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 629, Wed Apr 26 04:06:41 2000 UTC revision 679, Thu Jun 29 07:03:20 2000 UTC
# Line 12  Line 12 
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14  ----------------------------------------------------------------------  ----------------------------------------------------------------------
15    Name: Matthias Blumen
16    Date: 2000/06/29 16:04:00 JST
17    Tag: blume-20000629-yacctool
18    Description:
19    
20    Changes:
21    
22    1. Class "mlyacc" now takes separate arguments to pass options to
23       generated .sml- and .sig-files independently.
24    2. Corresponding CM manual updates.
25    3. BTrace module now also reports call sites.  (However, for loop clusters
26       it only shows from where the cluster was entered.)  There are associated
27       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
28    
29    ----------------------------------------------------------------------
30    Name: Matthias Blume
31    Date: 2000/06/27 16:51:00 JST
32    Tag: blume-20000627-noweb
33    Description:
34    
35    Changes:
36    
37     1. Implemented "subdir" and "witness" options for noweb tool.
38        This caused some slight internal changes in CM's tool implementation.
39     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
40        some remaining issues from earlier path anchor changes.
41     3. Updated CM manual accordingly.
42    
43     4. Changed implementation of back-tracing so that I now consider it
44        ready for prime-time.
45    
46        In particular, you don't have to explicitly trigger the back-trace
47        anymore.  Instead, if you are running BTrace-instrumented code and
48        there is an uncaught exception (regardless of whether or not it was
49        raised in instrumented code), the top-level evalloop will print
50        the back-trace.
51    
52        Features:
53    
54          - Instrumented and uninstrumented code work together seemlessly.
55            (Of course, uninstrumented code is never mentioned in actual
56             back-traces.)
57    
58          - Asymptotic time- and space-complexity of instrumented code is
59            equal to that of uninstrumented code.  (This means that
60            tail-recursion is preserved by the instrumentation phase.)
61    
62          - Modules whose code has been instrumented in different sessions
63            work together without problem.
64    
65          - There is no penalty whatsoever on uninstrumented code.
66    
67          - There is no penalty on "raise" expressions, even in
68            instrumented code.
69    
70        A potential bug (or perhaps it is a feature, too):
71    
72          A back-trace reaches no further than the outermost instrumented
73          non-trivial "raise".  Here, a "trivial" raise is one that is the
74          sole RHS of a "handle" rule.  Thus, back-traces reach trough
75    
76               <exp> handle e => raise e
77    
78          and even
79    
80               <exp> handle Foo => raise Bar
81    
82          and, of course, through
83    
84               <exp> handle Foo => ...
85    
86         if the exception was not Foo.
87    
88         Back-traces always reach right through any un-instrumented code
89         including any of its "handle" expressions, trivial or not.
90    
91       To try this out, do the following:
92    
93         - Erase all existing binfiles for your program.
94           (You may keep binfiles for those modules where you think you
95            definitely don't need back-tracing.)
96         - Turn on back-trace instrumentation:
97              SMLofNJ.Internals.BTrace.mode (SOME true);
98         - Recompile your program.  (I.e., run "CM.make" or "use".)
99         - You may now turn instrumentation off again (if you want):
100              SMLofNJ.Internals.BTrace.mode (SOME false);
101         - Run your program as usual.  If it raises an exception that
102           reaches the interactive toplevel, then a back-trace will
103           automatically be printed.  After that, the toplevel loop
104           will print the exception history as usual.
105    
106    ----------------------------------------------------------------------
107    Name: Matthias Blume
108    Date: 2000/06/26 09:56:46 JST
109    Tag: blume-20000626-setup
110    Description:
111    
112    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
113          ML code before and after compiling a file (e.g., to set compiler
114          flags)
115    
116    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
117              - associated changes to btrace.sml (BTrace instrumentation pass)
118              - cleaner implementation of btimp.sml (BTrace tracing and report
119                module)
120    
121    CM manual: * new path encoding documented
122               * description of setup-parameter to "sml" added
123    
124    The biggest user-visible change to back-tracing is that it is no
125    longer necessary to compile all traced modules within the same
126    session.  (This was a real limitation.)
