, if you are running BTrace-instrumented code and
195        there is an uncaught exception (regardless of whether or not it was
196        raised in instrumented code), the top-level evalloop will print
197        the back-trace.
198    
199        Features:
200    
201          - Instrumented and uninstrumented code work together seemlessly.
202            (Of course, uninstrumented code is never mentioned in actual
203             back-traces.)
204    
205          - Asymptotic time- and space-complexity of instrumented code is
206            equal to that of uninstrumented code.  (This means that
207            tail-recursion is preserved by the instrumentation phase.)
208    
209          - Modules whose code has been instrumented in different sessions
210            work together without problem.
211    
212          - There is no penalty whatsoever on uninstrumented code.
213    
214          - There is no penalty on "raise" expressions, even in
215            instrumented code.
216    
217        A potential bug (or perhaps it is a feature, too):
218    
219          A back-trace reaches no further than the outermost instrumented
220          non-trivial "raise".  Here, a "trivial" raise is one that is the
221          sole RHS of a "handle" rule.  Thus, back-traces reach trough
222    
223               <exp> handle e => raise e
224    
225          and even
226    
227               <exp> handle Foo => raise Bar
228    
229          and, of course, through
230    
231               <exp> handle Foo => ...
232    
233         if the exception was not Foo.
234    
235         Back-traces always reach right through any un-instrumented code
236         including any of its "handle" expressions, trivial or not.
237    
238       To try this out, do the following:
239    
240         - Erase all existing binfiles for your program.
241           (You may keep binfiles for those modules where you think you
242            definitely don't need back-tracing.)
243         - Turn on back-trace instrumentation:
244              SMLofNJ.Internals.BTrace.mode (SOME true);
245         - Recompile your program.  (I.e., run "CM.make" or "use".)
246         - You may now turn instrumentation off again (if you want):
247              SMLofNJ.Internals.BTrace.mode (SOME false);
248         - Run your program as usual.  If it raises an exception that
249           reaches the interactive toplevel, then a back-trace will
250           automatically be printed.  After that, the toplevel loop
251           will print the exception history as usual.
252    
253    ----------------------------------------------------------------------
254    Name: Matthias Blume
255    Date: 2000/06/26 09:56:46 JST
256    Tag: blume-20000626-setup
257    Description:
258    
259    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
260          ML code before and after compiling a file (e.g., to set compiler
261          flags)
262    
263    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
264              - associated changes to btrace.sml (BTrace instrumentation pass)
265              - cleaner implementation of btimp.sml (BTrace tracing and report
266                module)
267    
268    CM manual: * new path encoding documented
269               * description of setup-parameter to "sml" added
270    
271    The biggest user-visible change to back-tracing is that it is no
272    longer necessary to compile all traced modules within the same
273    session.  (This was a real limitation.)
274    
275    ----------------------------------------------------------------------
276    Name: Matthias Blume
277    Date: 2000/06/24 12:40:00 JST
278    Tag: blume-20000624-startup
279    Description:
280    
281    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
282    bit too often -- to put it mildly. :)
283    
284    ----------------------------------------------------------------------
285    Name: Matthias Blume
286    Date: 2000/06/23 18:20:00 JST
287    Tag: blume-20000623-btrace
288    Description:
289    
290    This updates adds a backtrace facility to aid programmers in debugging
291    their programs.  This involves the following changes:
292    
293    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
294       keeping track of the current call stack.  When programs are compiled
295       in a special mode, the compiler will insert calls to these hooks
296       into the user program.
297       "Hook" means that it is possible for different implementations of
298       back-tracing to register themselves (at different times).
299    
300    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
301       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
302       It can be turned on using this call:
303         SMLofNJ.Internals.BTrace.mode (SOME true);
304       Turning it off again:
305         SMLofNJ.Internals.BTrace.mode (SOME false);
306       Querying the current status:
307         SMLofNJ.Internals.BTrace.mode NONE;
308       Annotated programs are about twice as big as normal ones, and they
309       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
310       where all hooks do nothing).  The slowdown with a plugin that is
311       actually useful (such as the one supplied by default) is even greater,
312       but in the case of the default plugin it is still only an constant
313       factor (amortized).
314    
315    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
316       with a sub-structure BTrace for controlling back-tracing.  In particular,
317       the above-mentioned function "mode" controls whether the annotation
318       phase is invoked by the compiler.  Another important function is
319       "trigger": when called it aborts the current execution and causes
320       the top-level loop to print a full back-trace.
321    
322    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
323       for back-tracing.  It keeps track of the dynamic call stack and in
324       addition to that it keeps a partial history at each "level" of that
325       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
326       d and b (at separate times, dynamically), then the report will show:
327    
328       GOTO   d
329             /c
330       GOTO  \b
331       CALL   a
332    
333       This shows that there was an initial non-tail call of a, then a
334       tail-call to b or c, looping behavior in a cluster of functions that
335       consist of b and c, and then a goto from that cluster (i.e., either from
336       b or from c) to d.
337    
338       Note that (depending on the user program) the amount of information
339       that the back-trace module has to keep track of at each level is bounded
340       by a constant.  Thus, the whole implementation has the same asymptotical
341       complexity as the original program (both in space and in time).
342    
343    5. compiler/TopLevel/interact/evalloop.sml has been modified to
344       handle the special exception SMLofNJ.Internals.BTrace.BTrace
345       which is raised by the "trigger" function mentioned above.
346    
347    Notes on usage:
348    
349    - Annotated code works well together with unannotated code:
350    Unannotated calls simply do not show up at all in the backtrace.
351    
352    - It is not a good idea to let modules that were annotated during
353    different sessions run at the same time.  This is because the compiler
354    chooses small integers to identify individual functions, and there
355    will be clashes if different modules were compiled in separate sessions.
356    (Nothing will crash, and you will even be told about the clashes, but
357    back-trace information will in general not be useful.)
358    
359    - Back-tracing can be confused by callcc and capture.
360    
361    - The only way of getting a back-trace right now is to explicitly
362    invoke the "trigger" function from your user program.  Eventually, we
363    should make every exception carry back-trace information (if
364    available).  But since this creates more overhead at "raise"-time
365    (similar to the current exnHistory overhead), I have not yet
366    implemented this.  (The implementation will be rather easy.)  With
367    exceptions carrying back-trace information, this facility will be even
368    more useful because users don't need to modify their programs...
369    
370    - While it is possible to compile the compiler with back-trace
371    annotations turned on (I did it to get some confidence in
372    correctness), you must make absolutely sure that core.sml and
373    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
374    be compiled with annotation because there is no core access yet, but
375    if you compile btimp.sml with annotation, then the system will go into
376    an infinite recursion and crash.)
377    Since CM currently does not know about BTrace, the only way to turn
378    annotations on and off for different modules of the compiler is to
379    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
380    this is awkward and clumsy.
