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 635, Tue May 2 15:09:12 2000 UTC revision 680, Mon Jul 3 06:35:55 2000 UTC
# Line 12  Line 12 
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14  ----------------------------------------------------------------------  ----------------------------------------------------------------------
15    Name: Matthias Blume
16    Date: 2000/07/02 15:36:00 JST
17    Tag: blume-20000702-manual
18    Description:
20    Small corrections and updates to CM manual.
22    ----------------------------------------------------------------------
23    Name: Matthias Blumen
24    Date: 2000/06/29 16:04:00 JST
25    Tag: blume-20000629-yacctool
26    Description:
28    Changes:
30    1. Class "mlyacc" now takes separate arguments to pass options to
31       generated .sml- and .sig-files independently.
32    2. Corresponding CM manual updates.
33    3. BTrace module now also reports call sites.  (However, for loop clusters
34       it only shows from where the cluster was entered.)  There are associated
35       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
37    ----------------------------------------------------------------------
38    Name: Matthias Blume
39    Date: 2000/06/27 16:51:00 JST
40    Tag: blume-20000627-noweb
41    Description:
43    Changes:
45     1. Implemented "subdir" and "witness" options for noweb tool.
46        This caused some slight internal changes in CM's tool implementation.
47     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
48        some remaining issues from earlier path anchor changes.
49     3. Updated CM manual accordingly.
51     4. Changed implementation of back-tracing so that I now consider it
52        ready for prime-time.
54        In particular, you don't have to explicitly trigger the back-trace
55        anymore.  Instead, if you are running BTrace-instrumented code and
56        there is an uncaught exception (regardless of whether or not it was
57        raised in instrumented code), the top-level evalloop will print
58        the back-trace.
60        Features:
62          - Instrumented and uninstrumented code work together seemlessly.
63            (Of course, uninstrumented code is never mentioned in actual
64             back-traces.)
66          - Asymptotic time- and space-complexity of instrumented code is
67            equal to that of uninstrumented code.  (This means that
68            tail-recursion is preserved by the instrumentation phase.)
70          - Modules whose code has been instrumented in different sessions
71            work together without problem.
73          - There is no penalty whatsoever on uninstrumented code.
75          - There is no penalty on "raise" expressions, even in
76            instrumented code.
78        A potential bug (or perhaps it is a feature, too):
80          A back-trace reaches no further than the outermost instrumented
81          non-trivial "raise".  Here, a "trivial" raise is one that is the
82          sole RHS of a "handle" rule.  Thus, back-traces reach trough
84               <exp> handle e => raise e
86          and even
88               <exp> handle Foo => raise Bar
90          and, of course, through
92               <exp> handle Foo => ...
94         if the exception was not Foo.
96         Back-traces always reach right through any un-instrumented code
97         including any of its "handle" expressions, trivial or not.
99       To try this out, do the following:
101         - Erase all existing binfiles for your program.
102           (You may keep binfiles for those modules where you think you
103            definitely don't need back-tracing.)
104         - Turn on back-trace instrumentation:
105              SMLofNJ.Internals.BTrace.mode (SOME true);
106         - Recompile your program.  (I.e., run "CM.make" or "use".)
107         - You may now turn instrumentation off again (if you want):
108              SMLofNJ.Internals.BTrace.mode (SOME false);
109         - Run your program as usual.  If it raises an exception that
110           reaches the interactive toplevel, then a back-trace will
111           automatically be printed.  After that, the toplevel loop
112           will print the exception history as usual.
114    ----------------------------------------------------------------------
115    Name: Matthias Blume
116    Date: 2000/06/26 09:56:46 JST
117    Tag: blume-20000626-setup
118    Description:
120    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
121          ML code before and after compiling a file (e.g., to set compiler
122          flags)
124    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
125              - associated changes to btrace.sml (BTrace instrumentation pass)
126              - cleaner implementation of btimp.sml (BTrace tracing and report
127                module)
129    CM manual: * new path encoding documented
130               * description of setup-parameter to "sml" added
132    The biggest user-visible change to back-tracing is that it is no
133    longer necessary to compile all traced modules within the same
134    session.  (This was a real limitation.)
