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 695, Mon Aug 7 23:57:38 2000 UTC
# Line 12  Line 12 
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14  ----------------------------------------------------------------------  ----------------------------------------------------------------------
15    Name: Allen Leung
16    Date: 2000/08/7 19:31:00
17    Tag: leunga-20000807-a-whole-bunch-of-stuff
18    
19      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
20    and Mobdy.
21    
22    1.  x86
23    
24       a.  Fixes to peephole module by John and Dan.
25       b.  Assembly fix to SETcc by Allen.
26       c.  Fix to c-call by John.
27       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
28       e.  Instruction selection optimization to SETcc as suggested by John.
29    
30           For example,
31    
32            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
33    
34           should generate:
35    
36            MOVL a, x
37            SUBL b, x
38            SHRL 31, x
39    
40    2.  IR stuff
41    
42         A bunch of new DJ-graph related algorithms added.  These
43         speed up SSA construction.
44    
45    3.  SSA + Scheduling
46    
47         Added code for SSA and scheduling to the repository
48    
49    ----------------------------------------------------------------------
50    Name: Lal George
51    Date: 2000/07/27 11:53:14 EDT
52    
53    Tag: lal-20000727-linux-ppc
54    Description:
55    
56     Made changes to support Linux PPC.
57     p.s. I have confirmation that the 110.29 boot files work fine.
58    
59    ----------------------------------------------------------------------
60    Name: Matthias Blume
61    Date: 2000/07/27 17:40:00 JST
62    Tag: blume-20000727-scripts
63    Description:
64    
65    !!!! WARNING !!!!
66    You must recompile the runtime system!
67    !!!! WARNING !!!!
68    
69    This is basically another round of script-enhancements:
70    
71    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
72       and undefine CM preprocessor symbols.
73    
74    2. ml-build avoids generating a new heap image if it finds that the
75       existing one is still ok.  (The condition is that no ML file had to
76       be recompiled and all ML files are found to be older that the heap
77       file.)
78    
79       To make this work smoothly, I also hacked the runtime system as
80       well as SMLofNJ.SysInfo to get access to the heap image suffix
81       (.sparc-solaris, ...) that is currently being used.
82    
83       Moreover, the signature of CM.mk_standalone has changed.  See the
84       CM manual.
85    
86    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
87       CM manual for details.)
88    
89    4. More CM manual updates:
90        - all of the above has been documented.
91        - there is now a section describing the (CM-related) command line
92          arguments that are accepted by the "sml" command
93    
94    ----------------------------------------------------------------------
95    Name: Matthias Blume
96    Date: 2000/07/25 16:20:00 JST
97    Tag: blume-20000725-makedepend
98    Description:
99    
100    Added a script called ml-makedepend.  This can be used in makefiles
101    for Unix' make in a way very similar to the "makedepend" command for
102    C.
103    
104    The script internally uses function CM.sources.
105    
106    Synopsis:
107    
108        ml-makedepend [-f makefile] cmfile targetname
109    
110    The default for the makefile is "makefile" (or "Makefile" should
111    "makefile" not exist).
112    
113    ml-makedepend adds a cmfile/targetname-specific section to this
114    makefile (after removing the previous version of this section).  The
115    section contains a single dependency specification with targetname on
116    the LHS (targetname is an arbitrary name), and a list of files derived
117    from the cmfile on the RHS.  Some of the files on the RHS are
118    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
119    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
120    path names.  The makefile writer is responsible for making sure that
121    these variables have correct at the time "make" is invoked.
122    
123    ----------------------------------------------------------------------
124    Name: Matthias Blume
125    Date: 2000/07/22 23:30:00 JST
126    Tag: blume-20000722-urlupdate
127    Description:
128    
129    Changed BOOT and config/srcarchiveurl to point to BL server:
130    
131        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
132    
133    ----------------------------------------------------------------------
134    Name: Matthias Blume
135    Date: 2000/07/18 18:00:00 JST
136    Tag: blume-20000718-Version_110_29
137    Description:
138    
139    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
140    
141    2. Updated config/version to 110.29
142    
143    3. Updated config/srcarchiveurl
144    
145    3. New boot files!