381    
382    Sample sessions:
383    
384    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
385    - SMLofNJ.Internals.BTrace.mode (SOME true);
386    [autoloading]
387    [autoloading done]
388    val it = false : bool
389    - structure X = struct
390    -     fun main n = let
391    -         fun a (x, 0) = d x
392    -           | a (x, n) = b (x, n - 1)
393    -         and b (x, n) = c (x, n)
394    -         and c (x, n) = a (x, n)
395    -         and d x = e (x, 3)
396    -         and e (x, 0) = f x
397    -           | e (x, n) = e (x, n - 1)
398    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
399    -           | f n = n * g (n - 1)
400    -         and g n = a (n, 3)
401    -     in
402    -         f n
403    -     end
404    - end;
405    structure X : sig val main : int -> int end
406    - X.main 3;
407    *** BACK-TRACE ***
408    GOTO   stdIn:4.2-13.20: X.main[2].f
409    GOTO-( stdIn:4.2-13.20: X.main[2].e
410    GOTO   stdIn:4.2-13.20: X.main[2].d
411         / stdIn:4.2-13.20: X.main[2].a
412         | stdIn:4.2-13.20: X.main[2].b
413    GOTO-\ stdIn:4.2-13.20: X.main[2].c
414    CALL   stdIn:4.2-13.20: X.main[2].g
415    GOTO   stdIn:4.2-13.20: X.main[2].f
416    GOTO-( stdIn:4.2-13.20: X.main[2].e
417    GOTO   stdIn:4.2-13.20: X.main[2].d
418         / stdIn:4.2-13.20: X.main[2].a
419         | stdIn:4.2-13.20: X.main[2].b
420    GOTO-\ stdIn:4.2-13.20: X.main[2].c
421    CALL   stdIn:4.2-13.20: X.main[2].g
422    GOTO   stdIn:4.2-13.20: X.main[2].f
423    GOTO-( stdIn:4.2-13.20: X.main[2].e
424    GOTO   stdIn:4.2-13.20: X.main[2].d
425         / stdIn:4.2-13.20: X.main[2].a
426         | stdIn:4.2-13.20: X.main[2].b
427    GOTO-\ stdIn:4.2-13.20: X.main[2].c
428    CALL   stdIn:4.2-13.20: X.main[2].g
429    GOTO   stdIn:4.2-13.20: X.main[2].f
430    CALL   stdIn:2.15-17.4: X.main[2]
431    -
432    
433    (Note that because of a FLINt bug the above code currently does not
434    compile without BTrace turned on.)
435    
436    Here is another example, using my modified Tiger compiler:
437    
438    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
439    - SMLofNJ.Internals.BTrace.mode (SOME true);
440    [autoloading]
441    [autoloading done]
442    val it = false : bool
443    - CM.make "sources.cm";
444    [autoloading]
445    ...
446    [autoloading done]
447    [scanning sources.cm]
448    [parsing (sources.cm):parse.sml]
449    [creating directory CM/SKEL ...]
450    [parsing (sources.cm):tiger.lex.sml]
451    ...
452    [wrote CM/sparc-unix/semant.sml]
453    [compiling (sources.cm):main.sml]
454    [wrote CM/sparc-unix/main.sml]
455    [New bindings added.]
456    val it = true : bool
457    - Main.compile ("../testcases/merge.tig", "foo.out");
458    *** BACK-TRACE ***
459    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
460    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
461    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
462    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
463    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
464    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
465    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
466         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
467    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
468    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
469    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
470    CALL   main.sml:1.18-118.4: Main.compile[2]
471    -
472    
473    ----------------------------------------------------------------------
474    Name: Matthias Blumen
475    Date: 2000/06/21 18:00:00 JST
476    Tag: blume-20000621-manual
477    Description:
478    
479    CM manual update: Path environments documented.
480    
481    ----------------------------------------------------------------------
482    Name: Matthias Blume
483    Date: 2000/06/19 13:40:00
484    Tag: blume-20000619-manual
485    Description:
486    
487    CM manual and system/README update.  This only covers the fact that
488    there are no more implicit anchors.  (Path environments and the "bind"
489    option to "cm" have yet to be documented.)
490    
491    ----------------------------------------------------------------------
492    Name: Matthias Blume
493    Date: 2000/06/19 11:05:00 JST
494    Tag: blume-20000619-chdir-bugfix
495    Description:
496    
497    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
498    
499    ----------------------------------------------------------------------
500    Name: Matthias Blume
501    Date: 2000/06/18 22:00:10 JST
502    Tag: blume-20000618-implicit-anchors-really-gone
503    Description:
504    
505    I updates the previous HISTORY entry where I forgot to mention that
506    implicit anchors are no longer with us.
507    
508    The current update also gets rid of the (now useless) controller
509    CM.Control.implicit_anchors.
510    
511    ----------------------------------------------------------------------
512    Name: Matthias Blume
513    Date: 2000/06/16 17:30:00 JST
514    Tag: blume-20000616-anchorenv
515    Description:
516    
517    This patch implements the long anticipated (just kidding :) "anchor
518    environment" mechanism.  In the course of doing this, I also
519    re-implemented CM's internal "SrcPath" module from scratch.  The new
520    one should be more robust in certain boundary cases.  In any case, it
521    is a lot cleaner than its predecessor (IMHO).
522    
523    This time, although there is yet another boot file format change, I
524    kept the unpickler backward-compatible.  As a result, no new bootfiles
525    are necessary and bootstrapping is straightforward.  (You cannot read
526    new bootfiles into an old system, but the other way around is no
527    problem.)
528    
529    Visible changes:
530    
531    ** 0. Implicit path anchors (without the leading $-symbol) are no
532    longer recognized at all. This means that such path names are not
533    illegal either.  For example, the name basis.cm simply refers to a
534    local file called "basis.cm" (i.e, the name is an ordinary path
535    relative to .cm-files directory).  Or, to put it differently, only
536    names that start with $ are anchored paths.
537    
538    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
539    vanished.
540    
541    John (Reppy) had critizised this as soon as I originally proposed and
542    implemented it, but at that time I did not really deeply believe
543    him. :) Now I came full-circle because I need the $<singlearc> syntax
544    in another place where it cannot be seen as an abbreviation for
545    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
546    seems to mean (i.e., it "expands" into the corresponding anchor
547    value).
548    
549    However, when paths are used as members in CM description files, it
550    continues to be true that there must be at least another arc after the
551    anchor.  This is now enforced separately during semantic analysis
552    (i.e., from a lexical/syntactical point of view, the notation is ok.)
553    
554    ** 2. The "cm" class now accepts an option "bind".  The option's value
555    is a sub-option list of precisely two items -- one labeled "anchor"
556    and the other one labeled "value".  As you might expect, "anchor" is
557    used to specify an anchor name to be bound, and "value" specifies what
558    the anchor is being bound to.
559    
560    The value must be a directory name and can be given in either standard
561    syntax (including the possibility that it is itself an anchored path)
562    or native syntax.
563    
564    Examples:
565    
566       foo.cm (bind:(anchor:bar value:$mystuff/bar))
567       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
568    
569    and so on.
570    
571    The meaning of this is that the .cm-file will be processed with an
572    augmented anchor environment where the given anchor(s) is/are bound to
573    the given values(s).
574    
575    The rationale for having this feature is this: Suppose you are trying
576    to use two different (already stable) libraries a.cm and b.cm (that
577    you perhaps didn't write yourself).  Further, suppose each of these
578    two libraries internally uses its own auxiliary library $aux/lib.cm.
