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 585, Wed Mar 29 23:55:35 2000 UTC revision 682, Tue Jul 4 06:25:51 2000 UTC
# Line 11  Line 11 
11  Date:  Date:
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    ----------------------------------------------------------------------
15    Name: Matthias Blume
16    Date: 2000/07/04 15:26:00
17    Tag: blume-20000704-trigger
18    Description:
19    
20    1. Improvements to CM manual.
21    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
22       of getting a back-trace.  The function, when called, raises an
23       internal exception which explicitly carries the full back-trace history,
24       so it is unaffected by any intervening handle-raise pairs ("trivial"
25       or not).  The interactive loop will print that history once it arrives
26       at top level.
27       Short of having all exceptions implicitly carry the full history, the
28       recommended way of using this facility is:
29         - compile your program with instrumentation "on"
30         - run it, when it raises an exception, look at the history
31         - if the history is "cut off" because of some handler, go and modify
32           your program so that it explicitly calls BTrace.trigger
33         - recompile (still instrumented), and rerun; look at the full history
34    
35    ----------------------------------------------------------------------
36    Name: Matthias Blume
37    Date: 2000/07/03 15:36:00 JST
38    Tag: blume-20000702-manual
39    Description:
40    
41    Small corrections and updates to CM manual.
42    
43    ----------------------------------------------------------------------
44    Name: Matthias Blume
45    Date: 2000/06/29 16:04:00 JST
46    Tag: blume-20000629-yacctool
47    Description:
48    
49    Changes:
50    
51    1. Class "mlyacc" now takes separate arguments to pass options to
52       generated .sml- and .sig-files independently.
53    2. Corresponding CM manual updates.
54    3. BTrace module now also reports call sites.  (However, for loop clusters
55       it only shows from where the cluster was entered.)  There are associated
56       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
57    
58    ----------------------------------------------------------------------
59    Name: Matthias Blume
60    Date: 2000/06/27 16:51:00 JST
61    Tag: blume-20000627-noweb
62    Description:
63    
64    Changes:
65    
66     1. Implemented "subdir" and "witness" options for noweb tool.
67        This caused some slight internal changes in CM's tool implementation.
68     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
69        some remaining issues from earlier path anchor changes.
70     3. Updated CM manual accordingly.
71    
72     4. Changed implementation of back-tracing so that I now consider it
73        ready for prime-time.
74    
75        In particular, you don't have to explicitly trigger the back-trace
76        anymore.  Instead, if you are running BTrace-instrumented code and
77        there is an uncaught exception (regardless of whether or not it was
78        raised in instrumented code), the top-level evalloop will print
79        the back-trace.
80    
81        Features:
82    
83          - Instrumented and uninstrumented code work together seemlessly.
84            (Of course, uninstrumented code is never mentioned in actual
85             back-traces.)
86    
87          - Asymptotic time- and space-complexity of instrumented code is
88            equal to that of uninstrumented code.  (This means that
89            tail-recursion is preserved by the instrumentation phase.)
90    
91          - Modules whose code has been instrumented in different sessions
92            work together without problem.
93    
94          - There is no penalty whatsoever on uninstrumented code.
95    
96          - There is no penalty on "raise" expressions, even in
97            instrumented code.
98    
99        A potential bug (or perhaps it is a feature, too):
100    
101          A back-trace reaches no further than the outermost instrumented
102          non-trivial "raise".  Here, a "trivial" raise is one that is the
103          sole RHS of a "handle" rule.  Thus, back-traces reach trough
104    
105               <exp> handle e => raise e
106    
107          and even
108    
109               <exp> handle Foo => raise Bar
110    
111          and, of course, through
112    
113               <exp> handle Foo => ...
114    
115         if the exception was not Foo.
116    
117         Back-traces always reach right through any un-instrumented code
118         including any of its "handle" expressions, trivial or not.
119    
120       To try this out, do the following:
121    
122         - Erase all existing binfiles for your program.
123           (You may keep binfiles for those modules where you think you
124            definitely don't need back-tracing.)
125         - Turn on back-trace instrumentation:
126              SMLofNJ.Internals.BTrace.mode (SOME true);
127         - Recompile your program.  (I.e., run "CM.make" or "use".)
128         - You may now turn instrumentation off again (if you want):
129              SMLofNJ.Internals.BTrace.mode (SOME false);
130         - Run your program as usual.  If it raises an exception that
131           reaches the interactive toplevel, then a back-trace will
132           automatically be printed.  After that, the toplevel loop
133           will print the exception history as usual.
134    
135    ----------------------------------------------------------------------
136    Name: Matthias Blume
137    Date: 2000/06/26 09:56:46 JST
138    Tag: blume-20000626-setup
139    Description:
140    
141    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
142          ML code before and after compiling a file (e.g., to set compiler
143          flags)
144    
145    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
146              - associated changes to btrace.sml (BTrace instrumentation pass)
147              - cleaner implementation of btimp.sml (BTrace tracing and report
148                module)
149    
150    CM manual: * new path encoding documented
151               * description of setup-parameter to "sml" added
152    
153    The biggest user-visible change to back-tracing is that it is no
154    longer necessary to compile all traced modules within the same
155    session.  (This was a real limitation.)
156    
157    ----------------------------------------------------------------------
158    Name: Matthias Blume
159    Date: 2000/06/24 12:40:00 JST
160    Tag: blume-20000624-startup
161    Description:
162    
163    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
164    bit too often -- to put it mildly. :)
165    
166    ----------------------------------------------------------------------
167    Name: Matthias Blume
168    Date: 2000/06/23 18:20:00 JST
169    Tag: blume-20000623-btrace
170    Description:
171    
172    This updates adds a backtrace facility to aid programmers in debugging
173    their programs.  This involves the following changes:
174    
175    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
176       keeping track of the current call stack.  When programs are compiled
177       in a special mode, the compiler will insert calls to these hooks
178       into the user program.
179       "Hook" means that it is possible for different implementations of
180       back-tracing to register themselves (at different times).