146       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
147    
148    ----------------------------------------------------------------------
149    Name: Matthias Blume
150    Date: 2000/07/11 13:58:00 JST
151    Tag: blume-20000711-doctypo
152    Description:
153    
154    Fixed a few typos in CM manual.
155    
156    ----------------------------------------------------------------------
157    Name: Allen Leung
158    Date: 2000/06/15 00:38:00
159    Tag: leunga-20000704-sparc-x86
160    
161    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
162    2. fix to x86 RET bug [found by Dan Grossman]
163    3. sparc assembly bug fix for ticc instructions [found by Fermin]
164    
165       Affects c-- and moby only
166    
167    ----------------------------------------------------------------------
168    Name: Matthias Blume
169    Date: 2000/07/04 15:26:00
170    Tag: blume-20000704-trigger
171    Description:
172    
173    1. Improvements to CM manual.
174    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
175       of getting a back-trace.  The function, when called, raises an
176       internal exception which explicitly carries the full back-trace history,
177       so it is unaffected by any intervening handle-raise pairs ("trivial"
178       or not).  The interactive loop will print that history once it arrives
179       at top level.
180       Short of having all exceptions implicitly carry the full history, the
181       recommended way of using this facility is:
182         - compile your program with instrumentation "on"
183         - run it, when it raises an exception, look at the history
184         - if the history is "cut off" because of some handler, go and modify
185           your program so that it explicitly calls BTrace.trigger
186         - recompile (still instrumented), and rerun; look at the full history
187    
188    ----------------------------------------------------------------------
189    Name: Matthias Blume
190    Date: 2000/07/03 15:36:00 JST
191    Tag: blume-20000702-manual
192    Description:
193    
194    Small corrections and updates to CM manual.
195    
196    ----------------------------------------------------------------------
197    Name: Matthias Blume
198    Date: 2000/06/29 16:04:00 JST
199    Tag: blume-20000629-yacctool
200    Description:
201    
202    Changes:
203    
204    1. Class "mlyacc" now takes separate arguments to pass options to
205       generated .sml- and .sig-files independently.
206    2. Corresponding CM manual updates.
207    3. BTrace module now also reports call sites.  (However, for loop clusters
208       it only shows from where the cluster was entered.)  There are associated
209       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
210    
211    ----------------------------------------------------------------------
212    Name: Matthias Blume
213    Date: 2000/06/27 16:51:00 JST
214    Tag: blume-20000627-noweb
215    Description:
216    
217    Changes:
218    
219     1. Implemented "subdir" and "witness" options for noweb tool.
220        This caused some slight internal changes in CM's tool implementation.
221     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
222        some remaining issues from earlier path anchor changes.
223     3. Updated CM manual accordingly.
224    
225     4. Changed implementation of back-tracing so that I now consider it
226        ready for prime-time.
227    
228        In particular, you don't have to explicitly trigger the back-trace
229        anymore.  Instead, if you are running BTrace-instrumented code and
230        there is an uncaught exception (regardless of whether or not it was
231        raised in instrumented code), the top-level evalloop will print
232        the back-trace.
233    
234        Features:
235    
236          - Instrumented and uninstrumented code work together seemlessly.
237            (Of course, uninstrumented code is never mentioned in actual
238             back-traces.)
239    
240          - Asymptotic time- and space-complexity of instrumented code is
241            equal to that of uninstrumented code.  (This means that
242            tail-recursion is preserved by the instrumentation phase.)
243    
244          - Modules whose code has been instrumented in different sessions
245            work together without problem.
246    
247          - There is no penalty whatsoever on uninstrumented code.
248    
249          - There is no penalty on "raise" expressions, even in
250            instrumented code.
251    
252        A potential bug (or perhaps it is a feature, too):
253    
254          A back-trace reaches no further than the outermost instrumented
255          non-trivial "raise".  Here, a "trivial" raise is one that is the
256          sole RHS of a "handle" rule.  Thus, back-traces reach trough
257    
258               <exp> handle e => raise e
259    
260          and even
261    
262               <exp> handle Foo => raise Bar
263    
264          and, of course, through
265    
266               <exp> handle Foo => ...
267    
268         if the exception was not Foo.
269    
270         Back-traces always reach right through any un-instrumented code
271         including any of its "handle" expressions, trivial or not.
272    
273       To try this out, do the following:
274    
275         - Erase all existing binfiles for your program.
276           (You may keep binfiles for those modules where you think you
277            definitely don't need back-tracing.)
278         - Turn on back-trace instrumentation:
279              SMLofNJ.Internals.BTrace.mode (SOME true);
280         - Recompile your program.  (I.e., run "CM.make" or "use".)
281         - You may now turn instrumentation off again (if you want):
282              SMLofNJ.Internals.BTrace.mode (SOME false);
283         - Run your program as usual.  If it raises an exception that
284           reaches the interactive toplevel, then a back-trace will
285           automatically be printed.  After that, the toplevel loop
286           will print the exception history as usual.