579    Normally you would now have a problem because the anchor "lib" can not
580    be bound to more than one value globally.  Therefore, the project that
581    uses both a.cm and b.cm must locally redirect the anchor to some other
582    place:
583    
584       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
585       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
586    
587    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
588    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
589    
590    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
591    you when you do so at the time of CM.stabilize).  Therefore, you can
592    also use an anchored path as the value:
593    
594      a.cm (bind:(anchor:lib value:$a-lib))
595      b.cm (bind:(anchor:lib value:$b-lib))
596    
597    Now you can globally configure (using the usual CM.Anchor.anchor or
598    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
599    itself is always locally bound, setting it globally is no longer
600    meaningful or necessary (but it does not hurt either).  In fact, "lib"
601    can still be used as a global anchor for separate purposes.  As a
602    matter of fact, one can locally define "lib" in terms of a global
603    "lib":
604    
605      a.cm (bind:(anchor:lib value:$lib/a))
606      b.cm (bind:(anchor:lib value:$lib/b))
607    
608    ** 3: The encoding of path names has changed.  This affects the way
609    path names are shown in CM's progress report and also the internal
610    protocol encoding used for parallel make.
611    
612    The encoding now uses one or more ':'-separated segments.  Each
613    segments corresponds to a file that has been specified relative to the
614    file given by its preceding segment.  The first segment is either
615    relative to the CWD, absolute, or anchored.  Each segment itself is
616    basically a Unix pathname; all segments but the first are relative.
617    
618    Example:
619    
620       $foo/bar/baz.cm:a/b/c.sml
621    
622    This path denotes the file bar/a/b/c.sml relative to the directory
623    denoted by anchor "foo".  Notice that the encoding also includes
624    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
625    paths are resolved relative to the .cm-files directory, so baz.cm must
626    be ignored to get the "real" pathname.
627    
628    To make this fact more obvious, CM puts the names of such "virtual
629    arcs" into parentheses when they appear in progress reports. (No
630    parentheses will appear in the internal protocol encoding.)  Thus,
631    what you really see is:
632    
633      $foo/bar/(baz.cm):a/b/c.sml
634    
635    I find this notation to be much more informative than before.
636    
637    Another new feature of the encoding is that special characters
638    including parentheses, colons, (back)slashes, and white space are
639    written as \ddd (where ddd is the decimal encoding of the character).
640    
641    *** The CM manual still needs to be updated.
642    
643    ----------------------------------------------------------------------
644    Name: Allen Leung
645    Date: 2000/06/15 00:38:00
646    Tag: leunga-20000615-x86-peephole
647    
648    x86 Peephole fix by Fermin.  Affects c-- and moby only.
649    
650    ----------------------------------------------------------------------
651    Name: Matthias Blume
652    Date: 2000/06/12 11:40:00
653    Tag: blume-20000612-parmakefix
654    Description:
655    
656    More cleanup after changing the file naming scheme: This time I
657    repaired the parallel make mechanism for CMB.make which I broke earlier.
658    
659    ----------------------------------------------------------------------
660    Name: Allen Leung
661    Date: 2000/06/09 01:25:00
662    Tag: leunga-20000609-various
663    
664    None of these things should affect normal SML/NJ operations
665    
666    1. Peephole improvements provided by Fermin (c--)
667    2. New annotation DEFUSE for adding extra dependence (moby)
668    3. New X86 LOCK instructions (moby)
669    4. New machine description language for reservation tables (scheduling)
670    5. Fixes to various optimization/analysis modules (branch chaining, dominator
671       trees etc.)
672    6. I've changed the CM files so that they can work with versions
673       110.0.6, 110.25 and 110.28
674    
675    ----------------------------------------------------------------------
676    Name: Matthias Blume
677    Date: 2000/06/09 12:40:00
678    Tag: blume-20000609-log
679    Description:
680    
681    - Removed all(?) remaining RCS Log entries from sources.
682    
683    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
684      anchored paths).
685    
686    ----------------------------------------------------------------------
687    Name: Matthias Blume
688    Date: 2000/06/07 17:00:00 JST
689    Tag: blume-20000607-no-implicit-anchors
690    Description:
691    
692    1. This update changes the default setting for
693    CM.Control.implicit_anchors from true to false.  This means that
694    implicit anchors are no longer permitted by default.  I also tried to
695    make sure that nothing else still relies on implicit anchors.
696    (This is the next step on the schedule towards a CM that does not even
697    have the notion of implicit anchors anymore.)
698    
699    2. More CM manual updates.
700    
701    3. I managed to track down and fix the pickling bug I mentioned last
702    time.  Because of the previously existing workaround, this entails no
703    immediate practical changes.
704    
705    ----------------------------------------------------------------------
706    Name: Matthias Blume
707    Date: 2000/06/06 11:15:00 JST
708    Tag: blume-20000606-lazierpickle
709    Description:
710    
711    !!!! NEW BOOT FILES !!!!
712    
713    * The main purpose of this update is to make library pickles lazier in
714    order to reduce the initial space penalty for autoloading a library.
715    As a result, it is now possible to have $smlnj/compiler.cm
716    pre-registered.  This should take care of the many complaints or
717    inquiries about missing structure Compiler.  This required changes to
718    CM's internal data structures and small tweaks to some algorithms.
719    
720    As a neat additional effect, it is no longer necessary (for the sake
721    of lean heap image files) to distinguish between a "minimal" CM and a
722    "full" CM.  Now, there is only one CM (i.e., the "full" version:
723    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
724    interactive top level. ($smlnj/cm/minimal.cm is gone.)
725    
726    To make the life of compiler-hackers easier, "makeml" now also
727    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
728    words, after you bootstrap a new sml for the first time, you will not
729    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
730    around you will still have to do it, though.)
731    
732    * A second change consists of major updates to the CM manual.  There
733    are now several appendices with summary information and also a full
734    specification of the CM description file syntax.
735    
736    * In directory src/system I added the script "allcross".  This script
737    invokes sml and cross-compiles the compiler for all supported
738    architectures.  (Useful when providing a new set of boot files.)
739    
740    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
741    added a small tweak to pickle-util.sml to work around this problem,
742    but it is not a proper fix yet.  I will investigate further.  (The
743    effect of the bug was an inflation of library pickle size.)
744    
745    * Version number increased to 110.28.1 (to avoid compatibility problems).