181    
182    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
183       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
184       It can be turned on using this call:
185         SMLofNJ.Internals.BTrace.mode (SOME true);
186       Turning it off again:
187         SMLofNJ.Internals.BTrace.mode (SOME false);
188       Querying the current status:
189         SMLofNJ.Internals.BTrace.mode NONE;
190       Annotated programs are about twice as big as normal ones, and they
191       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
192       where all hooks do nothing).  The slowdown with a plugin that is
193       actually useful (such as the one supplied by default) is even greater,
194       but in the case of the default plugin it is still only an constant
195       factor (amortized).
196    
197    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
198       with a sub-structure BTrace for controlling back-tracing.  In particular,
199       the above-mentioned function "mode" controls whether the annotation
200       phase is invoked by the compiler.  Another important function is
201       "trigger": when called it aborts the current execution and causes
202       the top-level loop to print a full back-trace.
203    
204    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
205       for back-tracing.  It keeps track of the dynamic call stack and in
206       addition to that it keeps a partial history at each "level" of that
207       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
208       d and b (at separate times, dynamically), then the report will show:
209    
210       GOTO   d
211             /c
212       GOTO  \b
213       CALL   a
214    
215       This shows that there was an initial non-tail call of a, then a
216       tail-call to b or c, looping behavior in a cluster of functions that
217       consist of b and c, and then a goto from that cluster (i.e., either from
218       b or from c) to d.
219    
220       Note that (depending on the user program) the amount of information
221       that the back-trace module has to keep track of at each level is bounded
222       by a constant.  Thus, the whole implementation has the same asymptotical
223       complexity as the original program (both in space and in time).
224    
225    5. compiler/TopLevel/interact/evalloop.sml has been modified to
226       handle the special exception SMLofNJ.Internals.BTrace.BTrace
227       which is raised by the "trigger" function mentioned above.
228    
229    Notes on usage:
230    
231    - Annotated code works well together with unannotated code:
232    Unannotated calls simply do not show up at all in the backtrace.
233    
234    - It is not a good idea to let modules that were annotated during
235    different sessions run at the same time.  This is because the compiler
236    chooses small integers to identify individual functions, and there
237    will be clashes if different modules were compiled in separate sessions.
238    (Nothing will crash, and you will even be told about the clashes, but
239    back-trace information will in general not be useful.)
240    
241    - Back-tracing can be confused by callcc and capture.
242    
243    - The only way of getting a back-trace right now is to explicitly
244    invoke the "trigger" function from your user program.  Eventually, we
245    should make every exception carry back-trace information (if
246    available).  But since this creates more overhead at "raise"-time
247    (similar to the current exnHistory overhead), I have not yet
248    implemented this.  (The implementation will be rather easy.)  With
249    exceptions carrying back-trace information, this facility will be even
250    more useful because users don't need to modify their programs...
251    
252    - While it is possible to compile the compiler with back-trace
253    annotations turned on (I did it to get some confidence in
254    correctness), you must make absolutely sure that core.sml and
255    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
256    be compiled with annotation because there is no core access yet, but
257    if you compile btimp.sml with annotation, then the system will go into
258    an infinite recursion and crash.)
259    Since CM currently does not know about BTrace, the only way to turn
260    annotations on and off for different modules of the compiler is to
261    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
262    this is awkward and clumsy.
263    
264    Sample sessions:
265    
266    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
267    - SMLofNJ.Internals.BTrace.mode (SOME true);
268    [autoloading]
269    [autoloading done]
270    val it = false : bool
271    - structure X = struct
272    -     fun main n = let
273    -         fun a (x, 0) = d x
274    -           | a (x, n) = b (x, n - 1)
275    -         and b (x, n) = c (x, n)
276    -         and c (x, n) = a (x, n)
277    -         and d x = e (x, 3)
278    -         and e (x, 0) = f x
279    -           | e (x, n) = e (x, n - 1)
280    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
281    -           | f n = n * g (n - 1)
282    -         and g n = a (n, 3)
283    -     in
284    -         f n
285    -     end
286    - end;
287    structure X : sig val main : int -> int end
288    - X.main 3;
289    *** BACK-TRACE ***
290    GOTO   stdIn:4.2-13.20: X.main[2].f
291    GOTO-( stdIn:4.2-13.20: X.main[2].e
292    GOTO   stdIn:4.2-13.20: X.main[2].d
293         / stdIn:4.2-13.20: X.main[2].a
294         | stdIn:4.2-13.20: X.main[2].b
295    GOTO-\ stdIn:4.2-13.20: X.main[2].c
296    CALL   stdIn:4.2-13.20: X.main[2].g
297    GOTO   stdIn:4.2-13.20: X.main[2].f
298    GOTO-( stdIn:4.2-13.20: X.main[2].e
299    GOTO   stdIn:4.2-13.20: X.main[2].d
300         / stdIn:4.2-13.20: X.main[2].a
301         | stdIn:4.2-13.20: X.main[2].b
302    GOTO-\ stdIn:4.2-13.20: X.main[2].c
303    CALL   stdIn:4.2-13.20: X.main[2].g
304    GOTO   stdIn:4.2-13.20: X.main[2].f
305    GOTO-( stdIn:4.2-13.20: X.main[2].e
306    GOTO   stdIn:4.2-13.20: X.main[2].d
307         / stdIn:4.2-13.20: X.main[2].a
308         | stdIn:4.2-13.20: X.main[2].b
309    GOTO-\ stdIn:4.2-13.20: X.main[2].c
310    CALL   stdIn:4.2-13.20: X.main[2].g
311    GOTO   stdIn:4.2-13.20: X.main[2].f
312    CALL   stdIn:2.15-17.4: X.main[2]
313    -
314    
315    (Note that because of a FLINt bug the above code currently does not
316    compile without BTrace turned on.)
317    
318    Here is another example, using my modified Tiger compiler:
319    
320    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
321    - SMLofNJ.Internals.BTrace.mode (SOME true);
322    [autoloading]
323    [autoloading done]
324    val it = false : bool
325    - CM.make "sources.cm";
326    [autoloading]
327    ...