287    
288    ----------------------------------------------------------------------
289    Name: Matthias Blume
290    Date: 2000/06/26 09:56:46 JST
291    Tag: blume-20000626-setup
292    Description:
293    
294    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
295          ML code before and after compiling a file (e.g., to set compiler
296          flags)
297    
298    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
299              - associated changes to btrace.sml (BTrace instrumentation pass)
300              - cleaner implementation of btimp.sml (BTrace tracing and report
301                module)
302    
303    CM manual: * new path encoding documented
304               * description of setup-parameter to "sml" added
305    
306    The biggest user-visible change to back-tracing is that it is no
307    longer necessary to compile all traced modules within the same
308    session.  (This was a real limitation.)
309    
310    ----------------------------------------------------------------------
311    Name: Matthias Blume
312    Date: 2000/06/24 12:40:00 JST
313    Tag: blume-20000624-startup
314    Description:
315    
316    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
317    bit too often -- to put it mildly. :)
318    
319    ----------------------------------------------------------------------
320    Name: Matthias Blume
321    Date: 2000/06/23 18:20:00 JST
322    Tag: blume-20000623-btrace
323    Description:
324    
325    This updates adds a backtrace facility to aid programmers in debugging
326    their programs.  This involves the following changes:
327    
328    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
329       keeping track of the current call stack.  When programs are compiled
330       in a special mode, the compiler will insert calls to these hooks
331       into the user program.
332       "Hook" means that it is possible for different implementations of
333       back-tracing to register themselves (at different times).
334    
335    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
336       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
337       It can be turned on using this call:
338         SMLofNJ.Internals.BTrace.mode (SOME true);
339       Turning it off again:
340         SMLofNJ.Internals.BTrace.mode (SOME false);
341       Querying the current status:
342         SMLofNJ.Internals.BTrace.mode NONE;
343       Annotated programs are about twice as big as normal ones, and they
344       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
345       where all hooks do nothing).  The slowdown with a plugin that is
346       actually useful (such as the one supplied by default) is even greater,
347       but in the case of the default plugin it is still only an constant
348       factor (amortized).
349    
350    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
351       with a sub-structure BTrace for controlling back-tracing.  In particular,
352       the above-mentioned function "mode" controls whether the annotation
353       phase is invoked by the compiler.  Another important function is
354       "trigger": when called it aborts the current execution and causes
355       the top-level loop to print a full back-trace.
356    
357    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
358       for back-tracing.  It keeps track of the dynamic call stack and in
359       addition to that it keeps a partial history at each "level" of that
360       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
361       d and b (at separate times, dynamically), then the report will show:
362    
363       GOTO   d
364             /c
365       GOTO  \b
366       CALL   a
367    
368       This shows that there was an initial non-tail call of a, then a
369       tail-call to b or c, looping behavior in a cluster of functions that
370       consist of b and c, and then a goto from that cluster (i.e., either from
371       b or from c) to d.
372    
373       Note that (depending on the user program) the amount of information
374       that the back-trace module has to keep track of at each level is bounded
375       by a constant.  Thus, the whole implementation has the same asymptotical
376       complexity as the original program (both in space and in time).
377    
378    5. compiler/TopLevel/interact/evalloop.sml has been modified to
379       handle the special exception SMLofNJ.Internals.BTrace.BTrace
380       which is raised by the "trigger" function mentioned above.
381    
382    Notes on usage:
383    
384    - Annotated code works well together with unannotated code:
385    Unannotated calls simply do not show up at all in the backtrace.
386    
387    - It is not a good idea to let modules that were annotated during
388    different sessions run at the same time.  This is because the compiler
389    chooses small integers to identify individual functions, and there
390    will be clashes if different modules were compiled in separate sessions.
391    (Nothing will crash, and you will even be told about the clashes, but
392    back-trace information will in general not be useful.)
393    
394    - Back-tracing can be confused by callcc and capture.
395    
396    - The only way of getting a back-trace right now is to explicitly
397    invoke the "trigger" function from your user program.  Eventually, we
398    should make every exception carry back-trace information (if
399    available).  But since this creates more overhead at "raise"-time
400    (similar to the current exnHistory overhead), I have not yet
401    implemented this.  (The implementation will be rather easy.)  With
402    exceptions carrying back-trace information, this facility will be even
403    more useful because users don't need to modify their programs...