746    
747    ----------------------------------------------------------------------
748    Name: Allen Leung
749    Date: 2000/05/25 17:28 EDT
750    Tag: leunga-20000525-ra
751    Description:
752    
753      Fixed a bug in freezing phase of the register allocator.
754    
755    ----------------------------------------------------------------------
756    Name: Allen Leung
757    Date: 2000/05/15 22:53 EDT
758    Tag: leunga-20000515-alpha-x86-ra
759    Description:
760    
761      1. Alpha
762    
763          Slight cleanup.  Removed the instruction SGNXL
764    
765      2. X86
766    
767          Added the following instructions to the instruction set:
768    
769            ROLx, RORx,
770            BTx, BTSx, BTLx, BTRx,
771            XCHGx, and variants with the LOCK prefix
772    
773      3. Register Allocation
774    
775          The module ra-rewrite-with-renaming has been improved.
776    
777      These have no effect on SML/NJ.
778    
779    ----------------------------------------------------------------------
780    Name: Matthias Blume
781    Date: 2000/05/15 16:20:00 JST
782    Tag: blume-20000515-lightrebuild
783    Description:
784    
785    1. I added an alternative to "-rebuild" to "makeml".  The difference is
786       that prior to calling CMB.make' the CM-variable "LIGHT" will be
787       defined.  In effect, the command will not build any cross-compiler
788       backends and therefore finish more quickly.
789    
790       The "fixpt" script also takes a "-light" switch to be able to use
791       this new facility while compiling for a fixpoint.
792    
793    2. I replaced all mentions of anchored paths in group owner specifications
794       with simple relative paths (usually starting with "..").
795       The rationale is that a library's internal workings should not be
796       compromised by the lack of some anchor.  (An anchor is necessary
797       for someone who wants to refer to the library by an anchored path,
798       but it should not be necessary to build the same library in the first
799       place.)
800    
801    3. I changed the way CM's tool mechanism determines the shell command
802       string used for things like ml-yacc etc. so that it does not break
803       when CM.Control.implicit_anchors is turned off.
804    
805    ----------------------------------------------------------------------
806    Name: Matthias Blume
807    Date: 2000/05/12 18:20:00 JST
808    Tag: blume-20000512-ml-build
809    Description:
810    
811    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
812    getting installed properly (by config/install.sh).
813    
814    ----------------------------------------------------------------------
815    Name: Matthias Blume
816    Date: 2000/05/12 17:30:00 JST
817    Tag: blume-20000512-anchors
818    Description:
819    
820    !!! NEW BOOT FILES !!!
821    
822    This change is in preparation of fading out support for "implicitly
823    anchored path names".  I went through all sources and used the
824    explicit (and relatively new) $-notation.  See system/README and the
825    CM manual for more info on this.
826    
827    I also modified the anchoring scheme for some things such as "smlnj",
828    "MLRISC", "cm", etc. to take advantage of the fact that explicit
829    anchors are more expressive: anchor name and first arc do not have to
830    coincide.  This entails the following user-visible change:
831    
832    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
833    particular, when you fire up sml with a command-line argument, say,
834    e.g.:
835    
836       sml '$smlnj/cmb.cm'
837    
838    At the ML toplevel prompt:
839    
840       CM.autoload "$smlnj/cmb.cm";
841    
842    There is also a new controller in CM.Control that can be used to turn
843    off all remaining support for implicit anchors by saying:
844    
845        CM.autoload "$smlnj/
846        #set CM.Control.implicit_anchors false;
847    
848    This causes CM to reject implicitly anchored paths.  This is (for the
849    time being) less permissive than the "final" version where there will
850    be no more such implicit anchors and relative paths will be just that:
851    relative.
852    
853    The next step (version after next version?) will be to make the
854    default for CM.Control.implicit_anchors false.  After the dust has
855    settled, I can then produce the "final" version of this...
856    
857    Note: Since bootstrapping is a bit tricky, I provided new boot files.
858    
859    ----------------------------------------------------------------------
860    Name: Matthias Blume
861    Date: 2000/05/11 16:30:00 JST
862    Tag: blume-20000511-sources
863    Description:
864    
865    The main change is that I added function CM.sources as a generalized
866    version of the earlier CM.makedepend.  This entails the following
867    additional changes:
868    
869      - CM.makedepend has been dropped.
870    
871      - CM manual has been updated.
872    
873      - TOOLS signature and API have been changed.
874    
875    ----------------------------------------------------------------------
876    Name: Allen Leung
877    Date: 2000/05/10 21:17 EDT
878    Tag: leunga-20000510-moby-c--ssa
879    Description:
880    
881      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
882    None of these affect SML/NJ.
883    
884    1. Register Allocation
885    
886        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
887           This module tries to remove local (i.e. basic block level) redundancies
888           during spilling.
889    
890        b. A new framework for performing region based register allocation.
891           Not yet entirely functional.
892    
893    2. X86
894    
895       a. DefUse for POP was missing the stack pointer [found by Lal]
896       b. Reload for CALL was incorrect in X86Spill [found by John]
897       c. Various fixes in X86Spill so that it can be used correctly for
898          the new spilling module.
899    
900    3. SSA/IR
901    
902       a. New module ir/dj-dataflow.sml implements elimination based
903          data flow analysis.
904    
905    4. MLRiscGen
906    
907       a. Fix for gc type annotation
908    
909    5. MDGen
910    
911       Various fixes for machine description -> ml code translation.  For ssa
912       only.
913    
914    ----------------------------------------------------------------------
915    Name: Allen Leung
916    Date: 2000/05/08 22:17 EDT
917    Tag: leunga-20000508-labexp
918    Description:
919    
920      Fermin has found a few assembly problems with constant expressions
921      generated in LabelExp.  Mostly, the problems involve extra parentheses,
922      which choke on dumb assemblers.  This is his fix.
923    
924    ----------------------------------------------------------------------
925    Name: Dave MacQueen
926    Date: 2000/04/09 14:00 EDT
927    Tag: dbm-20000502-Version_110_28
928    Description:
929    
930    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
931    
932    2. Updated config/version to 110.28
933    
934    3. Updated config/srcarchiveurl
935    
936    3. New boot files!
937       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
938    
939    ----------------------------------------------------------------------
940    Name: Matthias Blume
941    Date: 2000/05/01 19:05:00 JST
942    Tag: blume-20000501-noweb
943    Description:
944    
945    A new noweb tool has been added.  The existing system is entirely
946    unaffected by this, but some CM users have asked for renewed noweb
947    support.  Everything is documented in the CM manual.
948    
949    New (plugin) libraries:
950    
951       noweb-tool.cm
952       nw-ext.cm
953    
954    ----------------------------------------------------------------------
955    Name: Dave MacQueen
956    Date: 2000/04/30 12:40PM EDT
957    Tag: dbm-20000430-bug_fixes
958    Description:
959    
960    1. Fix for bug 1498
961       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
962       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
963         added toRealArray function
964       smlnj/src/compiler/MiscUtil/print/ppobj.sml
965         added check for tag Obj.RealArray to array printing case in ppObj
966    
967    2. Fix for bug 1510
968       smlnj/src/compiler/Semant/types/typesutil.sml
969         fixed definition of dummyargs (used by equalTycon) so that
970         dummy args are distinct types
971    
972    ----------------------------------------------------------------------
973    Name: Matthias Blume
974    Date: 2000/04/30 01:00:00 JST
975    Tag: blume-20000430-versions
976    Description:
977    
978    1. CM version numbering added.  This is an implementation of Lal's
979       proposal for adding version numbers and version checking to .cm
980       files.  Lal said that his proposal was just that -- a proposal.
981       For the time being I went ahead and implemented it so that people
982       can comment on it.  Everything is completely backward-compatible
983       (except for the stable library format, i.e., new bootfiles!).