136    ----------------------------------------------------------------------
137    Name: Matthias Blume
138    Date: 2000/06/24 12:40:00 JST
139    Tag: blume-20000624-startup
140    Description:
142    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
143    bit too often -- to put it mildly. :)
145    ----------------------------------------------------------------------
146    Name: Matthias Blume
147    Date: 2000/06/23 18:20:00 JST
148    Tag: blume-20000623-btrace
149    Description:
151    This updates adds a backtrace facility to aid programmers in debugging
152    their programs.  This involves the following changes:
154    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
155       keeping track of the current call stack.  When programs are compiled
156       in a special mode, the compiler will insert calls to these hooks
157       into the user program.
158       "Hook" means that it is possible for different implementations of
159       back-tracing to register themselves (at different times).
161    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
162       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
163       It can be turned on using this call:
164         SMLofNJ.Internals.BTrace.mode (SOME true);
165       Turning it off again:
166         SMLofNJ.Internals.BTrace.mode (SOME false);
167       Querying the current status:
168         SMLofNJ.Internals.BTrace.mode NONE;
169       Annotated programs are about twice as big as normal ones, and they
170       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
171       where all hooks do nothing).  The slowdown with a plugin that is
172       actually useful (such as the one supplied by default) is even greater,
173       but in the case of the default plugin it is still only an constant
174       factor (amortized).
176    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
177       with a sub-structure BTrace for controlling back-tracing.  In particular,
178       the above-mentioned function "mode" controls whether the annotation
179       phase is invoked by the compiler.  Another important function is
180       "trigger": when called it aborts the current execution and causes
181       the top-level loop to print a full back-trace.
183    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
184       for back-tracing.  It keeps track of the dynamic call stack and in
185       addition to that it keeps a partial history at each "level" of that
186       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
187       d and b (at separate times, dynamically), then the report will show:
189       GOTO   d
190             /c
191       GOTO  \b
192       CALL   a
194       This shows that there was an initial non-tail call of a, then a
195       tail-call to b or c, looping behavior in a cluster of functions that
196       consist of b and c, and then a goto from that cluster (i.e., either from
197       b or from c) to d.
199       Note that (depending on the user program) the amount of information
200       that the back-trace module has to keep track of at each level is bounded
201       by a constant.  Thus, the whole implementation has the same asymptotical
202       complexity as the original program (both in space and in time).
204    5. compiler/TopLevel/interact/evalloop.sml has been modified to
205       handle the special exception SMLofNJ.Internals.BTrace.BTrace
206       which is raised by the "trigger" function mentioned above.
208    Notes on usage:
210    - Annotated code works well together with unannotated code:
211    Unannotated calls simply do not show up at all in the backtrace.
213    - It is not a good idea to let modules that were annotated during
214    different sessions run at the same time.  This is because the compiler
215    chooses small integers to identify individual functions, and there
216    will be clashes if different modules were compiled in separate sessions.
217    (Nothing will crash, and you will even be told about the clashes, but
218    back-trace information will in general not be useful.)
220    - Back-tracing can be confused by callcc and capture.
222    - The only way of getting a back-trace right now is to explicitly
223    invoke the "trigger" function from your user program.  Eventually, we
224    should make every exception carry back-trace information (if
225    available).  But since this creates more overhead at "raise"-time
226    (similar to the current exnHistory overhead), I have not yet
227    implemented this.  (The implementation will be rather easy.)  With
228    exceptions carrying back-trace information, this facility will be even
229    more useful because users don't need to modify their programs...
231    - While it is possible to compile the compiler with back-trace
232    annotations turned on (I did it to get some confidence in
233    correctness), you must make absolutely sure that core.sml and
234    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
235    be compiled with annotation because there is no core access yet, but
236    if you compile btimp.sml with annotation, then the system will go into
237    an infinite recursion and crash.)
238    Since CM currently does not know about BTrace, the only way to turn
239    annotations on and off for different modules of the compiler is to
240    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
241    this is awkward and clumsy.