404    
405    - While it is possible to compile the compiler with back-trace
406    annotations turned on (I did it to get some confidence in
407    correctness), you must make absolutely sure that core.sml and
408    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
409    be compiled with annotation because there is no core access yet, but
410    if you compile btimp.sml with annotation, then the system will go into
411    an infinite recursion and crash.)
412    Since CM currently does not know about BTrace, the only way to turn
413    annotations on and off for different modules of the compiler is to
414    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
415    this is awkward and clumsy.
416    
417    Sample sessions:
418    
419    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
420    - SMLofNJ.Internals.BTrace.mode (SOME true);
421    [autoloading]
422    [autoloading done]
423    val it = false : bool
424    - structure X = struct
425    -     fun main n = let
426    -         fun a (x, 0) = d x
427    -           | a (x, n) = b (x, n - 1)
428    -         and b (x, n) = c (x, n)
429    -         and c (x, n) = a (x, n)
430    -         and d x = e (x, 3)
431    -         and e (x, 0) = f x
432    -           | e (x, n) = e (x, n - 1)
433    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
434    -           | f n = n * g (n - 1)
435    -         and g n = a (n, 3)
436    -     in
437    -         f n
438    -     end
439    - end;
440    structure X : sig val main : int -> int end
441    - X.main 3;
442    *** BACK-TRACE ***
443    GOTO   stdIn:4.2-13.20: X.main[2].f
444    GOTO-( stdIn:4.2-13.20: X.main[2].e
445    GOTO   stdIn:4.2-13.20: X.main[2].d
446         / stdIn:4.2-13.20: X.main[2].a
447         | stdIn:4.2-13.20: X.main[2].b
448    GOTO-\ stdIn:4.2-13.20: X.main[2].c
449    CALL   stdIn:4.2-13.20: X.main[2].g
450    GOTO   stdIn:4.2-13.20: X.main[2].f
451    GOTO-( stdIn:4.2-13.20: X.main[2].e
452    GOTO   stdIn:4.2-13.20: X.main[2].d
453         / stdIn:4.2-13.20: X.main[2].a
454         | stdIn:4.2-13.20: X.main[2].b
455    GOTO-\ stdIn:4.2-13.20: X.main[2].c
456    CALL   stdIn:4.2-13.20: X.main[2].g
457    GOTO   stdIn:4.2-13.20: X.main[2].f
458    GOTO-( stdIn:4.2-13.20: X.main[2].e
459    GOTO   stdIn:4.2-13.20: X.main[2].d
460         / stdIn:4.2-13.20: X.main[2].a
461         | stdIn:4.2-13.20: X.main[2].b
462    GOTO-\ stdIn:4.2-13.20: X.main[2].c
463    CALL   stdIn:4.2-13.20: X.main[2].g
464    GOTO   stdIn:4.2-13.20: X.main[2].f
465    CALL   stdIn:2.15-17.4: X.main[2]
466    -
467    
468    (Note that because of a FLINt bug the above code currently does not
469    compile without BTrace turned on.)
470    
471    Here is another example, using my modified Tiger compiler:
472    
473    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
474    - SMLofNJ.Internals.BTrace.mode (SOME true);
475    [autoloading]
476    [autoloading done]
477    val it = false : bool
478    - CM.make "sources.cm";
479    [autoloading]
480    ...
481    [autoloading done]
482    [scanning sources.cm]
483    [parsing (sources.cm):parse.sml]
484    [creating directory CM/SKEL ...]
485    [parsing (sources.cm):tiger.lex.sml]
486    ...
487    [wrote CM/sparc-unix/semant.sml]
488    [compiling (sources.cm):main.sml]
489    [wrote CM/sparc-unix/main.sml]
490    [New bindings added.]
491    val it = true : bool
492    - Main.compile ("../testcases/merge.tig", "foo.out");
493    *** BACK-TRACE ***
494    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
495    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
496    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
497    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
498    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
499    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
500    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
501         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
502    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
503    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
504    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
505    CALL   main.sml:1.18-118.4: Main.compile[2]
506    -
507    
508    ----------------------------------------------------------------------
509    Name: Matthias Blumen
510    Date: 2000/06/21 18:00:00 JST
511    Tag: blume-20000621-manual
512    Description:
513    
514    CM manual update: Path environments documented.
515    
516    ----------------------------------------------------------------------
517    Name: Matthias Blume
518    Date: 2000/06/19 13:40:00
519    Tag: blume-20000619-manual
520    Description:
521    
522    CM manual and system/README update.  This only covers the fact that
523    there are no more implicit anchors.  (Path environments and the "bind"
524    option to "cm" have yet to be documented.)