328    [autoloading done]
329    [scanning sources.cm]
330    [parsing (sources.cm):parse.sml]
331    [creating directory CM/SKEL ...]
332    [parsing (sources.cm):tiger.lex.sml]
333    ...
334    [wrote CM/sparc-unix/semant.sml]
335    [compiling (sources.cm):main.sml]
336    [wrote CM/sparc-unix/main.sml]
337    [New bindings added.]
338    val it = true : bool
339    - Main.compile ("../testcases/merge.tig", "foo.out");
340    *** BACK-TRACE ***
341    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
342    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
343    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
344    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
345    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
346    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
347    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
348         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
349    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
350    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
351    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
352    CALL   main.sml:1.18-118.4: Main.compile[2]
353    -
354    
355    ----------------------------------------------------------------------
356    Name: Matthias Blumen
357    Date: 2000/06/21 18:00:00 JST
358    Tag: blume-20000621-manual
359    Description:
360    
361    CM manual update: Path environments documented.
362    
363    ----------------------------------------------------------------------
364    Name: Matthias Blume
365    Date: 2000/06/19 13:40:00
366    Tag: blume-20000619-manual
367    Description:
368    
369    CM manual and system/README update.  This only covers the fact that
370    there are no more implicit anchors.  (Path environments and the "bind"
371    option to "cm" have yet to be documented.)
372    
373    ----------------------------------------------------------------------
374    Name: Matthias Blume
375    Date: 2000/06/19 11:05:00 JST
376    Tag: blume-20000619-chdir-bugfix
377    Description:
378    
379    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
380    
381    ----------------------------------------------------------------------
382    Name: Matthias Blume
383    Date: 2000/06/18 22:00:10 JST
384    Tag: blume-20000618-implicit-anchors-really-gone
385    Description:
386    
387    I updates the previous HISTORY entry where I forgot to mention that
388    implicit anchors are no longer with us.
389    
390    The current update also gets rid of the (now useless) controller
391    CM.Control.implicit_anchors.
392    
393    ----------------------------------------------------------------------
394    Name: Matthias Blume
395    Date: 2000/06/16 17:30:00 JST
396    Tag: blume-20000616-anchorenv
397    Description:
398    
399    This patch implements the long anticipated (just kidding :) "anchor
400    environment" mechanism.  In the course of doing this, I also
401    re-implemented CM's internal "SrcPath" module from scratch.  The new
402    one should be more robust in certain boundary cases.  In any case, it
403    is a lot cleaner than its predecessor (IMHO).
404    
405    This time, although there is yet another boot file format change, I
406    kept the unpickler backward-compatible.  As a result, no new bootfiles
407    are necessary and bootstrapping is straightforward.  (You cannot read
408    new bootfiles into an old system, but the other way around is no
409    problem.)
410    
411    Visible changes:
412    
413    ** 0. Implicit path anchors (without the leading $-symbol) are no
414    longer recognized at all. This means that such path names are not
415    illegal either.  For example, the name basis.cm simply refers to a
416    local file called "basis.cm" (i.e, the name is an ordinary path
417    relative to .cm-files directory).  Or, to put it differently, only
418    names that start with $ are anchored paths.
419    
420    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
421    vanished.
422    
423    John (Reppy) had critizised this as soon as I originally proposed and
424    implemented it, but at that time I did not really deeply believe
425    him. :) Now I came full-circle because I need the $<singlearc> syntax
426    in another place where it cannot be seen as an abbreviation for
427    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
428    seems to mean (i.e., it "expands" into the corresponding anchor
429    value).
430    
431    However, when paths are used as members in CM description files, it
432    continues to be true that there must be at least another arc after the
433    anchor.  This is now enforced separately during semantic analysis
434    (i.e., from a lexical/syntactical point of view, the notation is ok.)
435    
436    ** 2. The "cm" class now accepts an option "bind".  The option's value
437    is a sub-option list of precisely two items -- one labeled "anchor"
438    and the other one labeled "value".  As you might expect, "anchor" is
439    used to specify an anchor name to be bound, and "value" specifies what
440    the anchor is being bound to.
441    
442    The value must be a directory name and can be given in either standard
443    syntax (including the possibility that it is itself an anchored path)
444    or native syntax.
445    
446    Examples:
447    
448       foo.cm (bind:(anchor:bar value:$mystuff/bar))
449       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
450    
451    and so on.
452    
453    The meaning of this is that the .cm-file will be processed with an
454    augmented anchor environment where the given anchor(s) is/are bound to
455    the given values(s).
456    
457    The rationale for having this feature is this: Suppose you are trying
458    to use two different (already stable) libraries a.cm and b.cm (that
459    you perhaps didn't write yourself).  Further, suppose each of these
460    two libraries internally uses its own auxiliary library $aux/lib.cm.
461    Normally you would now have a problem because the anchor "lib" can not
462    be bound to more than one value globally.  Therefore, the project that
463    uses both a.cm and b.cm must locally redirect the anchor to some other
464    place:
465    
466       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
467       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
468    
469    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
470    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
471    
472    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
473    you when you do so at the time of CM.stabilize).  Therefore, you can
474    also use an anchored path as the value:
475    
476      a.cm (bind:(anchor:lib value:$a-lib))
477      b.cm (bind:(anchor:lib value:$b-lib))
478    
479    Now you can globally configure (using the usual CM.Anchor.anchor or
480    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
481    itself is always locally bound, setting it globally is no longer
482    meaningful or necessary (but it does not hurt either).  In fact, "lib"
483    can still be used as a global anchor for separate purposes.  As a
484    matter of fact, one can locally define "lib" in terms of a global
485    "lib":
486    
487      a.cm (bind:(anchor:lib value:$lib/a))
488      b.cm (bind:(anchor:lib value:$lib/b))
489    
490    ** 3: The encoding of path names has changed.  This affects the way
491    path names are shown in CM's progress report and also the internal
492    protocol encoding used for parallel make.
493    
494    The encoding now uses one or more ':'-separated segments.  Each
495    segments corresponds to a file that has been specified relative to the
496    file given by its preceding segment.  The first segment is either
497    relative to the CWD, absolute, or anchored.  Each segment itself is
498    basically a Unix pathname; all segments but the first are relative.