243    Sample sessions:
245    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
246    - SMLofNJ.Internals.BTrace.mode (SOME true);
247    [autoloading]
248    [autoloading done]
249    val it = false : bool
250    - structure X = struct
251    -     fun main n = let
252    -         fun a (x, 0) = d x
253    -           | a (x, n) = b (x, n - 1)
254    -         and b (x, n) = c (x, n)
255    -         and c (x, n) = a (x, n)
256    -         and d x = e (x, 3)
257    -         and e (x, 0) = f x
258    -           | e (x, n) = e (x, n - 1)
259    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
260    -           | f n = n * g (n - 1)
261    -         and g n = a (n, 3)
262    -     in
263    -         f n
264    -     end
265    - end;
266    structure X : sig val main : int -> int end
267    - X.main 3;
268    *** BACK-TRACE ***
269    GOTO   stdIn:4.2-13.20: X.main[2].f
270    GOTO-( stdIn:4.2-13.20: X.main[2].e
271    GOTO   stdIn:4.2-13.20: X.main[2].d
272         / stdIn:4.2-13.20: X.main[2].a
273         | stdIn:4.2-13.20: X.main[2].b
274    GOTO-\ stdIn:4.2-13.20: X.main[2].c
275    CALL   stdIn:4.2-13.20: X.main[2].g
276    GOTO   stdIn:4.2-13.20: X.main[2].f
277    GOTO-( stdIn:4.2-13.20: X.main[2].e
278    GOTO   stdIn:4.2-13.20: X.main[2].d
279         / stdIn:4.2-13.20: X.main[2].a
280         | stdIn:4.2-13.20: X.main[2].b
281    GOTO-\ stdIn:4.2-13.20: X.main[2].c
282    CALL   stdIn:4.2-13.20: X.main[2].g
283    GOTO   stdIn:4.2-13.20: X.main[2].f
284    GOTO-( stdIn:4.2-13.20: X.main[2].e
285    GOTO   stdIn:4.2-13.20: X.main[2].d
286         / stdIn:4.2-13.20: X.main[2].a
287         | stdIn:4.2-13.20: X.main[2].b
288    GOTO-\ stdIn:4.2-13.20: X.main[2].c
289    CALL   stdIn:4.2-13.20: X.main[2].g
290    GOTO   stdIn:4.2-13.20: X.main[2].f
291    CALL   stdIn:2.15-17.4: X.main[2]
292    -
294    (Note that because of a FLINt bug the above code currently does not
295    compile without BTrace turned on.)
297    Here is another example, using my modified Tiger compiler:
299    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
300    - SMLofNJ.Internals.BTrace.mode (SOME true);
301    [autoloading]
302    [autoloading done]
303    val it = false : bool
304    - CM.make "sources.cm";
305    [autoloading]
306    ...
307    [autoloading done]
308    [scanning sources.cm]
309    [parsing (sources.cm):parse.sml]
310    [creating directory CM/SKEL ...]
311    [parsing (sources.cm):tiger.lex.sml]
312    ...
313    [wrote CM/sparc-unix/semant.sml]
314    [compiling (sources.cm):main.sml]
315    [wrote CM/sparc-unix/main.sml]
316    [New bindings added.]
317    val it = true : bool
318    - Main.compile ("../testcases/merge.tig", "foo.out");
319    *** BACK-TRACE ***
320    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
321    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
322    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
323    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
324    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
325    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
326    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
327         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
328    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
329    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
330    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
331    CALL   main.sml:1.18-118.4: Main.compile[2]
332    -
334    ----------------------------------------------------------------------
335    Name: Matthias Blumen
336    Date: 2000/06/21 18:00:00 JST
337    Tag: blume-20000621-manual
338    Description:
340    CM manual update: Path environments documented.
342    ----------------------------------------------------------------------
343    Name: Matthias Blume
344    Date: 2000/06/19 13:40:00
345    Tag: blume-20000619-manual
346    Description:
348    CM manual and system/README update.  This only covers the fact that
349    there are no more implicit anchors.  (Path environments and the "bind"
350    option to "cm" have yet to be documented.)