525    
526    ----------------------------------------------------------------------
527    Name: Matthias Blume
528    Date: 2000/06/19 11:05:00 JST
529    Tag: blume-20000619-chdir-bugfix
530    Description:
531    
532    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
533    
534    ----------------------------------------------------------------------
535    Name: Matthias Blume
536    Date: 2000/06/18 22:00:10 JST
537    Tag: blume-20000618-implicit-anchors-really-gone
538    Description:
539    
540    I updates the previous HISTORY entry where I forgot to mention that
541    implicit anchors are no longer with us.
542    
543    The current update also gets rid of the (now useless) controller
544    CM.Control.implicit_anchors.
545    
546    ----------------------------------------------------------------------
547    Name: Matthias Blume
548    Date: 2000/06/16 17:30:00 JST
549    Tag: blume-20000616-anchorenv
550    Description:
551    
552    This patch implements the long anticipated (just kidding :) "anchor
553    environment" mechanism.  In the course of doing this, I also
554    re-implemented CM's internal "SrcPath" module from scratch.  The new
555    one should be more robust in certain boundary cases.  In any case, it
556    is a lot cleaner than its predecessor (IMHO).
557    
558    This time, although there is yet another boot file format change, I
559    kept the unpickler backward-compatible.  As a result, no new bootfiles
560    are necessary and bootstrapping is straightforward.  (You cannot read
561    new bootfiles into an old system, but the other way around is no
562    problem.)
563    
564    Visible changes:
565    
566    ** 0. Implicit path anchors (without the leading $-symbol) are no
567    longer recognized at all. This means that such path names are not
568    illegal either.  For example, the name basis.cm simply refers to a
569    local file called "basis.cm" (i.e, the name is an ordinary path
570    relative to .cm-files directory).  Or, to put it differently, only
571    names that start with $ are anchored paths.
572    
573    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
574    vanished.
575    
576    John (Reppy) had critizised this as soon as I originally proposed and
577    implemented it, but at that time I did not really deeply believe
578    him. :) Now I came full-circle because I need the $<singlearc> syntax
579    in another place where it cannot be seen as an abbreviation for
580    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
581    seems to mean (i.e., it "expands" into the corresponding anchor
582    value).
583    
584    However, when paths are used as members in CM description files, it
585    continues to be true that there must be at least another arc after the
586    anchor.  This is now enforced separately during semantic analysis
587    (i.e., from a lexical/syntactical point of view, the notation is ok.)
588    
589    ** 2. The "cm" class now accepts an option "bind".  The option's value
590    is a sub-option list of precisely two items -- one labeled "anchor"
591    and the other one labeled "value".  As you might expect, "anchor" is
592    used to specify an anchor name to be bound, and "value" specifies what
593    the anchor is being bound to.
594    
595    The value must be a directory name and can be given in either standard
596    syntax (including the possibility that it is itself an anchored path)
597    or native syntax.
598    
599    Examples:
600    
601       foo.cm (bind:(anchor:bar value:$mystuff/bar))
602       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
603    
604    and so on.
605    
606    The meaning of this is that the .cm-file will be processed with an
607    augmented anchor environment where the given anchor(s) is/are bound to
608    the given values(s).
609    
610    The rationale for having this feature is this: Suppose you are trying
611    to use two different (already stable) libraries a.cm and b.cm (that
612    you perhaps didn't write yourself).  Further, suppose each of these
613    two libraries internally uses its own auxiliary library $aux/lib.cm.
614    Normally you would now have a problem because the anchor "lib" can not
615    be bound to more than one value globally.  Therefore, the project that
616    uses both a.cm and b.cm must locally redirect the anchor to some other
617    place:
618    
619       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
620       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
621    
622    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
623    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
624    
625    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
626    you when you do so at the time of CM.stabilize).  Therefore, you can
627    also use an anchored path as the value:
628    
629      a.cm (bind:(anchor:lib value:$a-lib))
630      b.cm (bind:(anchor:lib value:$b-lib))
631    
632    Now you can globally configure (using the usual CM.Anchor.anchor or
633    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
634    itself is always locally bound, setting it globally is no longer
635    meaningful or necessary (but it does not hurt either).  In fact, "lib"
636    can still be used as a global anchor for separate purposes.  As a
637    matter of fact, one can locally define "lib" in terms of a global
638    "lib":
639    
640      a.cm (bind:(anchor:lib value:$lib/a))
641      b.cm (bind:(anchor:lib value:$lib/b))
642    
643    ** 3: The encoding of path names has changed.  This affects the way
644    path names are shown in CM's progress report and also the internal
645    protocol encoding used for parallel make.