984    
985       As usual, see the CM manual for details.
986    
987    2. An alternative syntax for anchored paths has been implemented.
988       Dave has recently voiced the same concerns that I had when I did
989       this, so there should be some support.  My take is that eventually
990       I will let support for the current syntax (where anchors are
991       "implicit") fade out in favor of the new, explicit syntax.
992       In order to be backward-compatible, both old and new syntax are
993       currently supported.
994    
995       Again, see the CM manual for details.
996    
997    3. Parallel make is trying to be slightly smarter:  When the master
998       process finds a "bottleneck", i.e., when there is only one
999       compilation unit that can be compiled and everybody else is
1000       waiting on it, then it will simply compile it directly instead
1001       of clumsily telling one of the slaves to do it.
1002    
1003    4. Support for "unsharing" added.  This is necessary in order to be
1004       able to have two different versions of the same library running
1005       at the same time (e.g., for trying out a new MLRISC while still
1006       having the old MLRISC linked into the current compiler, etc.)
1007       See the CM manual.
1008    
1009    5. Simple "makedepend" functionality added for generating Makefile
1010       dependency information.  (This is rather crude at the moment.
1011       Expect some changes here in the future.)
1012    
1013    6. ".fun" added as a recognized suffix for ML files. Also documented
1014       explicitly in the manual that the fallback behavior (unknown suffix
1015       -> ML file) is not an official feature!
1016    
1017    7. Small changes to the pickler for stable libraries.
1018    
1019    8. Several internal changes to CM (for cleanup/improvement).
1020    
1021    
1022    !!!! NEW BINFILES !!!!
1023    
1024    ----------------------------------------------------------------------
1025    Name: Matthias Blume
1026    Date: 2000/04/28 17:30:00 JST
1027    Tag: blume-20000428-pathconfig
1028    Description:
1029    
1030    1. I changed config/install.sh to remove duplicate entries from the
1031       lib/pathconfig file at the end.  Moreover, the final version of
1032       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1033       in src/system/installml.
1034    
1035    2. The config/install.sh script now consistently uses relative
1036       pathnames in lib/pathconfig whenever the anchor is in the lib
1037       directory.  (So far this was true for the libraries that come
1038       pre-compiled and bundled as part of the bootfiles but not for
1039       libraries that are compiled by the script itself.)
1040    
1041    ----------------------------------------------------------------------
1042    Name: Matthias Blume
1043    Date: 2000/04/26 13:10:00 JST
1044    Tag: blume-20000426-fun_suffix
1045    Description:
1046    
1047    Added ".fun" as a recognized file name suffix (for ML code).
1048    
1049    ----------------------------------------------------------------------
1050    Name: Allen Leung
1051    Date: 2000/04/25 17:00:00 EST
1052    Tag: leunga-20000425-alpha-ra
1053    Description:
1054    
1055    1. Alpha
1056    
1057        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1058    in C--.
1059    
1060    2. RA
1061    
1062       Added a flag "ra-dump-size" to print out the size of the flowgraph
1063       and the interference graph.
1064    
1065    ----------------------------------------------------------------------
1066    Name: Dave MacQueen
1067    Date: 2000/04/25/
1068    Tag: dbm-20000425-mlyacc_doc_examples
1069    Description:
1070      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1071      Updated all three examples in src/ml-yacc/examples to run
1072      under 110.* using CM.make.
1073    
1074    ----------------------------------------------------------------------
1075    Name: Allen Leung
1076    Date: 2000/04/20 23:04:00 EST
1077    Tag: leunga-20000420-ssa-c---stuff
1078    Description:
1079    
1080      This update synchronizes my repository with Yale's.  Most of these
1081    changes, however, do not affect SML/NJ at all (the RA is an exception).
1082    
1083    1. Register Allocator
1084    
1085       a. An improvement in the interference graph construction:
1086          Given a copy
1087    
1088                s <- t
1089    
1090          no interference edge between s and t is added for this definition of s.
1091    
1092       b. I've added two new spill heuristic modules that Fermin and I developed
1093          (in the new library RA.cm). These are unused in SML/NJ but maybe
1094          useful for others (Moby?)
1095    
1096    2. X86
1097    
1098       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1099    
1100    3. Alpha
1101    
1102       a. Added the BSR instruction and code generation that goes with it [C--]
1103       b. Other fixes too numerous to recount provided by Fermin [C--]
1104    
1105    4. Regmaps
1106    
1107       a. The regmaps are not initialized with the identity physical bindings
1108          at creation time.  This is unneeded.
1109    
1110    5. MLRISC Optimizations
1111    
1112       a. The DJ-Graph module can now compute the iterated dominance frontiers
1113          intersects with liveness incrementally in linear time! Woohoo!
1114          This is now used in my new SSA construction algorithm.
1115    
1116       b. THe branch reorganization module is now smarter about linear chains of
1117          basic blocks.
1118    
1119    
1120    ----------------------------------------------------------------------
1121    Name: Matthias Blume
1122    Date: 2000/04/12 13:52:00 JST
1123    Tag: blume_main_v110p27_1
1124    Description:
1125    
1126    Changed install.sh script to handle archive files without version number
1127    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1128    name of the boot file archive.
1129    
1130    ----------------------------------------------------------------------
1131    Name: Dave MacQueen
1132    Date: 2000/04/09 14:00 EDT
1133    Tag: dbm-20000410-Version_110_27
1134    Description:
1135    
1136    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1137    
1138    2. Updated src/config/version to 110.27
1139    
1140    3. New boot files!
1141    
1142    ----------------------------------------------------------------------
1143    Name: Allen Leung
1144    Date: 2000/04/09 19:09:00 EST
1145    Tag: leunga-20000409-misc
1146    Description:
1147    
1148    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1149    
1150    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1151    
1152    ----------------------------------------------------------------------
1153    Name: Stefan
1154    Date: 2000/04/07 10:00:00 EDT
1155    Tag: monnier-20000406-branch-handling
1156    Description:
1157    
1158    Improved handling of branches (mostly those generated from
1159    polymorphic equality), removed switchoff and changed the
1160    default optimization settings (more cpsopt and less flintopt).
1161    
1162    ----------------------------------------------------------------------
1163    Name: Allen Leung
1164    Date: 2000/04/06 01:30:00 EST
1165    Tag: leunga-20000406-peephole-x86-SSA-2
1166    Description:
1167    
1168       Forgot a few files.
1169    
1170    ----------------------------------------------------------------------
1171    Name: Allen Leung
1172    Date: 2000/04/06 00:36:00 EST
1173    Tag: leunga-20000406-peephole-x86-SSA
1174    Description:
1175    
1176    1.  New Peephole code
1177    
1178    2.  Minor improvement to X86 instruction selection
1179    
1180    3.  Various fixes to SSA and machine description -> code translator
1181    
1182    ----------------------------------------------------------------------
1183    Name: Matthias Blume
1184    Date: 2000/04/05 12:30:00 JST
1185    Tag: blume_main_v110p26p2_3
1186    Description:
1187    
1188    This update just merges three minor cosmetic updates to CM's sources
1189    to get ready for the 110.27 code freeze on Friday.  No functionality
1190    has changed.