499    
500    Example:
501    
502       $foo/bar/baz.cm:a/b/c.sml
503    
504    This path denotes the file bar/a/b/c.sml relative to the directory
505    denoted by anchor "foo".  Notice that the encoding also includes
506    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
507    paths are resolved relative to the .cm-files directory, so baz.cm must
508    be ignored to get the "real" pathname.
509    
510    To make this fact more obvious, CM puts the names of such "virtual
511    arcs" into parentheses when they appear in progress reports. (No
512    parentheses will appear in the internal protocol encoding.)  Thus,
513    what you really see is:
514    
515      $foo/bar/(baz.cm):a/b/c.sml
516    
517    I find this notation to be much more informative than before.
518    
519    Another new feature of the encoding is that special characters
520    including parentheses, colons, (back)slashes, and white space are
521    written as \ddd (where ddd is the decimal encoding of the character).
522    
523    *** The CM manual still needs to be updated.
524    
525    ----------------------------------------------------------------------
526    Name: Allen Leung
527    Date: 2000/06/15 00:38:00
528    Tag: leunga-20000615-x86-peephole
529    
530    x86 Peephole fix by Fermin.  Affects c-- and moby only.
531    
532    ----------------------------------------------------------------------
533    Name: Matthias Blume
534    Date: 2000/06/12 11:40:00
535    Tag: blume-20000612-parmakefix
536    Description:
537    
538    More cleanup after changing the file naming scheme: This time I
539    repaired the parallel make mechanism for CMB.make which I broke earlier.
540    
541    ----------------------------------------------------------------------
542    Name: Allen Leung
543    Date: 2000/06/09 01:25:00
544    Tag: leunga-20000609-various
545    
546    None of these things should affect normal SML/NJ operations
547    
548    1. Peephole improvements provided by Fermin (c--)
549    2. New annotation DEFUSE for adding extra dependence (moby)
550    3. New X86 LOCK instructions (moby)
551    4. New machine description language for reservation tables (scheduling)
552    5. Fixes to various optimization/analysis modules (branch chaining, dominator
553       trees etc.)
554    6. I've changed the CM files so that they can work with versions
555       110.0.6, 110.25 and 110.28
556    
557    ----------------------------------------------------------------------
558    Name: Matthias Blume
559    Date: 2000/06/09 12:40:00
560    Tag: blume-20000609-log
561    Description:
562    
563    - Removed all(?) remaining RCS Log entries from sources.
564    
565    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
566      anchored paths).
567    
568    ----------------------------------------------------------------------
569    Name: Matthias Blume
570    Date: 2000/06/07 17:00:00 JST
571    Tag: blume-20000607-no-implicit-anchors
572    Description:
573    
574    1. This update changes the default setting for
575    CM.Control.implicit_anchors from true to false.  This means that
576    implicit anchors are no longer permitted by default.  I also tried to
577    make sure that nothing else still relies on implicit anchors.
578    (This is the next step on the schedule towards a CM that does not even
579    have the notion of implicit anchors anymore.)
580    
581    2. More CM manual updates.
582    
583    3. I managed to track down and fix the pickling bug I mentioned last
584    time.  Because of the previously existing workaround, this entails no
585    immediate practical changes.
586    
587    ----------------------------------------------------------------------
588    Name: Matthias Blume
589    Date: 2000/06/06 11:15:00 JST
590    Tag: blume-20000606-lazierpickle
591    Description:
592    
593    !!!! NEW BOOT FILES !!!!
594    
595    * The main purpose of this update is to make library pickles lazier in
596    order to reduce the initial space penalty for autoloading a library.
597    As a result, it is now possible to have $smlnj/compiler.cm
598    pre-registered.  This should take care of the many complaints or
599    inquiries about missing structure Compiler.  This required changes to
600    CM's internal data structures and small tweaks to some algorithms.
601    
602    As a neat additional effect, it is no longer necessary (for the sake
603    of lean heap image files) to distinguish between a "minimal" CM and a
604    "full" CM.  Now, there is only one CM (i.e., the "full" version:
605    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
606    interactive top level. ($smlnj/cm/minimal.cm is gone.)
607    
608    To make the life of compiler-hackers easier, "makeml" now also
609    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
610    words, after you bootstrap a new sml for the first time, you will not
611    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
612    around you will still have to do it, though.)
613    
614    * A second change consists of major updates to the CM manual.  There
615    are now several appendices with summary information and also a full
616    specification of the CM description file syntax.
617    
618    * In directory src/system I added the script "allcross".  This script
619    invokes sml and cross-compiles the compiler for all supported
620    architectures.  (Useful when providing a new set of boot files.)
621    
622    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
623    added a small tweak to pickle-util.sml to work around this problem,
624    but it is not a proper fix yet.  I will investigate further.  (The
625    effect of the bug was an inflation of library pickle size.)
626    
627    * Version number increased to 110.28.1 (to avoid compatibility problems).
628    
629    ----------------------------------------------------------------------
630    Name: Allen Leung
631    Date: 2000/05/25 17:28 EDT
632    Tag: leunga-20000525-ra
633    Description:
634    
635      Fixed a bug in freezing phase of the register allocator.
636    
637    ----------------------------------------------------------------------
638    Name: Allen Leung
639    Date: 2000/05/15 22:53 EDT
640    Tag: leunga-20000515-alpha-x86-ra
641    Description:
642    
643      1. Alpha
644    
645          Slight cleanup.  Removed the instruction SGNXL
646    
647      2. X86
648    
649          Added the following instructions to the instruction set:
650    
651            ROLx, RORx,
652            BTx, BTSx, BTLx, BTRx,
653            XCHGx, and variants with the LOCK prefix
654    
655      3. Register Allocation
656    
657          The module ra-rewrite-with-renaming has been improved.
658    
659      These have no effect on SML/NJ.
660    
661    ----------------------------------------------------------------------
662    Name: Matthias Blume
663    Date: 2000/05/15 16:20:00 JST
664    Tag: blume-20000515-lightrebuild
665    Description:
666    
667    1. I added an alternative to "-rebuild" to "makeml".  The difference is
668       that prior to calling CMB.make' the CM-variable "LIGHT" will be
669       defined.  In effect, the command will not build any cross-compiler
670       backends and therefore finish more quickly.