352    ----------------------------------------------------------------------
353    Name: Matthias Blume
354    Date: 2000/06/19 11:05:00 JST
355    Tag: blume-20000619-chdir-bugfix
356    Description:
358    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
360    ----------------------------------------------------------------------
361    Name: Matthias Blume
362    Date: 2000/06/18 22:00:10 JST
363    Tag: blume-20000618-implicit-anchors-really-gone
364    Description:
366    I updates the previous HISTORY entry where I forgot to mention that
367    implicit anchors are no longer with us.
369    The current update also gets rid of the (now useless) controller
370    CM.Control.implicit_anchors.
372    ----------------------------------------------------------------------
373    Name: Matthias Blume
374    Date: 2000/06/16 17:30:00 JST
375    Tag: blume-20000616-anchorenv
376    Description:
378    This patch implements the long anticipated (just kidding :) "anchor
379    environment" mechanism.  In the course of doing this, I also
380    re-implemented CM's internal "SrcPath" module from scratch.  The new
381    one should be more robust in certain boundary cases.  In any case, it
382    is a lot cleaner than its predecessor (IMHO).
384    This time, although there is yet another boot file format change, I
385    kept the unpickler backward-compatible.  As a result, no new bootfiles
386    are necessary and bootstrapping is straightforward.  (You cannot read
387    new bootfiles into an old system, but the other way around is no
388    problem.)
390    Visible changes:
392    ** 0. Implicit path anchors (without the leading $-symbol) are no
393    longer recognized at all. This means that such path names are not
394    illegal either.  For example, the name basis.cm simply refers to a
395    local file called "basis.cm" (i.e, the name is an ordinary path
396    relative to .cm-files directory).  Or, to put it differently, only
397    names that start with $ are anchored paths.
399    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
400    vanished.
402    John (Reppy) had critizised this as soon as I originally proposed and
403    implemented it, but at that time I did not really deeply believe
404    him. :) Now I came full-circle because I need the $<singlearc> syntax
405    in another place where it cannot be seen as an abbreviation for
406    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
407    seems to mean (i.e., it "expands" into the corresponding anchor
408    value).
410    However, when paths are used as members in CM description files, it
411    continues to be true that there must be at least another arc after the
412    anchor.  This is now enforced separately during semantic analysis
413    (i.e., from a lexical/syntactical point of view, the notation is ok.)
415    ** 2. The "cm" class now accepts an option "bind".  The option's value
416    is a sub-option list of precisely two items -- one labeled "anchor"
417    and the other one labeled "value".  As you might expect, "anchor" is
418    used to specify an anchor name to be bound, and "value" specifies what
419    the anchor is being bound to.
421    The value must be a directory name and can be given in either standard
422    syntax (including the possibility that it is itself an anchored path)
423    or native syntax.
425    Examples:
427       foo.cm (bind:(anchor:bar value:$mystuff/bar))
428       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
430    and so on.
432    The meaning of this is that the .cm-file will be processed with an
433    augmented anchor environment where the given anchor(s) is/are bound to
434    the given values(s).
436    The rationale for having this feature is this: Suppose you are trying
437    to use two different (already stable) libraries a.cm and b.cm (that
438    you perhaps didn't write yourself).  Further, suppose each of these
439    two libraries internally uses its own auxiliary library $aux/lib.cm.
440    Normally you would now have a problem because the anchor "lib" can not
441    be bound to more than one value globally.  Therefore, the project that
442    uses both a.cm and b.cm must locally redirect the anchor to some other
443    place:
445       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
446       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
448    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
449    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
451    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
452    you when you do so at the time of CM.stabilize).  Therefore, you can
453    also use an anchored path as the value:
455      a.cm (bind:(anchor:lib value:$a-lib))
456      b.cm (bind:(anchor:lib value:$b-lib))
458    Now you can globally configure (using the usual CM.Anchor.anchor or
459    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
460    itself is always locally bound, setting it globally is no longer
461    meaningful or necessary (but it does not hurt either).  In fact, "lib"
462    can still be used as a global anchor for separate purposes.  As a
463    matter of fact, one can locally define "lib" in terms of a global
464    "lib":
466      a.cm (bind:(anchor:lib value:$lib/a))
467      b.cm (bind:(anchor:lib value:$lib/b))
469    ** 3: The encoding of path names has changed.  This affects the way
470    path names are shown in CM's progress report and also the internal
471    protocol encoding used for parallel make.