1191    
1192    ----------------------------------------------------------------------
1193    Name: Allen Leung
1194    Date: 2000/04/04 19:39:00 EST
1195    Tag: leunga-20000404-x86-asm
1196    Description:
1197    
1198    1.  Fixed a problem in X86 assembly.
1199    
1200        Things like
1201    
1202           jmp %eax
1203           jmp (%eax)
1204    
1205        should be output as
1206    
1207           jmp *%eax
1208           jmp *(%eax)
1209    
1210    2.  Assembly output
1211    
1212          Added a new flag
1213    
1214              "asm-indent-copies" (default to false)
1215    
1216          When this flag is on, parallel copies will be indented an extra level.
1217    
1218    ----------------------------------------------------------------------
1219    Name: Allen Leung
1220    Date: 2000/04/04 03:18:00 EST
1221    Tag: leunga-20000404-C--Moby
1222    Description:
1223    
1224        All of these fixes are related to C--, Moby, and my own optimization
1225        stuff; so they shouldn't affect SML/NJ.
1226    
1227    1.  X86
1228    
1229        Various fixes related floating point, and extensions.
1230    
1231    2.  Alpha
1232    
1233        Some extra patterns related to loads with signed/zero extension
1234        provided by Fermin.
1235    
1236    3.  Assembly
1237    
1238        When generating assembly, resolve the value of client defined constants,
1239        instead of generating symbolic values.  This is controlled by the
1240        new flag "asm-resolve-constants", which is default to true.
1241    
1242    4.  Machine Descriptions
1243    
1244        a. The precedence parser was slightly broken when parsing infixr symbols.
1245        b. The type generalizing code had the bound variables reversed, resulting
1246           in a problem during arity raising.
1247        c. Various fixes in machine descriptions.
1248    
1249    ----------------------------------------------------------------------
1250    Name: Matthias Blume
1251    Date: 2000/04/03 16:05:00 JST
1252    Tag: blume_main_v110p26p2_2
1253    Description:
1254    
1255    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
1256    now done via the ordinary static environment that is context to each
1257    compilation unit.
1258    
1259    To this end, I arranged that instead of "structure Core" as "structure
1260    _Core" is bound in the pervasive environment.  Core access is done via
1261    _Core (which can never be accidentally rebound because _Core is not a
1262    legal surface-syntax symbol).
1263    
1264    The current solution is much cleaner because the core environment is
1265    now simply part of the pervasive environment which is part of every
1266    compilation unit's context anyway.  In particular, this eliminates all
1267    special-case handling that was necessary until now in order to deal
1268    with dynamic and symbolic parts of the core environment.
1269    
1270    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1271    compilation manager's bootstrap compiler (actually: in the "init group"
1272    handling).  See the comments in src/system/smlnj/init/init.cmi for
1273    more details.
1274    
1275    I also tried to track down all mentions of "Core" (as string argument
1276    to Symbol.strSymbol) in the compiler and replaced them with a
1277    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
1278    name appears in one place only.
1279    
1280    Binfile and bootfile format have not changed, but the switchover from
1281    the old "init.cmi" to the new one is a bit tricky, so I supplied new
1282    bootfiles anyway.
1283    
1284    ----------------------------------------------------------------------
1285    Name: Allen Leung
1286    Date: 2000/04/02 21:17:00 EST
1287    Tag: leunga-20000402-mltree
1288    Description:
1289    
1290       1. Renamed the constructor CALL in MLTREE by popular demand.
1291       2. Added a bunch of files from my repository.  These are currently
1292          used by other non-SMLNJ backends.
1293    
1294    ----------------------------------------------------------------------
1295    Name: Allen Leung
1296    Date: 2000/03/31 21:15:00 EST
1297    Tag: leunga-20000331-aliasing
1298    Description:
1299    
1300    This update contains a rewritten (and hopefully more correct) module
1301    for extracting aliasing information from CPS.
1302    
1303       To turn on this feature:
1304    
1305            Compiler.Control.CG.memDisambiguate := true
1306    
1307       To pretty print the region information with assembly
1308    
1309           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
1310    
1311       To control how many levels of aliasing information are printed, use:
1312    
1313           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
1314    
1315       The default of n is 3.
1316    
1317    ----------------------------------------------------------------------
1318    Name: David MacQueen
1319    Date: 2000/03/31 11:15:00 EST
1320    Tag: dbm-20000331-runtime_fix
1321    Description:
1322    
1323    This update contains:
1324    
1325    1. runtime/c-lib/c-libraries.c
1326       includes added in revision 1.2 caused compilation errors on hppa-hpux
1327    
1328    2. fix for bug 1556
1329       system/Basis/Implementation/NJ/internal-signals.sml
1330    
1331    ----------------------------------------------------------------------
1332    Name: Matthias Blume
1333    Date: 2000/03/31 18:00:00 JST
1334    Tag: blume_main_v110p26p2_1
1335    Description:
1336    
1337    This update contains:
1338    
1339    1. A small change to CM's handling of stable libraries:
1340       CM now maintains one "global" modmap that is used for all stable
1341       libraries.  The use of such a global modmap maximizes sharing and
1342       minimizes the need for re-traversing parts of environments during
1343       modmap construction.  (However, this has minor impact since modmap
1344       construction seems to account for just one percent or less of total
1345       compile time.)
1346    
1347    2. I added a "genmap" phase to the statistics.  This is where I got the
1348       "one percent" number (see above).
1349    
1350    3. CM's new tool parameter mechanism just became _even_ better. :)
1351       - The parser understands named parameters and recursive options.
1352       - The "make" and "shell" tools use these new features.
1353         (This makes it a lot easier to cascade these tools.)
1354       - There is a small syntax change: named parameters use a
1355    
1356           <name> : ( <option> ... )            or
1357           <name> : <string>
1358    
1359         syntax.  Previously, named parameters were implemented in an
1360         ad-hoc fashion by each tool individually (by parsing strings)
1361         and had the form
1362    
1363           <name>=<string>
1364    
1365       See the CM manual for a full description of these issues.
1366    
1367    ----------------------------------------------------------------------
1368    Name: Matthias Blume
1369    Date: 2000/03/30 18:00:00 JST
1370    Tag: blume_main_v110p26p2_0
1371    Description:
1372    
1373    !!!!! WARNING !!!!!!
1374    !!  New binfiles  !!
1375    !!!!!!!!!!!!!!!!!!!!
1376    
1377    This update contains:
1378    
1379    1. Moderate changes to CM:
1380    
1381       - Changes to CM's tools mechanism.  In particular, it is now possible
1382       to have tools that accept additional "command line" parameters
1383       (specified in the .cm file at each instance where the tool's class is
1384       used).
1385    
1386       This was done to accommodate the new "make" and "shell" tools which
1387       facilitate fairly seamless hookup to portions of code managed using
1388       Makefiles or Shell scripts.
1389    
1390       There are no classes "shared" or "private" anymore.  Instead, the
1391       sharing annotation is now a parameter to the "sml" class.