127    
128    ----------------------------------------------------------------------
129    Name: Matthias Blume
130    Date: 2000/06/24 12:40:00 JST
131    Tag: blume-20000624-startup
132    Description:
133    
134    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
135    bit too often -- to put it mildly. :)
136    
137    ----------------------------------------------------------------------
138    Name: Matthias Blume
139    Date: 2000/06/23 18:20:00 JST
140    Tag: blume-20000623-btrace
141    Description:
142    
143    This updates adds a backtrace facility to aid programmers in debugging
144    their programs.  This involves the following changes:
145    
146    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
147       keeping track of the current call stack.  When programs are compiled
148       in a special mode, the compiler will insert calls to these hooks
149       into the user program.
150       "Hook" means that it is possible for different implementations of
151       back-tracing to register themselves (at different times).
152    
153    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
154       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
155       It can be turned on using this call:
156         SMLofNJ.Internals.BTrace.mode (SOME true);
157       Turning it off again:
158         SMLofNJ.Internals.BTrace.mode (SOME false);
159       Querying the current status:
160         SMLofNJ.Internals.BTrace.mode NONE;
161       Annotated programs are about twice as big as normal ones, and they
162       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
163       where all hooks do nothing).  The slowdown with a plugin that is
164       actually useful (such as the one supplied by default) is even greater,
165       but in the case of the default plugin it is still only an constant
166       factor (amortized).
167    
168    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
169       with a sub-structure BTrace for controlling back-tracing.  In particular,
170       the above-mentioned function "mode" controls whether the annotation
171       phase is invoked by the compiler.  Another important function is
172       "trigger": when called it aborts the current execution and causes
173       the top-level loop to print a full back-trace.
174    
175    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
176       for back-tracing.  It keeps track of the dynamic call stack and in
177       addition to that it keeps a partial history at each "level" of that
178       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
179       d and b (at separate times, dynamically), then the report will show:
180    
181       GOTO   d
182             /c
183       GOTO  \b
184       CALL   a
185    
186       This shows that there was an initial non-tail call of a, then a
187       tail-call to b or c, looping behavior in a cluster of functions that
188       consist of b and c, and then a goto from that cluster (i.e., either from
189       b or from c) to d.
190    
191       Note that (depending on the user program) the amount of information
192       that the back-trace module has to keep track of at each level is bounded
193       by a constant.  Thus, the whole implementation has the same asymptotical
194       complexity as the original program (both in space and in time).
195    
196    5. compiler/TopLevel/interact/evalloop.sml has been modified to
197       handle the special exception SMLofNJ.Internals.BTrace.BTrace
198       which is raised by the "trigger" function mentioned above.
199    
200    Notes on usage:
201    
202    - Annotated code works well together with unannotated code:
203    Unannotated calls simply do not show up at all in the backtrace.
204    
205    - It is not a good idea to let modules that were annotated during
206    different sessions run at the same time.  This is because the compiler
207    chooses small integers to identify individual functions, and there
208    will be clashes if different modules were compiled in separate sessions.
209    (Nothing will crash, and you will even be told about the clashes, but
210    back-trace information will in general not be useful.)
211    
212    - Back-tracing can be confused by callcc and capture.
213    
214    - The only way of getting a back-trace right now is to explicitly
215    invoke the "trigger" function from your user program.  Eventually, we
216    should make every exception carry back-trace information (if
217    available).  But since this creates more overhead at "raise"-time
218    (similar to the current exnHistory overhead), I have not yet
219    implemented this.  (The implementation will be rather easy.)  With
220    exceptions carrying back-trace information, this facility will be even
221    more useful because users don't need to modify their programs...
222    
223    - While it is possible to compile the compiler with back-trace
224    annotations turned on (I did it to get some confidence in
225    correctness), you must make absolutely sure that core.sml and
226    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
227    be compiled with annotation because there is no core access yet, but
228    if you compile btimp.sml with annotation, then the system will go into
229    an infinite recursion and crash.)
230    Since CM currently does not know about BTrace, the only way to turn
231    annotations on and off for different modules of the compiler is to
232    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
233    this is awkward and clumsy.