671    
672       The "fixpt" script also takes a "-light" switch to be able to use
673       this new facility while compiling for a fixpoint.
674    
675    2. I replaced all mentions of anchored paths in group owner specifications
676       with simple relative paths (usually starting with "..").
677       The rationale is that a library's internal workings should not be
678       compromised by the lack of some anchor.  (An anchor is necessary
679       for someone who wants to refer to the library by an anchored path,
680       but it should not be necessary to build the same library in the first
681       place.)
682    
683    3. I changed the way CM's tool mechanism determines the shell command
684       string used for things like ml-yacc etc. so that it does not break
685       when CM.Control.implicit_anchors is turned off.
686    
687    ----------------------------------------------------------------------
688    Name: Matthias Blume
689    Date: 2000/05/12 18:20:00 JST
690    Tag: blume-20000512-ml-build
691    Description:
692    
693    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
694    getting installed properly (by config/install.sh).
695    
696    ----------------------------------------------------------------------
697    Name: Matthias Blume
698    Date: 2000/05/12 17:30:00 JST
699    Tag: blume-20000512-anchors
700    Description:
701    
702    !!! NEW BOOT FILES !!!
703    
704    This change is in preparation of fading out support for "implicitly
705    anchored path names".  I went through all sources and used the
706    explicit (and relatively new) $-notation.  See system/README and the
707    CM manual for more info on this.
708    
709    I also modified the anchoring scheme for some things such as "smlnj",
710    "MLRISC", "cm", etc. to take advantage of the fact that explicit
711    anchors are more expressive: anchor name and first arc do not have to
712    coincide.  This entails the following user-visible change:
713    
714    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
715    particular, when you fire up sml with a command-line argument, say,
716    e.g.:
717    
718       sml '$smlnj/cmb.cm'
719    
720    At the ML toplevel prompt:
721    
722       CM.autoload "$smlnj/cmb.cm";
723    
724    There is also a new controller in CM.Control that can be used to turn
725    off all remaining support for implicit anchors by saying:
726    
727        CM.autoload "$smlnj/
728        #set CM.Control.implicit_anchors false;
729    
730    This causes CM to reject implicitly anchored paths.  This is (for the
731    time being) less permissive than the "final" version where there will
732    be no more such implicit anchors and relative paths will be just that:
733    relative.
734    
735    The next step (version after next version?) will be to make the
736    default for CM.Control.implicit_anchors false.  After the dust has
737    settled, I can then produce the "final" version of this...
738    
739    Note: Since bootstrapping is a bit tricky, I provided new boot files.
740    
741    ----------------------------------------------------------------------
742    Name: Matthias Blume
743    Date: 2000/05/11 16:30:00 JST
744    Tag: blume-20000511-sources
745    Description:
746    
747    The main change is that I added function CM.sources as a generalized
748    version of the earlier CM.makedepend.  This entails the following
749    additional changes:
750    
751      - CM.makedepend has been dropped.
752    
753      - CM manual has been updated.
754    
755      - TOOLS signature and API have been changed.
756    
757    ----------------------------------------------------------------------
758    Name: Allen Leung
759    Date: 2000/05/10 21:17 EDT
760    Tag: leunga-20000510-moby-c--ssa
761    Description:
762    
763      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
764    None of these affect SML/NJ.
765    
766    1. Register Allocation
767    
768        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
769           This module tries to remove local (i.e. basic block level) redundancies
770           during spilling.
771    
772        b. A new framework for performing region based register allocation.
773           Not yet entirely functional.
774    
775    2. X86
776    
777       a. DefUse for POP was missing the stack pointer [found by Lal]
778       b. Reload for CALL was incorrect in X86Spill [found by John]
779       c. Various fixes in X86Spill so that it can be used correctly for
780          the new spilling module.
781    
782    3. SSA/IR
783    
784       a. New module ir/dj-dataflow.sml implements elimination based
785          data flow analysis.
786    
787    4. MLRiscGen
788    
789       a. Fix for gc type annotation
790    
791    5. MDGen
792    
793       Various fixes for machine description -> ml code translation.  For ssa
794       only.
795    
796    ----------------------------------------------------------------------
797    Name: Allen Leung
798    Date: 2000/05/08 22:17 EDT
799    Tag: leunga-20000508-labexp
800    Description:
801    
802      Fermin has found a few assembly problems with constant expressions
803      generated in LabelExp.  Mostly, the problems involve extra parentheses,
804      which choke on dumb assemblers.  This is his fix.
805    
806    ----------------------------------------------------------------------
807    Name: Dave MacQueen
808    Date: 2000/04/09 14:00 EDT
809    Tag: dbm-20000502-Version_110_28
810    Description:
811    
812    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
813    
814    2. Updated config/version to 110.28
815    
816    3. Updated config/srcarchiveurl
817    
818    3. New boot files!
819       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
820    
821    ----------------------------------------------------------------------
822    Name: Matthias Blume
823    Date: 2000/05/01 19:05:00 JST
824    Tag: blume-20000501-noweb
825    Description:
826    
827    A new noweb tool has been added.  The existing system is entirely
828    unaffected by this, but some CM users have asked for renewed noweb
829    support.  Everything is documented in the CM manual.