1392    
1393       There is a bit of generic machinery for implementing one's own
1394       tools that accept command-line parameters.  However, I am not yet fully
1395       satisfied with that part, so expect changes here in the future.
1396    
1397       All existing tools are described in the CM manual.
1398    
1399       - Slightly better error handling.  (CM now suppresses many followup
1400       error messages that tended to be more annoying than helpful.)
1401    
1402    2. Major changes to the compiler's static environment data structures.
1403    
1404       - no CMStaticEnv anymore.
1405            - no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
1406              but it is called Environment), no conversions between different
1407              kinds of static environments
1408    
1409       - There is still a notion of a "modmap", but such modmaps are generated
1410         on demand at the time when they are needed.  This sounds slow, but I
1411         sped up the code that generates modmaps enough for this not to lead to
1412         a slowdown of the compiler (at least I didn't detect any).
1413    
1414       - To facilitate rapid modmap generation, static environments now
1415         contain an (optional) "modtree" structure.  Modtree annotations are
1416         constructed by the unpickler during unpickling.  (This means that
1417         the elaborator does not have to worry about modtrees at all.)
1418         Modtrees have the advantage that they are compositional in the same
1419         way as the environment data structure itself is compositional.
1420         As a result, modtrees never hang on to parts of an environment that
1421         has already been rendered "stale" by filtering or rebinding.
1422    
1423       - I went through many, many trials and errors before arriving at the
1424         current solution.  (The initial idea of "linkpaths" did not work.)
1425         But the result of all this is that I have touched a lot of files that
1426         depend on the "modules" and "types" data structures (most of the
1427         elaborator). There were a lot of changes during my "linkpath" trials
1428         that could have been reverted to their original state but weren't.
1429         Please, don't be too harsh on me for messing with this code a bit more
1430         than what was strictly necessary...  (I _did_ resist the tempation
1431         of doing any "global reformatting" to avoid an untimely death at
1432         Dave's hands. :)
1433    
1434       - One positive aspect of the previous point:  At least I made sure that
1435         all files that I touched now compile without warnings (other than
1436         "polyEqual").
1437    
1438       - compiler now tends to run "leaner" (i.e., ties up less memory in
1439         redundant modmaps)
1440    
1441    ----------------------------------------------------------------------
1442    Name: Allen Leung
1443    Date: 2000/03/29 18:00:00
1444    Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
1445    Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
1446    Description:
1447    
1448       This update contains *MAJOR* changes to the way code is generated from CPS
1449    in the module mlriscGen, and in various backend modules.
1450    
1451    CHANGES
1452    =======
1453    
1454    1. MLRiscGen: forward propagation fix.
1455    
1456       There was a bug in forward propagation introduced at about the same time
1457       as the MLRISC x86 backend, which prohibits coalescing to be
1458       performed effectively in loops.
1459    
1460       Effect: speed up of loops in RISC architectures.
1461               By itself, this actually slowed down certain benchmarks on the x86.
1462    
1463    2. MLRiscGen:  forward propagating addresses from consing.
1464    
1465       I've changed the way consing code is generated.  Basically I separated
1466       out the initialization part:
1467    
1468            store tag,   offset(allocptr)
1469            store elem1, offset+4(allocptr)
1470            store elem2, offset+8(allocptr)
1471            ...
1472            store elemn, offset+4n(allocptr)
1473    
1474       and the address computation part:
1475    
1476            celladdr <- offset+4+alloctpr
1477    
1478       and move the address computation part
1479    
1480       Effect:  register pressure is generally lower as a result.  This
1481                makes compilation of certain expressions much faster, such as
1482                long lists with non-trivial elements.
1483    
1484                 [(0,0), (0,0), .... (0,0)]
1485    
1486    3. MLRiscGen: base pointer elimination.
1487    
1488        As part of the linkage mechanism, we generate the sequence:
1489    
1490         L:  ...  <- start of the code fragment
1491    
1492         L1:
1493             base pointer <- linkreg - L1 + L
1494    
1495         The base pointer was then used for computing relocatable addresses
1496       in the code fragment.  Frequently (such as in lots of continuations)
1497       this is not needed.  We now eliminate this sequence whenever possible.
1498    
1499         For compile time efficiency, I'm using a very stupid local heuristic.
1500       But in general, this should be done as a control flow analysis.
1501    
1502       Effect:  Smaller code size.  Speed up of most programs.
1503    
1504    4. Hppa back end
1505    
1506         Long jumps in span dependence resolution used to depend on the existence
1507      of the base pointer.
1508    
1509         A jump to a long label L was expanded into the following sequence:
1510    
1511          LDIL %hi(L-8192), %r29
1512          LDO  %lo(L-8192)(%r29), %r29
1513          ADD  %r29, baseptr, %r29
1514          BV,n %r0(%r29)
1515    
1516         In the presence of change (3) above, this will not work.  I've changed
1517       it so that the following sequence of instructions are generated, which
1518       doesn't mention the base pointer at all:
1519    
1520             BL,n  L', %r29           /* branch and link, L' + 4 -> %r29 */
1521        L':  ADDIL L-(L'+4), %r29     /* Compute address of L */
1522             BV,n  %r0(%r29)          /* Jump */
1523    
1524    5. Alpha back end
1525    
1526          New alpha instructions LDB/LDW have been added, as per Fermin's
1527       suggestions.   This is unrelated to all other changes.
1528    
1529    6. X86 back end
1530    
1531         I've changed andl to testl in the floating point test sequence
1532         whenever appropriate.  The Intel optimization guide states that
1533         testl is preferable to andl.
1534    
1535    7. RA (x86 only)
1536    
1537         I've improved the spill propagation algorithm, using an approximation
1538       of maximal weighted independent sets.   This seems to be necessary to
1539       alleviate the negative effect in light of the slow down in (1).
1540    
1541         I'll write down the algorithm one of these days.
1542    
1543    8. MLRiscGen: frequencies
1544    
1545         I've added an annotation that states that all call gc blocks have zero
1546       execution frequencies.  This improves register allocation on the x86.
1547    
1548    BENCHMARKS
1549    ==========
1550    
1551       I've only perform the comparison on 110.25.
1552    
1553       The platforms are:
1554    
1555        HPPA  A four processor HP machine (E9000) with 5G of memory.
1556        X86   A 300Hhz Pentium II with 128M of memory, and
1557        SPARC An Ultra sparc 2 with 512M of memory.
1558    
1559       I used the following parameters for the SML benchmarks:
1560    
1561                 @SMLalloc
1562         HPPA    256k
1563         SPARC   512k
1564         X86     256k
1565    
1566    COMPILATION TIME
1567    ----------------
1568       Here are the numbers comparing the compilation times of the compilers.
1569       I've only compared 110.25 compiling the new sources versus
1570       a fixpoint version of the new compiler compiling the same.