234    
235    Sample sessions:
236    
237    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
238    - SMLofNJ.Internals.BTrace.mode (SOME true);
239    [autoloading]
240    [autoloading done]
241    val it = false : bool
242    - structure X = struct
243    -     fun main n = let
244    -         fun a (x, 0) = d x
245    -           | a (x, n) = b (x, n - 1)
246    -         and b (x, n) = c (x, n)
247    -         and c (x, n) = a (x, n)
248    -         and d x = e (x, 3)
249    -         and e (x, 0) = f x
250    -           | e (x, n) = e (x, n - 1)
251    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
252    -           | f n = n * g (n - 1)
253    -         and g n = a (n, 3)
254    -     in
255    -         f n
256    -     end
257    - end;
258    structure X : sig val main : int -> int end
259    - X.main 3;
260    *** BACK-TRACE ***
261    GOTO   stdIn:4.2-13.20: X.main[2].f
262    GOTO-( stdIn:4.2-13.20: X.main[2].e
263    GOTO   stdIn:4.2-13.20: X.main[2].d
264         / stdIn:4.2-13.20: X.main[2].a
265         | stdIn:4.2-13.20: X.main[2].b
266    GOTO-\ stdIn:4.2-13.20: X.main[2].c
267    CALL   stdIn:4.2-13.20: X.main[2].g
268    GOTO   stdIn:4.2-13.20: X.main[2].f
269    GOTO-( stdIn:4.2-13.20: X.main[2].e
270    GOTO   stdIn:4.2-13.20: X.main[2].d
271         / stdIn:4.2-13.20: X.main[2].a
272         | stdIn:4.2-13.20: X.main[2].b
273    GOTO-\ stdIn:4.2-13.20: X.main[2].c
274    CALL   stdIn:4.2-13.20: X.main[2].g
275    GOTO   stdIn:4.2-13.20: X.main[2].f
276    GOTO-( stdIn:4.2-13.20: X.main[2].e
277    GOTO   stdIn:4.2-13.20: X.main[2].d
278         / stdIn:4.2-13.20: X.main[2].a
279         | stdIn:4.2-13.20: X.main[2].b
280    GOTO-\ stdIn:4.2-13.20: X.main[2].c
281    CALL   stdIn:4.2-13.20: X.main[2].g
282    GOTO   stdIn:4.2-13.20: X.main[2].f
283    CALL   stdIn:2.15-17.4: X.main[2]
284    -
285    
286    (Note that because of a FLINt bug the above code currently does not
287    compile without BTrace turned on.)
288    
289    Here is another example, using my modified Tiger compiler:
290    
291    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
292    - SMLofNJ.Internals.BTrace.mode (SOME true);
293    [autoloading]
294    [autoloading done]
295    val it = false : bool
296    - CM.make "sources.cm";
297    [autoloading]
298    ...
299    [autoloading done]
300    [scanning sources.cm]
301    [parsing (sources.cm):parse.sml]
302    [creating directory CM/SKEL ...]
303    [parsing (sources.cm):tiger.lex.sml]
304    ...
305    [wrote CM/sparc-unix/semant.sml]
306    [compiling (sources.cm):main.sml]
307    [wrote CM/sparc-unix/main.sml]
308    [New bindings added.]
309    val it = true : bool
310    - Main.compile ("../testcases/merge.tig", "foo.out");
311    *** BACK-TRACE ***
312    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
313    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
314    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
315    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
316    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
317    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
318    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
319         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
320    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
321    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
322    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
323    CALL   main.sml:1.18-118.4: Main.compile[2]
324    -
325    
326    ----------------------------------------------------------------------
327    Name: Matthias Blumen
328    Date: 2000/06/21 18:00:00 JST
329    Tag: blume-20000621-manual
330    Description:
331    
332    CM manual update: Path environments documented.
333    
334    ----------------------------------------------------------------------
335    Name: Matthias Blume
336    Date: 2000/06/19 13:40:00
337    Tag: blume-20000619-manual
338    Description:
339    
340    CM manual and system/README update.  This only covers the fact that
341    there are no more implicit anchors.  (Path environments and the "bind"
342    option to "cm" have yet to be documented.)
343    
344    ----------------------------------------------------------------------
345    Name: Matthias Blume
346    Date: 2000/06/19 11:05:00 JST
347    Tag: blume-20000619-chdir-bugfix
348    Description:
349    
350    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
351    
352    ----------------------------------------------------------------------
353    Name: Matthias Blume
354    Date: 2000/06/18 22:00:10 JST
355    Tag: blume-20000618-implicit-anchors-really-gone
356    Description:
357    
358    I updates the previous HISTORY entry where I forgot to mention that
359    implicit anchors are no longer with us.