473    The encoding now uses one or more ':'-separated segments.  Each
474    segments corresponds to a file that has been specified relative to the
475    file given by its preceding segment.  The first segment is either
476    relative to the CWD, absolute, or anchored.  Each segment itself is
477    basically a Unix pathname; all segments but the first are relative.
479    Example:
481       $foo/bar/baz.cm:a/b/c.sml
483    This path denotes the file bar/a/b/c.sml relative to the directory
484    denoted by anchor "foo".  Notice that the encoding also includes
485    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
486    paths are resolved relative to the .cm-files directory, so baz.cm must
487    be ignored to get the "real" pathname.
489    To make this fact more obvious, CM puts the names of such "virtual
490    arcs" into parentheses when they appear in progress reports. (No
491    parentheses will appear in the internal protocol encoding.)  Thus,
492    what you really see is:
494      $foo/bar/(baz.cm):a/b/c.sml
496    I find this notation to be much more informative than before.
498    Another new feature of the encoding is that special characters
499    including parentheses, colons, (back)slashes, and white space are
500    written as \ddd (where ddd is the decimal encoding of the character).
502    *** The CM manual still needs to be updated.
504    ----------------------------------------------------------------------
505    Name: Allen Leung
506    Date: 2000/06/15 00:38:00
507    Tag: leunga-20000615-x86-peephole
509    x86 Peephole fix by Fermin.  Affects c-- and moby only.
511    ----------------------------------------------------------------------
512    Name: Matthias Blume
513    Date: 2000/06/12 11:40:00
514    Tag: blume-20000612-parmakefix
515    Description:
517    More cleanup after changing the file naming scheme: This time I
518    repaired the parallel make mechanism for CMB.make which I broke earlier.
520    ----------------------------------------------------------------------
521    Name: Allen Leung
522    Date: 2000/06/09 01:25:00
523    Tag: leunga-20000609-various
525    None of these things should affect normal SML/NJ operations
527    1. Peephole improvements provided by Fermin (c--)
528    2. New annotation DEFUSE for adding extra dependence (moby)
529    3. New X86 LOCK instructions (moby)
530    4. New machine description language for reservation tables (scheduling)
531    5. Fixes to various optimization/analysis modules (branch chaining, dominator
532       trees etc.)
533    6. I've changed the CM files so that they can work with versions
534       110.0.6, 110.25 and 110.28
536    ----------------------------------------------------------------------
537    Name: Matthias Blume
538    Date: 2000/06/09 12:40:00
539    Tag: blume-20000609-log
540    Description:
542    - Removed all(?) remaining RCS Log entries from sources.
544    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
545      anchored paths).
547    ----------------------------------------------------------------------
548    Name: Matthias Blume
549    Date: 2000/06/07 17:00:00 JST
550    Tag: blume-20000607-no-implicit-anchors
551    Description:
553    1. This update changes the default setting for
554    CM.Control.implicit_anchors from true to false.  This means that
555    implicit anchors are no longer permitted by default.  I also tried to
556    make sure that nothing else still relies on implicit anchors.
557    (This is the next step on the schedule towards a CM that does not even
558    have the notion of implicit anchors anymore.)
560    2. More CM manual updates.
562    3. I managed to track down and fix the pickling bug I mentioned last
563    time.  Because of the previously existing workaround, this entails no
564    immediate practical changes.
566    ----------------------------------------------------------------------
567    Name: Matthias Blume
568    Date: 2000/06/06 11:15:00 JST
569    Tag: blume-20000606-lazierpickle
570    Description:
572    !!!! NEW BOOT FILES !!!!
574    * The main purpose of this update is to make library pickles lazier in
575    order to reduce the initial space penalty for autoloading a library.
576    As a result, it is now possible to have $smlnj/compiler.cm
577    pre-registered.  This should take care of the many complaints or
578    inquiries about missing structure Compiler.  This required changes to
579    CM's internal data structures and small tweaks to some algorithms.
581    As a neat additional effect, it is no longer necessary (for the sake
582    of lean heap image files) to distinguish between a "minimal" CM and a
583    "full" CM.  Now, there is only one CM (i.e., the "full" version:
584    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
585    interactive top level. ($smlnj/cm/minimal.cm is gone.)