1571    
1572                     110.25                                  New
1573               Total  Time in RA  Spill+Reload   Total  Time In RA Spill+Reload
1574         HPPA   627s    116s        2684+3584     599s    95s       1003+1879
1575         SPARC  892s    173s        2891+3870     708s    116s      1004+1880
1576         X86    999s    315s       94006+130691   987s    296s    108877+141957
1577    
1578                   110.25         New
1579                Code Size      Code Size
1580         HPPA   8596736         8561421
1581         SPARC  8974299         8785143
1582         X86    9029180         8716783
1583    
1584       So in summary, things are at least as good as before.   Dramatic
1585       reduction in compilation is obtained on the Sparc; I can't explain it,
1586       but it is reproducible.  Perhaps someone should try to reproduce this
1587       on their own machines.
1588    
1589    SML BENCHMARKS
1590    --------------
1591    
1592        On the average, all benchmarks perform at least as well as before.
1593    
1594          HPPA         Compilation Time     Spill+Reload      Run Time
1595                     110.25  New            110.25    New   110.25  New
1596    
1597          barnesHut  3.158  3.015  4.75%    1+1       0+0   2.980  2.922   2.00%
1598              boyer  6.152  5.708  7.77%    0+0       0+0   0.218  0.213   2.34%
1599       count-graphs  1.168  1.120  4.32%    0+0       0+0  22.705 23.073  -1.60%
1600                fft  0.877  0.792 10.74%    1+3       1+3   0.602  0.587   2.56%
1601        knuthBendix  3.180  2.857 11.32%    0+0       0+0   0.675  0.662   2.02%
1602             lexgen  6.190  5.290 17.01%    0+0       0+0   0.913  0.788  15.86%
1603               life  0.803  0.703 14.22%   25+25      0+0   0.153  0.140   9.52%
1604              logic  2.048  2.007  2.08%    6+6       1+1   4.133  4.008   3.12%
1605         mandelbrot  0.077  0.080 -4.17%    0+0       0+0   0.765  0.712   7.49%
1606             mlyacc 22.932 20.937  9.53%  154+181    32+57  0.468  0.430   8.91%
1607            nucleic  5.183  5.060  2.44%    2+2       0+0   0.125  0.120   4.17%
1608      ratio-regions  3.357  3.142  6.84%    0+0       0+0  116.225 113.173 2.70%
1609                ray  1.283  1.290 -0.52%    0+0       0+0   2.887  2.855   1.11%
1610             simple  6.307  6.032  4.56%   28+30      5+7   3.705  3.658   1.28%
1611                tsp  0.888  0.862  3.09%    0+0       0+0   7.040  6.893   2.13%
1612               vliw 24.378 23.455  3.94%  106+127    25+45  2.758  2.707   1.91%
1613      --------------------------------------------------------------------------
1614       Average                     6.12%                                   4.09%
1615    
1616          SPARC        Compilation Time     Spill+Reload      Run Time
1617                     110.25  New            110.25    New   110.25  New
1618    
1619          barnesHut  3.778  3.592  5.20%    2+2       0+0   3.648  3.453    5.65%
1620              boyer  6.632  6.110  8.54%    0+0       0+0   0.258  0.242    6.90%
1621       count-graphs  1.435  1.325  8.30%    0+0       0+0  33.672 34.737   -3.07%
1622                fft  0.980  0.940  4.26%    3+9       2+6   0.838  0.827    1.41%
1623        knuthBendix  3.590  3.138 14.39%    0+0       0+0   0.962  0.967   -0.52%
1624             lexgen  6.593  6.072  8.59%    1+1       0+0   1.077  1.078   -0.15%
1625               life  0.972  0.868 11.90%   26+26      0+0   0.143  0.140    2.38%
1626              logic  2.525  2.387  5.80%    7+7       1+1   5.625  5.158    9.05%
1627         mandelbrot  0.090  0.093 -3.57%    0+0       0+0   0.855  0.728   17.39%
1628             mlyacc 26.732 23.827 12.19%  162+189    32+57  0.550  0.560   -1.79%
1629            nucleic  6.233  6.197  0.59%    3+3       0+0   0.163  0.173   -5.77%
1630      ratio-regions  3.780  3.507  7.79%    0+0       0+0 133.993 131.035   2.26%
1631                ray  1.595  1.550  2.90%    1+1       0+0   3.440  3.418    0.63%
1632             simple  6.972  6.487  7.48%   29+32      5+7   3.523  3.525   -0.05%
1633                tsp  1.115  1.063  4.86%    0+0       0+0   7.393  7.265    1.77%
1634               vliw 27.765 24.818 11.87%  110+135    25+45  2.265  2.135    6.09%
1635      ----------------------------------------------------------------------------
1636       Average                     6.94%                                    2.64%
1637    
1638          X86          Compilation Time     Spill+Reload      Run Time
1639                     110.25  New            110.25    New   110.25  New
1640    
1641          barnesHut  5.530  5.420  2.03%  593+893   597+915   3.532  3.440   2.66%
1642              boyer  8.768  7.747 13.19%  493+199   301+289   0.327  0.297  10.11%
1643       count-graphs  2.040  2.010  1.49%  298+394   315+457  26.578 28.660  -7.26%
1644                fft  1.327  1.302  1.92%  112+209   115+210   1.055  0.962   9.71%
1645        knuthBendix  5.218  5.475 -4.69%  451+598   510+650   0.928  0.932  -0.36%
1646             lexgen  9.970  9.623  3.60% 1014+841  1157+885   0.947  0.928   1.97%
1647               life  1.183  1.183  0.00%  162+182   145+148   0.127  0.103  22.58%
1648              logic  3.285  3.512 -6.45%  514+684   591+836   5.682  5.577   1.88%
1649         mandelbrot  0.147  0.143  2.33%   38+41     33+54    0.703  0.690   1.93%
1650             mlyacc 35.457 32.763  8.22% 3496+4564 3611+4860  0.552  0.550   0.30%
1651            nucleic  7.100  6.888  3.07%  239+168   201+158   0.175  0.173   0.96%
1652      ratio-regions  6.388  6.843 -6.65% 1182+257   981+300  120.142 120.345 -0.17%
1653                ray  2.332  2.338 -0.29%  346+398   402+494   3.593  3.540   1.51%
1654             simple  9.912  9.903  0.08% 1475+941  1579+1168  3.057  3.178  -3.83%
1655                tsp  1.623  1.532  5.98%  266+200   250+211   8.045  7.878   2.12%
1656               vliw 33.947 35.470 -4.29% 2629+2774 2877+3171  2.072  1.890   9.61%
1657      ----------------------------------------------------------------------------
1658       Average                     1.22%                                     3.36%
1659    
1660    ----------------------------------------------------------------------
1661  Name: Allen Leung  Name: Allen Leung
1662  Date: 2000/03/23 16:25:00  Date: 2000/03/23 16:25:00
1663  Tag: leunga-20000323-fix_x86_alpha  Tag: leunga-20000323-fix_x86_alpha
# Line 66  Line 1712 
1712    
1713  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
1714  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
1715  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
1716  explanation.  explanation.
1717    
1718  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
1719  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
1720  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
1721  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 1759 
1759  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
1760  <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,
1761  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
1762  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
1763  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
1764  and, thus, saves a lot of time.  and, thus, saves a lot of time.
1765    
# Line 163  Line 1809 
1809    it from that remote directory.    it from that remote directory.
1810    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
1811    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
1812    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
1813    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
1814    
1815    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.694

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