360    
361    The current update also gets rid of the (now useless) controller
362    CM.Control.implicit_anchors.
363    
364    ----------------------------------------------------------------------
365    Name: Matthias Blume
366    Date: 2000/06/16 17:30:00 JST
367    Tag: blume-20000616-anchorenv
368    Description:
369    
370    This patch implements the long anticipated (just kidding :) "anchor
371    environment" mechanism.  In the course of doing this, I also
372    re-implemented CM's internal "SrcPath" module from scratch.  The new
373    one should be more robust in certain boundary cases.  In any case, it
374    is a lot cleaner than its predecessor (IMHO).
375    
376    This time, although there is yet another boot file format change, I
377    kept the unpickler backward-compatible.  As a result, no new bootfiles
378    are necessary and bootstrapping is straightforward.  (You cannot read
379    new bootfiles into an old system, but the other way around is no
380    problem.)
381    
382    Visible changes:
383    
384    ** 0. Implicit path anchors (without the leading $-symbol) are no
385    longer recognized at all. This means that such path names are not
386    illegal either.  For example, the name basis.cm simply refers to a
387    local file called "basis.cm" (i.e, the name is an ordinary path
388    relative to .cm-files directory).  Or, to put it differently, only
389    names that start with $ are anchored paths.
390    
391    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
392    vanished.
393    
394    John (Reppy) had critizised this as soon as I originally proposed and
395    implemented it, but at that time I did not really deeply believe
396    him. :) Now I came full-circle because I need the $<singlearc> syntax
397    in another place where it cannot be seen as an abbreviation for
398    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
399    seems to mean (i.e., it "expands" into the corresponding anchor
400    value).
401    
402    However, when paths are used as members in CM description files, it
403    continues to be true that there must be at least another arc after the
404    anchor.  This is now enforced separately during semantic analysis
405    (i.e., from a lexical/syntactical point of view, the notation is ok.)
406    
407    ** 2. The "cm" class now accepts an option "bind".  The option's value
408    is a sub-option list of precisely two items -- one labeled "anchor"
409    and the other one labeled "value".  As you might expect, "anchor" is
410    used to specify an anchor name to be bound, and "value" specifies what
411    the anchor is being bound to.
412    
413    The value must be a directory name and can be given in either standard
414    syntax (including the possibility that it is itself an anchored path)
415    or native syntax.
416    
417    Examples:
418    
419       foo.cm (bind:(anchor:bar value:$mystuff/bar))
420       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
421    
422    and so on.
423    
424    The meaning of this is that the .cm-file will be processed with an
425    augmented anchor environment where the given anchor(s) is/are bound to
426    the given values(s).
427    
428    The rationale for having this feature is this: Suppose you are trying
429    to use two different (already stable) libraries a.cm and b.cm (that
430    you perhaps didn't write yourself).  Further, suppose each of these
431    two libraries internally uses its own auxiliary library $aux/lib.cm.
432    Normally you would now have a problem because the anchor "lib" can not
433    be bound to more than one value globally.  Therefore, the project that
434    uses both a.cm and b.cm must locally redirect the anchor to some other
435    place:
436    
437       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
438       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
439    
440    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
441    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
442    
443    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
444    you when you do so at the time of CM.stabilize).  Therefore, you can
445    also use an anchored path as the value:
446    
447      a.cm (bind:(anchor:lib value:$a-lib))
448      b.cm (bind:(anchor:lib value:$b-lib))
449    
450    Now you can globally configure (using the usual CM.Anchor.anchor or
451    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
452    itself is always locally bound, setting it globally is no longer
453    meaningful or necessary (but it does not hurt either).  In fact, "lib"
454    can still be used as a global anchor for separate purposes.  As a
455    matter of fact, one can locally define "lib" in terms of a global
456    "lib":
457    
458      a.cm (bind:(anchor:lib value:$lib/a))
459      b.cm (bind:(anchor:lib value:$lib/b))
460    
461    ** 3: The encoding of path names has changed.  This affects the way
462    path names are shown in CM's progress report and also the internal
463    protocol encoding used for parallel make.