587    To make the life of compiler-hackers easier, "makeml" now also
588    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
589    words, after you bootstrap a new sml for the first time, you will not
590    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
591    around you will still have to do it, though.)
593    * A second change consists of major updates to the CM manual.  There
594    are now several appendices with summary information and also a full
595    specification of the CM description file syntax.
597    * In directory src/system I added the script "allcross".  This script
598    invokes sml and cross-compiles the compiler for all supported
599    architectures.  (Useful when providing a new set of boot files.)
601    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
602    added a small tweak to pickle-util.sml to work around this problem,
603    but it is not a proper fix yet.  I will investigate further.  (The
604    effect of the bug was an inflation of library pickle size.)
606    * Version number increased to 110.28.1 (to avoid compatibility problems).
608    ----------------------------------------------------------------------
609    Name: Allen Leung
610    Date: 2000/05/25 17:28 EDT
611    Tag: leunga-20000525-ra
612    Description:
614      Fixed a bug in freezing phase of the register allocator.
616    ----------------------------------------------------------------------
617    Name: Allen Leung
618    Date: 2000/05/15 22:53 EDT
619    Tag: leunga-20000515-alpha-x86-ra
620    Description:
622      1. Alpha
624          Slight cleanup.  Removed the instruction SGNXL
626      2. X86
628          Added the following instructions to the instruction set:
630            ROLx, RORx,
631            BTx, BTSx, BTLx, BTRx,
632            XCHGx, and variants with the LOCK prefix
634      3. Register Allocation
636          The module ra-rewrite-with-renaming has been improved.
638      These have no effect on SML/NJ.
640    ----------------------------------------------------------------------
641    Name: Matthias Blume
642    Date: 2000/05/15 16:20:00 JST
643    Tag: blume-20000515-lightrebuild
644    Description:
646    1. I added an alternative to "-rebuild" to "makeml".  The difference is
647       that prior to calling CMB.make' the CM-variable "LIGHT" will be
648       defined.  In effect, the command will not build any cross-compiler
649       backends and therefore finish more quickly.
651       The "fixpt" script also takes a "-light" switch to be able to use
652       this new facility while compiling for a fixpoint.
654    2. I replaced all mentions of anchored paths in group owner specifications
655       with simple relative paths (usually starting with "..").
656       The rationale is that a library's internal workings should not be
657       compromised by the lack of some anchor.  (An anchor is necessary
658       for someone who wants to refer to the library by an anchored path,
659       but it should not be necessary to build the same library in the first
660       place.)
662    3. I changed the way CM's tool mechanism determines the shell command
663       string used for things like ml-yacc etc. so that it does not break
664       when CM.Control.implicit_anchors is turned off.
666    ----------------------------------------------------------------------
667    Name: Matthias Blume
668    Date: 2000/05/12 18:20:00 JST
669    Tag: blume-20000512-ml-build
670    Description:
672    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
673    getting installed properly (by config/install.sh).
675    ----------------------------------------------------------------------
676    Name: Matthias Blume
677    Date: 2000/05/12 17:30:00 JST
678    Tag: blume-20000512-anchors
679    Description:
681    !!! NEW BOOT FILES !!!
683    This change is in preparation of fading out support for "implicitly
684    anchored path names".  I went through all sources and used the
685    explicit (and relatively new) $-notation.  See system/README and the
686    CM manual for more info on this.
688    I also modified the anchoring scheme for some things such as "smlnj",
689    "MLRISC", "cm", etc. to take advantage of the fact that explicit
690    anchors are more expressive: anchor name and first arc do not have to
691    coincide.  This entails the following user-visible change:
693    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
694    particular, when you fire up sml with a command-line argument, say,
695    e.g.:
697       sml '$smlnj/cmb.cm'
699    At the ML toplevel prompt:
701       CM.autoload "$smlnj/cmb.cm";
703    There is also a new controller in CM.Control that can be used to turn
704    off all remaining support for implicit anchors by saying:
706        CM.autoload "$smlnj/
707        #set CM.Control.implicit_anchors false;
709    This causes CM to reject implicitly anchored paths.  This is (for the
710    time being) less permissive than the "final" version where there will
711    be no more such implicit anchors and relative paths will be just that:
712    relative.