830    
831    New (plugin) libraries:
832    
833       noweb-tool.cm
834       nw-ext.cm
835    
836    ----------------------------------------------------------------------
837    Name: Dave MacQueen
838    Date: 2000/04/30 12:40PM EDT
839    Tag: dbm-20000430-bug_fixes
840    Description:
841    
842    1. Fix for bug 1498
843       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
844       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
845         added toRealArray function
846       smlnj/src/compiler/MiscUtil/print/ppobj.sml
847         added check for tag Obj.RealArray to array printing case in ppObj
848    
849    2. Fix for bug 1510
850       smlnj/src/compiler/Semant/types/typesutil.sml
851         fixed definition of dummyargs (used by equalTycon) so that
852         dummy args are distinct types
853    
854    ----------------------------------------------------------------------
855    Name: Matthias Blume
856    Date: 2000/04/30 01:00:00 JST
857    Tag: blume-20000430-versions
858    Description:
859    
860    1. CM version numbering added.  This is an implementation of Lal's
861       proposal for adding version numbers and version checking to .cm
862       files.  Lal said that his proposal was just that -- a proposal.
863       For the time being I went ahead and implemented it so that people
864       can comment on it.  Everything is completely backward-compatible
865       (except for the stable library format, i.e., new bootfiles!).
866    
867       As usual, see the CM manual for details.
868    
869    2. An alternative syntax for anchored paths has been implemented.
870       Dave has recently voiced the same concerns that I had when I did
871       this, so there should be some support.  My take is that eventually
872       I will let support for the current syntax (where anchors are
873       "implicit") fade out in favor of the new, explicit syntax.
874       In order to be backward-compatible, both old and new syntax are
875       currently supported.
876    
877       Again, see the CM manual for details.
878    
879    3. Parallel make is trying to be slightly smarter:  When the master
880       process finds a "bottleneck", i.e., when there is only one
881       compilation unit that can be compiled and everybody else is
882       waiting on it, then it will simply compile it directly instead
883       of clumsily telling one of the slaves to do it.
884    
885    4. Support for "unsharing" added.  This is necessary in order to be
886       able to have two different versions of the same library running
887       at the same time (e.g., for trying out a new MLRISC while still
888       having the old MLRISC linked into the current compiler, etc.)
889       See the CM manual.
890    
891    5. Simple "makedepend" functionality added for generating Makefile
892       dependency information.  (This is rather crude at the moment.
893       Expect some changes here in the future.)
894    
895    6. ".fun" added as a recognized suffix for ML files. Also documented
896       explicitly in the manual that the fallback behavior (unknown suffix
897       -> ML file) is not an official feature!
898    
899    7. Small changes to the pickler for stable libraries.
900    
901    8. Several internal changes to CM (for cleanup/improvement).
902    
903    
904    !!!! NEW BINFILES !!!!
905    
906    ----------------------------------------------------------------------
907    Name: Matthias Blume
908    Date: 2000/04/28 17:30:00 JST
909    Tag: blume-20000428-pathconfig
910    Description:
911    
912    1. I changed config/install.sh to remove duplicate entries from the
913       lib/pathconfig file at the end.  Moreover, the final version of
914       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
915       in src/system/installml.
916    
917    2. The config/install.sh script now consistently uses relative
918       pathnames in lib/pathconfig whenever the anchor is in the lib
919       directory.  (So far this was true for the libraries that come
920       pre-compiled and bundled as part of the bootfiles but not for
921       libraries that are compiled by the script itself.)
922    
923    ----------------------------------------------------------------------
924    Name: Matthias Blume
925    Date: 2000/04/26 13:10:00 JST
926    Tag: blume-20000426-fun_suffix
927    Description:
928    
929    Added ".fun" as a recognized file name suffix (for ML code).
930    
931    ----------------------------------------------------------------------
932    Name: Allen Leung
933    Date: 2000/04/25 17:00:00 EST
934    Tag: leunga-20000425-alpha-ra
935    Description:
936    
937    1. Alpha
938    
939        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
940    in C--.
941    
942    2. RA
943    
944       Added a flag "ra-dump-size" to print out the size of the flowgraph
945       and the interference graph.
946    
947    ----------------------------------------------------------------------
948    Name: Dave MacQueen
949    Date: 2000/04/25/
950    Tag: dbm-20000425-mlyacc_doc_examples
951    Description:
952      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
953      Updated all three examples in src/ml-yacc/examples to run
954      under 110.* using CM.make.
955    
956    ----------------------------------------------------------------------
957    Name: Allen Leung
958    Date: 2000/04/20 23:04:00 EST
959    Tag: leunga-20000420-ssa-c---stuff
960    Description:
961    
962      This update synchronizes my repository with Yale's.  Most of these
963    changes, however, do not affect SML/NJ at all (the RA is an exception).
964    
965    1. Register Allocator
966    
967       a. An improvement in the interference graph construction:
968          Given a copy
969    
970                s <- t
971    
972          no interference edge between s and t is added for this definition of s.
973    
974       b. I've added two new spill heuristic modules that Fermin and I developed
975          (in the new library RA.cm). These are unused in SML/NJ but maybe
976          useful for others (Moby?)
977    
978    2. X86
979    
980       a. Various fixes in the backend provided by Fermin [C--] and Lal.
981    
982    3. Alpha
983    
984       a. Added the BSR instruction and code generation that goes with it [C--]
985       b. Other fixes too numerous to recount provided by Fermin [C--]
986    
987    4. Regmaps
988    
989       a. The regmaps are not initialized with the identity physical bindings
990          at creation time.  This is unneeded.
991    
992    5. MLRISC Optimizations
993    
994       a. The DJ-Graph module can now compute the iterated dominance frontiers
995          intersects with liveness incrementally in linear time! Woohoo!
996          This is now used in my new SSA construction algorithm.
997    
998       b. THe branch reorganization module is now smarter about linear chains of
999          basic blocks.
1000    
1001    
1002    ----------------------------------------------------------------------
1003    Name: Matthias Blume
1004    Date: 2000/04/12 13:52:00 JST
1005    Tag: blume_main_v110p27_1
1006    Description:
1007    
1008    Changed install.sh script to handle archive files without version number
1009    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1010    name of the boot file archive.
1011    
1012    ----------------------------------------------------------------------
1013    Name: Dave MacQueen
1014    Date: 2000/04/09 14:00 EDT
1015    Tag: dbm-20000410-Version_110_27
1016    Description:
1017    
1018    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1019    
1020    2. Updated src/config/version to 110.27
1021    
1022    3. New boot files!