464    
465    The encoding now uses one or more ':'-separated segments.  Each
466    segments corresponds to a file that has been specified relative to the
467    file given by its preceding segment.  The first segment is either
468    relative to the CWD, absolute, or anchored.  Each segment itself is
469    basically a Unix pathname; all segments but the first are relative.
470    
471    Example:
472    
473       $foo/bar/baz.cm:a/b/c.sml
474    
475    This path denotes the file bar/a/b/c.sml relative to the directory
476    denoted by anchor "foo".  Notice that the encoding also includes
477    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
478    paths are resolved relative to the .cm-files directory, so baz.cm must
479    be ignored to get the "real" pathname.
480    
481    To make this fact more obvious, CM puts the names of such "virtual
482    arcs" into parentheses when they appear in progress reports. (No
483    parentheses will appear in the internal protocol encoding.)  Thus,
484    what you really see is:
485    
486      $foo/bar/(baz.cm):a/b/c.sml
487    
488    I find this notation to be much more informative than before.
489    
490    Another new feature of the encoding is that special characters
491    including parentheses, colons, (back)slashes, and white space are
492    written as \ddd (where ddd is the decimal encoding of the character).
493    
494    *** The CM manual still needs to be updated.
495    
496    ----------------------------------------------------------------------
497    Name: Allen Leung
498    Date: 2000/06/15 00:38:00
499    Tag: leunga-20000615-x86-peephole
500    
501    x86 Peephole fix by Fermin.  Affects c-- and moby only.
502    
503    ----------------------------------------------------------------------
504    Name: Matthias Blume
505    Date: 2000/06/12 11:40:00
506    Tag: blume-20000612-parmakefix
507    Description:
508    
509    More cleanup after changing the file naming scheme: This time I
510    repaired the parallel make mechanism for CMB.make which I broke earlier.
511    
512    ----------------------------------------------------------------------
513    Name: Allen Leung
514    Date: 2000/06/09 01:25:00
515    Tag: leunga-20000609-various
516    
517    None of these things should affect normal SML/NJ operations
518    
519    1. Peephole improvements provided by Fermin (c--)
520    2. New annotation DEFUSE for adding extra dependence (moby)
521    3. New X86 LOCK instructions (moby)
522    4. New machine description language for reservation tables (scheduling)
523    5. Fixes to various optimization/analysis modules (branch chaining, dominator
524       trees etc.)
525    6. I've changed the CM files so that they can work with versions
526       110.0.6, 110.25 and 110.28
527    
528    ----------------------------------------------------------------------
529    Name: Matthias Blume
530    Date: 2000/06/09 12:40:00
531    Tag: blume-20000609-log
532    Description:
533    
534    - Removed all(?) remaining RCS Log entries from sources.
535    
536    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
537      anchored paths).
538    
539    ----------------------------------------------------------------------
540    Name: Matthias Blume
541    Date: 2000/06/07 17:00:00 JST
542    Tag: blume-20000607-no-implicit-anchors
543    Description:
544    
545    1. This update changes the default setting for
546    CM.Control.implicit_anchors from true to false.  This means that
547    implicit anchors are no longer permitted by default.  I also tried to
548    make sure that nothing else still relies on implicit anchors.
549    (This is the next step on the schedule towards a CM that does not even
550    have the notion of implicit anchors anymore.)
551    
552    2. More CM manual updates.
553    
554    3. I managed to track down and fix the pickling bug I mentioned last
555    time.  Because of the previously existing workaround, this entails no
556    immediate practical changes.
557    
558    ----------------------------------------------------------------------
559    Name: Matthias Blume
560    Date: 2000/06/06 11:15:00 JST
561    Tag: blume-20000606-lazierpickle
562    Description:
563    
564    !!!! NEW BOOT FILES !!!!
565    
566    * The main purpose of this update is to make library pickles lazier in
567    order to reduce the initial space penalty for autoloading a library.
568    As a result, it is now possible to have $smlnj/compiler.cm
569    pre-registered.  This should take care of the many complaints or
570    inquiries about missing structure Compiler.  This required changes to
571    CM's internal data structures and small tweaks to some algorithms.
572    
573    As a neat additional effect, it is no longer necessary (for the sake
574    of lean heap image files) to distinguish between a "minimal" CM and a
575    "full" CM.  Now, there is only one CM (i.e., the "full" version:
576    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
577    interactive top level. ($smlnj/cm/minimal.cm is gone.)