714    The next step (version after next version?) will be to make the
715    default for CM.Control.implicit_anchors false.  After the dust has
716    settled, I can then produce the "final" version of this...
718    Note: Since bootstrapping is a bit tricky, I provided new boot files.
720    ----------------------------------------------------------------------
721    Name: Matthias Blume
722    Date: 2000/05/11 16:30:00 JST
723    Tag: blume-20000511-sources
724    Description:
726    The main change is that I added function CM.sources as a generalized
727    version of the earlier CM.makedepend.  This entails the following
728    additional changes:
730      - CM.makedepend has been dropped.
732      - CM manual has been updated.
734      - TOOLS signature and API have been changed.
736    ----------------------------------------------------------------------
737    Name: Allen Leung
738    Date: 2000/05/10 21:17 EDT
739    Tag: leunga-20000510-moby-c--ssa
740    Description:
742      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
743    None of these affect SML/NJ.
745    1. Register Allocation
747        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
748           This module tries to remove local (i.e. basic block level) redundancies
749           during spilling.
751        b. A new framework for performing region based register allocation.
752           Not yet entirely functional.
754    2. X86
756       a. DefUse for POP was missing the stack pointer [found by Lal]
757       b. Reload for CALL was incorrect in X86Spill [found by John]
758       c. Various fixes in X86Spill so that it can be used correctly for
759          the new spilling module.
761    3. SSA/IR
763       a. New module ir/dj-dataflow.sml implements elimination based
764          data flow analysis.
766    4. MLRiscGen
768       a. Fix for gc type annotation
770    5. MDGen
772       Various fixes for machine description -> ml code translation.  For ssa
773       only.
775    ----------------------------------------------------------------------
776    Name: Allen Leung
777    Date: 2000/05/08 22:17 EDT
778    Tag: leunga-20000508-labexp
779    Description:
781      Fermin has found a few assembly problems with constant expressions
782      generated in LabelExp.  Mostly, the problems involve extra parentheses,
783      which choke on dumb assemblers.  This is his fix.
785    ----------------------------------------------------------------------
786  Name: Dave MacQueen  Name: Dave MacQueen
787  Date: 2000/04/09 14:00 EDT  Date: 2000/04/09 14:00 EDT
788  Tag: dbm-20000502-Version_110_28  Tag: dbm-20000502-Version_110_28
# Line 325  Line 1096 
1097  3.  Assembly  3.  Assembly
1099      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
1100      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
1101      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
# Line 348  Line 1119 
1120  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
1121  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
1122  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
1123  legal surface-syntax symbol).  legal surface-syntax symbol).
1125  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 358  Line 1129 
1129  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
1131  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1132  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
1133  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
1134  more details.  more details.
# Line 473  Line 1244 
1244     (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
1245     used).     used).
1247     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
1248     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
1249     Makefiles or Shell scripts.     Makefiles or Shell scripts.
1251     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 486  Line 1257 
1258     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
1260     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
1261     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
1263  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 620  Line 1391 
1392       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
1393       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
1394       testl is perferable to andl.       testl is preferable to andl.
1396  7. RA (x86 only)  7. RA (x86 only)
# Line 802  Line 1573 
1574  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
1575  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
1576  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
1577  explanation.  explanation.
1579  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
1580  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
1581  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
1582  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 849  Line 1620 
1620  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
1621  <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,
1622  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
1623  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
1624  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
1625  and, thus, saves a lot of time.  and, thus, saves a lot of time.
# Line 899  Line 1670 
1670    it from that remote directory.    it from that remote directory.
1671    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
1672    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
1673    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
1674    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
1676    For CVS users, this mechanism is not relevant for source archives, but    For CVS users, this mechanism is not relevant for source archives, but

Removed from v.635  
changed lines
  Added in v.680

ViewVC Help
Powered by ViewVC 1.0.0