1023    
1024    ----------------------------------------------------------------------
1025    Name: Allen Leung
1026    Date: 2000/04/09 19:09:00 EST
1027    Tag: leunga-20000409-misc
1028    Description:
1029    
1030    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1031    
1032    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1033    
1034    ----------------------------------------------------------------------
1035    Name: Stefan
1036    Date: 2000/04/07 10:00:00 EDT
1037    Tag: monnier-20000406-branch-handling
1038    Description:
1039    
1040    Improved handling of branches (mostly those generated from
1041    polymorphic equality), removed switchoff and changed the
1042    default optimization settings (more cpsopt and less flintopt).
1043    
1044    ----------------------------------------------------------------------
1045    Name: Allen Leung
1046    Date: 2000/04/06 01:30:00 EST
1047    Tag: leunga-20000406-peephole-x86-SSA-2
1048    Description:
1049    
1050       Forgot a few files.
1051    
1052    ----------------------------------------------------------------------
1053    Name: Allen Leung
1054    Date: 2000/04/06 00:36:00 EST
1055    Tag: leunga-20000406-peephole-x86-SSA
1056    Description:
1057    
1058    1.  New Peephole code
1059    
1060    2.  Minor improvement to X86 instruction selection
1061    
1062    3.  Various fixes to SSA and machine description -> code translator
1063    
1064    ----------------------------------------------------------------------
1065    Name: Matthias Blume
1066    Date: 2000/04/05 12:30:00 JST
1067    Tag: blume_main_v110p26p2_3
1068    Description:
1069    
1070    This update just merges three minor cosmetic updates to CM's sources
1071    to get ready for the 110.27 code freeze on Friday.  No functionality
1072    has changed.
1073    
1074    ----------------------------------------------------------------------
1075    Name: Allen Leung
1076    Date: 2000/04/04 19:39:00 EST
1077    Tag: leunga-20000404-x86-asm
1078    Description:
1079    
1080    1.  Fixed a problem in X86 assembly.
1081    
1082        Things like
1083    
1084           jmp %eax
1085           jmp (%eax)
1086    
1087        should be output as
1088    
1089           jmp *%eax
1090           jmp *(%eax)
1091    
1092    2.  Assembly output
1093    
1094          Added a new flag
1095    
1096              "asm-indent-copies" (default to false)
1097    
1098          When this flag is on, parallel copies will be indented an extra level.
1099    
1100    ----------------------------------------------------------------------
1101    Name: Allen Leung
1102    Date: 2000/04/04 03:18:00 EST
1103    Tag: leunga-20000404-C--Moby
1104    Description:
1105    
1106        All of these fixes are related to C--, Moby, and my own optimization
1107        stuff; so they shouldn't affect SML/NJ.
1108    
1109    1.  X86
1110    
1111        Various fixes related floating point, and extensions.
1112    
1113    2.  Alpha
1114    
1115        Some extra patterns related to loads with signed/zero extension
1116        provided by Fermin.
1117    
1118    3.  Assembly
1119    
1120        When generating assembly, resolve the value of client defined constants,
1121        instead of generating symbolic values.  This is controlled by the
1122        new flag "asm-resolve-constants", which is default to true.
1123    
1124    4.  Machine Descriptions
1125    
1126        a. The precedence parser was slightly broken when parsing infixr symbols.
1127        b. The type generalizing code had the bound variables reversed, resulting
1128           in a problem during arity raising.
1129        c. Various fixes in machine descriptions.
1130    
1131    ----------------------------------------------------------------------
1132    Name: Matthias Blume
1133    Date: 2000/04/03 16:05:00 JST
1134    Tag: blume_main_v110p26p2_2
1135    Description:
1136    
1137    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
1138    now done via the ordinary static environment that is context to each
1139    compilation unit.
1140    
1141    To this end, I arranged that instead of "structure Core" as "structure
1142    _Core" is bound in the pervasive environment.  Core access is done via
1143    _Core (which can never be accidentally rebound because _Core is not a
1144    legal surface-syntax symbol).
1145    
1146    The current solution is much cleaner because the core environment is
1147    now simply part of the pervasive environment which is part of every
1148    compilation unit's context anyway.  In particular, this eliminates all
1149    special-case handling that was necessary until now in order to deal
1150    with dynamic and symbolic parts of the core environment.
1151    
1152    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1153    compilation manager's bootstrap compiler (actually: in the "init group"
1154    handling).  See the comments in src/system/smlnj/init/init.cmi for
1155    more details.
1156    
1157    I also tried to track down all mentions of "Core" (as string argument
1158    to Symbol.strSymbol) in the compiler and replaced them with a
1159    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
1160    name appears in one place only.
1161    
1162    Binfile and bootfile format have not changed, but the switchover from
1163    the old "init.cmi" to the new one is a bit tricky, so I supplied new
1164    bootfiles anyway.
1165    
1166    ----------------------------------------------------------------------
1167    Name: Allen Leung
1168    Date: 2000/04/02 21:17:00 EST
1169    Tag: leunga-20000402-mltree
1170    Description:
1171    
1172       1. Renamed the constructor CALL in MLTREE by popular demand.
1173       2. Added a bunch of files from my repository.  These are currently
1174          used by other non-SMLNJ backends.
1175    
1176    ----------------------------------------------------------------------
1177    Name: Allen Leung
1178    Date: 2000/03/31 21:15:00 EST
1179    Tag: leunga-20000331-aliasing
1180    Description:
1181    
1182    This update contains a rewritten (and hopefully more correct) module
1183    for extracting aliasing information from CPS.