578    
579    To make the life of compiler-hackers easier, "makeml" now also
580    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
581    words, after you bootstrap a new sml for the first time, you will not
582    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
583    around you will still have to do it, though.)
584    
585    * A second change consists of major updates to the CM manual.  There
586    are now several appendices with summary information and also a full
587    specification of the CM description file syntax.
588    
589    * In directory src/system I added the script "allcross".  This script
590    invokes sml and cross-compiles the compiler for all supported
591    architectures.  (Useful when providing a new set of boot files.)
592    
593    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
594    added a small tweak to pickle-util.sml to work around this problem,
595    but it is not a proper fix yet.  I will investigate further.  (The
596    effect of the bug was an inflation of library pickle size.)
597    
598    * Version number increased to 110.28.1 (to avoid compatibility problems).
599    
600    ----------------------------------------------------------------------
601    Name: Allen Leung
602    Date: 2000/05/25 17:28 EDT
603    Tag: leunga-20000525-ra
604    Description:
605    
606      Fixed a bug in freezing phase of the register allocator.
607    
608    ----------------------------------------------------------------------
609    Name: Allen Leung
610    Date: 2000/05/15 22:53 EDT
611    Tag: leunga-20000515-alpha-x86-ra
612    Description:
613    
614      1. Alpha
615    
616          Slight cleanup.  Removed the instruction SGNXL
617    
618      2. X86
619    
620          Added the following instructions to the instruction set:
621    
622            ROLx, RORx,
623            BTx, BTSx, BTLx, BTRx,
624            XCHGx, and variants with the LOCK prefix
625    
626      3. Register Allocation
627    
628          The module ra-rewrite-with-renaming has been improved.
629    
630      These have no effect on SML/NJ.
631    
632    ----------------------------------------------------------------------
633    Name: Matthias Blume
634    Date: 2000/05/15 16:20:00 JST
635    Tag: blume-20000515-lightrebuild
636    Description:
637    
638    1. I added an alternative to "-rebuild" to "makeml".  The difference is
639       that prior to calling CMB.make' the CM-variable "LIGHT" will be
640       defined.  In effect, the command will not build any cross-compiler
641       backends and therefore finish more quickly.
642    
643       The "fixpt" script also takes a "-light" switch to be able to use
644       this new facility while compiling for a fixpoint.
645    
646    2. I replaced all mentions of anchored paths in group owner specifications
647       with simple relative paths (usually starting with "..").
648       The rationale is that a library's internal workings should not be
649       compromised by the lack of some anchor.  (An anchor is necessary
650       for someone who wants to refer to the library by an anchored path,
651       but it should not be necessary to build the same library in the first
652       place.)
653    
654    3. I changed the way CM's tool mechanism determines the shell command
655       string used for things like ml-yacc etc. so that it does not break
656       when CM.Control.implicit_anchors is turned off.
657    
658    ----------------------------------------------------------------------
659    Name: Matthias Blume
660    Date: 2000/05/12 18:20:00 JST
661    Tag: blume-20000512-ml-build
662    Description:
663    
664    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
665    getting installed properly (by config/install.sh).
666    
667    ----------------------------------------------------------------------
668    Name: Matthias Blume
669    Date: 2000/05/12 17:30:00 JST
670    Tag: blume-20000512-anchors
671    Description:
672    
673    !!! NEW BOOT FILES !!!
674    
675    This change is in preparation of fading out support for "implicitly
676    anchored path names".  I went through all sources and used the
677    explicit (and relatively new) $-notation.  See system/README and the
678    CM manual for more info on this.
679    
680    I also modified the anchoring scheme for some things such as "smlnj",
681    "MLRISC", "cm", etc. to take advantage of the fact that explicit
682    anchors are more expressive: anchor name and first arc do not have to
683    coincide.  This entails the following user-visible change:
684    
685    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
686    particular, when you fire up sml with a command-line argument, say,
687    e.g.:
688    
689       sml '$smlnj/cmb.cm'
690    
691    At the ML toplevel prompt:
692    
693       CM.autoload "$smlnj/cmb.cm";
694    
695    There is also a new controller in CM.Control that can be used to turn
696    off all remaining support for implicit anchors by saying:
697    
698        CM.autoload "$smlnj/
699        #set CM.Control.implicit_anchors false;
700    
701    This causes CM to reject implicitly anchored paths.  This is (for the
702    time being) less permissive than the "final" version where there will
703    be no more such implicit anchors and relative paths will be just that:
704    relative.
705    
706    The next step (version after next version?) will be to make the
707    default for CM.Control.implicit_anchors false.  After the dust has
708    settled, I can then produce the "final" version of this...
709    
710    Note: Since bootstrapping is a bit tricky, I provided new boot files.
711    
712    ----------------------------------------------------------------------
713    Name: Matthias Blume
714    Date: 2000/05/11 16:30:00 JST
715    Tag: blume-20000511-sources
716    Description:
717    
718    The main change is that I added function CM.sources as a generalized
719    version of the earlier CM.makedepend.  This entails the following
720    additional changes:
721    
722      - CM.makedepend has been dropped.
723    
724      - CM manual has been updated.
725    
726      - TOOLS signature and API have been changed.
727    
728    ----------------------------------------------------------------------
729    Name: Allen Leung
730    Date: 2000/05/10 21:17 EDT
731    Tag: leunga-20000510-moby-c--ssa
732    Description:
733    
734      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
735    None of these affect SML/NJ.
736    
737    1. Register Allocation
738    
739        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
740           This module tries to remove local (i.e. basic block level) redundancies
741           during spilling.
742    
743        b. A new framework for performing region based register allocation.
744           Not yet entirely functional.
745    
746    2. X86
747    
748       a. DefUse for POP was missing the stack pointer [found by Lal]
749       b. Reload for CALL was incorrect in X86Spill [found by John]
750       c. Various fixes in X86Spill so that it can be used correctly for
751          the new spilling module.
752    
753    3. SSA/IR
754    
755       a. New module ir/dj-dataflow.sml implements elimination based
756          data flow analysis.
757    
758    4. MLRiscGen
759    
760       a. Fix for gc type annotation
761    
762    5. MDGen
763    
764       Various fixes for machine description -> ml code translation.  For ssa
765       only.
766    
767    ----------------------------------------------------------------------
768    Name: Allen Leung
769    Date: 2000/05/08 22:17 EDT
770    Tag: leunga-20000508-labexp
771    Description:
772    
773      Fermin has found a few assembly problems with constant expressions
774      generated in LabelExp.  Mostly, the problems involve extra parentheses,
775      which choke on dumb assemblers.  This is his fix.
776    
777    ----------------------------------------------------------------------
778    Name: Dave MacQueen
779    Date: 2000/04/09 14:00 EDT
780    Tag: dbm-20000502-Version_110_28
781    Description:
782    
783    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
784    
785    2. Updated config/version to 110.28
786    
787    3. Updated config/srcarchiveurl
788    
789    3. New boot files!
790       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
791    
792    ----------------------------------------------------------------------
793    Name: Matthias Blume
794    Date: 2000/05/01 19:05:00 JST
795    Tag: blume-20000501-noweb
796    Description:
797    
798    A new noweb tool has been added.  The existing system is entirely
799    unaffected by this, but some CM users have asked for renewed noweb
800    support.  Everything is documented in the CM manual.
801    
802    New (plugin) libraries:
803    
804       noweb-tool.cm
805       nw-ext.cm
806    
807    ----------------------------------------------------------------------
808    Name: Dave MacQueen
809    Date: 2000/04/30 12:40PM EDT
810    Tag: dbm-20000430-bug_fixes
811    Description:
812    
813    1. Fix for bug 1498
814       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
815       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
816         added toRealArray function
817       smlnj/src/compiler/MiscUtil/print/ppobj.sml
818         added check for tag Obj.RealArray to array printing case in ppObj
819    
820    2. Fix for bug 1510
821       smlnj/src/compiler/Semant/types/typesutil.sml
822         fixed definition of dummyargs (used by equalTycon) so that
823         dummy args are distinct types
824    
825    ----------------------------------------------------------------------
826    Name: Matthias Blume
827    Date: 2000/04/30 01:00:00 JST
828    Tag: blume-20000430-versions
829    Description:
830    
831    1. CM version numbering added.  This is an implementation of Lal's
832       proposal for adding version numbers and version checking to .cm
833       files.  Lal said that his proposal was just that -- a proposal.