1184    
1185       To turn on this feature:
1186    
1187            Compiler.Control.CG.memDisambiguate := true
1188    
1189       To pretty print the region information with assembly
1190    
1191           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
1192    
1193       To control how many levels of aliasing information are printed, use:
1194    
1195           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
1196    
1197       The default of n is 3.
1198    
1199    ----------------------------------------------------------------------
1200    Name: David MacQueen
1201    Date: 2000/03/31 11:15:00 EST
1202    Tag: dbm-20000331-runtime_fix
1203    Description:
1204    
1205    This update contains:
1206    
1207    1. runtime/c-lib/c-libraries.c
1208       includes added in revision 1.2 caused compilation errors on hppa-hpux
1209    
1210    2. fix for bug 1556
1211       system/Basis/Implementation/NJ/internal-signals.sml
1212    
1213    ----------------------------------------------------------------------
1214    Name: Matthias Blume
1215    Date: 2000/03/31 18:00:00 JST
1216    Tag: blume_main_v110p26p2_1
1217    Description:
1218    
1219    This update contains:
1220    
1221    1. A small change to CM's handling of stable libraries:
1222       CM now maintains one "global" modmap that is used for all stable
1223       libraries.  The use of such a global modmap maximizes sharing and
1224       minimizes the need for re-traversing parts of environments during
1225       modmap construction.  (However, this has minor impact since modmap
1226       construction seems to account for just one percent or less of total
1227       compile time.)
1228    
1229    2. I added a "genmap" phase to the statistics.  This is where I got the
1230       "one percent" number (see above).
1231    
1232    3. CM's new tool parameter mechanism just became _even_ better. :)
1233       - The parser understands named parameters and recursive options.
1234       - The "make" and "shell" tools use these new features.
1235         (This makes it a lot easier to cascade these tools.)
1236       - There is a small syntax change: named parameters use a
1237    
1238           <name> : ( <option> ... )            or
1239           <name> : <string>
1240    
1241         syntax.  Previously, named parameters were implemented in an
1242         ad-hoc fashion by each tool individually (by parsing strings)
1243         and had the form
1244    
1245           <name>=<string>
1246    
1247       See the CM manual for a full description of these issues.
1248    
1249    ----------------------------------------------------------------------
1250    Name: Matthias Blume
1251    Date: 2000/03/30 18:00:00 JST
1252    Tag: blume_main_v110p26p2_0
1253    Description:
1254    
1255    !!!!! WARNING !!!!!!
1256    !!  New binfiles  !!
1257    !!!!!!!!!!!!!!!!!!!!
1258    
1259    This update contains:
1260    
1261    1. Moderate changes to CM:
1262    
1263       - Changes to CM's tools mechanism.  In particular, it is now possible
1264       to have tools that accept additional "command line" parameters
1265       (specified in the .cm file at each instance where the tool's class is
1266       used).
1267    
1268       This was done to accommodate the new "make" and "shell" tools which
1269       facilitate fairly seamless hookup to portions of code managed using
1270       Makefiles or Shell scripts.
1271    
1272       There are no classes "shared" or "private" anymore.  Instead, the
1273       sharing annotation is now a parameter to the "sml" class.
1274    
1275       There is a bit of generic machinery for implementing one's own
1276       tools that accept command-line parameters.  However, I am not yet fully
1277       satisfied with that part, so expect changes here in the future.
1278    
1279       All existing tools are described in the CM manual.
1280    
1281       - Slightly better error handling.  (CM now suppresses many followup
1282       error messages that tended to be more annoying than helpful.)
1283    
1284    2. Major changes to the compiler's static environment data structures.
1285    
1286       - no CMStaticEnv anymore.
1287            - no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
1288              but it is called Environment), no conversions between different
1289              kinds of static environments
1290    
1291       - There is still a notion of a "modmap", but such modmaps are generated
1292         on demand at the time when they are needed.  This sounds slow, but I
1293         sped up the code that generates modmaps enough for this not to lead to
1294         a slowdown of the compiler (at least I didn't detect any).
1295    
1296       - To facilitate rapid modmap generation, static environments now
1297         contain an (optional) "modtree" structure.  Modtree annotations are
1298         constructed by the unpickler during unpickling.  (This means that
1299         the elaborator does not have to worry about modtrees at all.)
1300         Modtrees have the advantage that they are compositional in the same
1301         way as the environment data structure itself is compositional.
1302         As a result, modtrees never hang on to parts of an environment that
1303         has already been rendered "stale" by filtering or rebinding.
1304    
1305       - I went through many, many trials and errors before arriving at the
1306         current solution.  (The initial idea of "linkpaths" did not work.)
1307         But the result of all this is that I have touched a lot of files that
1308         depend on the "modules" and "types" data structures (most of the
1309         elaborator). There were a lot of changes during my "linkpath" trials
1310         that could have been reverted to their original state but weren't.
1311         Please, don't be too harsh on me for messing with this code a bit more
1312         than what was strictly necessary...  (I _did_ resist the tempation
1313         of doing any "global reformatting" to avoid an untimely death at
1314         Dave's hands. :)
1315    
1316       - One positive aspect of the previous point:  At least I made sure that
1317         all files that I touched now compile without warnings (other than
1318         "polyEqual").
1319    
1320       - compiler now tends to run "leaner" (i.e., ties up less memory in
1321         redundant modmaps)
1322    
1323  ----------------------------------------------------------------------  ----------------------------------------------------------------------
1324  Name: Allen Leung  Name: Allen Leung
1325  Date: 2000/03/29 18:00:00  Date: 2000/03/29 18:00:00
1326  Tag: leunga-20000327-mlriscGen_hppa_alpha_x86  Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
1327    Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
1328  Description:  Description:
1329    
1330     This update contains *MAJOR* changes to the way code is generated from CPS     This update contains *MAJOR* changes to the way code is generated from CPS
# Line 103  Line 1412 
1412    
1413       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
1414       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
1415       testl is perferable to andl.       testl is preferable to andl.
1416    
1417  7. RA (x86 only)  7. RA (x86 only)
1418    
# Line 285  Line 1594 
1594    
1595  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
1596  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
1597  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
1598  explanation.  explanation.
1599    
1600  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
1601  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
1602  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
1603  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 332  Line 1641 
1641  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
1642  <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,
1643  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
1644  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
1645  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
1646  and, thus, saves a lot of time.  and, thus, saves a lot of time.
1647    
# Line 382  Line 1691 
1691    it from that remote directory.    it from that remote directory.
1692    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
1693    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
1694    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
1695    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
1696    
1697    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.585  
changed lines
  Added in v.682

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