834       For the time being I went ahead and implemented it so that people
835       can comment on it.  Everything is completely backward-compatible
836       (except for the stable library format, i.e., new bootfiles!).
837    
838       As usual, see the CM manual for details.
839    
840    2. An alternative syntax for anchored paths has been implemented.
841       Dave has recently voiced the same concerns that I had when I did
842       this, so there should be some support.  My take is that eventually
843       I will let support for the current syntax (where anchors are
844       "implicit") fade out in favor of the new, explicit syntax.
845       In order to be backward-compatible, both old and new syntax are
846       currently supported.
847    
848       Again, see the CM manual for details.
849    
850    3. Parallel make is trying to be slightly smarter:  When the master
851       process finds a "bottleneck", i.e., when there is only one
852       compilation unit that can be compiled and everybody else is
853       waiting on it, then it will simply compile it directly instead
854       of clumsily telling one of the slaves to do it.
855    
856    4. Support for "unsharing" added.  This is necessary in order to be
857       able to have two different versions of the same library running
858       at the same time (e.g., for trying out a new MLRISC while still
859       having the old MLRISC linked into the current compiler, etc.)
860       See the CM manual.
861    
862    5. Simple "makedepend" functionality added for generating Makefile
863       dependency information.  (This is rather crude at the moment.
864       Expect some changes here in the future.)
865    
866    6. ".fun" added as a recognized suffix for ML files. Also documented
867       explicitly in the manual that the fallback behavior (unknown suffix
868       -> ML file) is not an official feature!
869    
870    7. Small changes to the pickler for stable libraries.
871    
872    8. Several internal changes to CM (for cleanup/improvement).
873    
874    
875    !!!! NEW BINFILES !!!!
876    
877    ----------------------------------------------------------------------
878    Name: Matthias Blume
879    Date: 2000/04/28 17:30:00 JST
880    Tag: blume-20000428-pathconfig
881    Description:
882    
883    1. I changed config/install.sh to remove duplicate entries from the
884       lib/pathconfig file at the end.  Moreover, the final version of
885       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
886       in src/system/installml.
887    
888    2. The config/install.sh script now consistently uses relative
889       pathnames in lib/pathconfig whenever the anchor is in the lib
890       directory.  (So far this was true for the libraries that come
891       pre-compiled and bundled as part of the bootfiles but not for
892       libraries that are compiled by the script itself.)
893    
894    ----------------------------------------------------------------------
895  Name: Matthias Blume  Name: Matthias Blume
896  Date: 2000/04/26 13:10:00 JST  Date: 2000/04/26 13:10:00 JST
897  Tag: blume-20000426-fun_suffix  Tag: blume-20000426-fun_suffix
# Line 208  Line 1088 
1088    
1089  3.  Assembly  3.  Assembly
1090    
1091      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
1092      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
1093      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
1094    
# Line 231  Line 1111 
1111    
1112  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
1113  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
1114  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
1115  legal surface-syntax symbol).  legal surface-syntax symbol).
1116    
1117  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 241  Line 1121 
1121  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
1122    
1123  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1124  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
1125  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
1126  more details.  more details.
1127    
# Line 356  Line 1236 
1236     (specified in the .cm file at each instance where the tool's class is     (specified in the .cm file at each instance where the tool's class is
1237     used).     used).
1238    
1239     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
1240     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
1241     Makefiles or Shell scripts.     Makefiles or Shell scripts.
1242    
1243     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 369  Line 1249 
1249    
1250     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
1251    
1252     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
1253     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
1254    
1255  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 503  Line 1383 
1383    
1384       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
1385       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
1386       testl is perferable to andl.       testl is preferable to andl.
1387    
1388  7. RA (x86 only)  7. RA (x86 only)
1389    
# Line 685  Line 1565 
1565    
1566  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
1567  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
1568  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
1569  explanation.  explanation.
1570    
1571  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
1572  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
1573  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
1574  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 732  Line 1612 
1612  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
1613  <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,
1614  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
1615  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
1616  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
1617  and, thus, saves a lot of time.  and, thus, saves a lot of time.
1618    
# Line 782  Line 1662 
1662    it from that remote directory.    it from that remote directory.
1663    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
1664    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
1665    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
1666    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
1667    
1668    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.629  
changed lines
  Added in v.679

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