646    
647    The encoding now uses one or more ':'-separated segments.  Each
648    segments corresponds to a file that has been specified relative to the
649    file given by its preceding segment.  The first segment is either
650    relative to the CWD, absolute, or anchored.  Each segment itself is
651    basically a Unix pathname; all segments but the first are relative.
652    
653    Example:
654    
655       $foo/bar/baz.cm:a/b/c.sml
656    
657    This path denotes the file bar/a/b/c.sml relative to the directory
658    denoted by anchor "foo".  Notice that the encoding also includes
659    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
660    paths are resolved relative to the .cm-files directory, so baz.cm must
661    be ignored to get the "real" pathname.
662    
663    To make this fact more obvious, CM puts the names of such "virtual
664    arcs" into parentheses when they appear in progress reports. (No
665    parentheses will appear in the internal protocol encoding.)  Thus,
666    what you really see is:
667    
668      $foo/bar/(baz.cm):a/b/c.sml
669    
670    I find this notation to be much more informative than before.
671    
672    Another new feature of the encoding is that special characters
673    including parentheses, colons, (back)slashes, and white space are
674    written as \ddd (where ddd is the decimal encoding of the character).
675    
676    *** The CM manual still needs to be updated.
677    
678    ----------------------------------------------------------------------
679    Name: Allen Leung
680    Date: 2000/06/15 00:38:00
681    Tag: leunga-20000615-x86-peephole
682    
683    x86 Peephole fix by Fermin.  Affects c-- and moby only.
684    
685    ----------------------------------------------------------------------
686    Name: Matthias Blume
687    Date: 2000/06/12 11:40:00
688    Tag: blume-20000612-parmakefix
689    Description:
690    
691    More cleanup after changing the file naming scheme: This time I
692    repaired the parallel make mechanism for CMB.make which I broke earlier.
693    
694    ----------------------------------------------------------------------
695    Name: Allen Leung
696    Date: 2000/06/09 01:25:00
697    Tag: leunga-20000609-various
698    
699    None of these things should affect normal SML/NJ operations
700    
701    1. Peephole improvements provided by Fermin (c--)
702    2. New annotation DEFUSE for adding extra dependence (moby)
703    3. New X86 LOCK instructions (moby)
704    4. New machine description language for reservation tables (scheduling)
705    5. Fixes to various optimization/analysis modules (branch chaining, dominator
706       trees etc.)
707    6. I've changed the CM files so that they can work with versions
708       110.0.6, 110.25 and 110.28
709    
710    ----------------------------------------------------------------------
711    Name: Matthias Blume
712    Date: 2000/06/09 12:40:00
713    Tag: blume-20000609-log
714    Description:
715    
716    - Removed all(?) remaining RCS Log entries from sources.
717    
718    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
719      anchored paths).
720    
721    ----------------------------------------------------------------------
722    Name: Matthias Blume
723    Date: 2000/06/07 17:00:00 JST
724    Tag: blume-20000607-no-implicit-anchors
725    Description:
726    
727    1. This update changes the default setting for
728    CM.Control.implicit_anchors from true to false.  This means that
729    implicit anchors are no longer permitted by default.  I also tried to
730    make sure that nothing else still relies on implicit anchors.
731    (This is the next step on the schedule towards a CM that does not even
732    have the notion of implicit anchors anymore.)
733    
734    2. More CM manual updates.
735    
736    3. I managed to track down and fix the pickling bug I mentioned last
737    time.  Because of the previously existing workaround, this entails no
738    immediate practical changes.
739    
740    ----------------------------------------------------------------------
741    Name: Matthias Blume
742    Date: 2000/06/06 11:15:00 JST
743    Tag: blume-20000606-lazierpickle
744    Description:
745    
746    !!!! NEW BOOT FILES !!!!
747    
748    * The main purpose of this update is to make library pickles lazier in
749    order to reduce the initial space penalty for autoloading a library.
750    As a result, it is now possible to have $smlnj/compiler.cm
751    pre-registered.  This should take care of the many complaints or
752    inquiries about missing structure Compiler.  This required changes to
753    CM's internal data structures and small tweaks to some algorithms.
754    
755    As a neat additional effect, it is no longer necessary (for the sake
756    of lean heap image files) to distinguish between a "minimal" CM and a
757    "full" CM.  Now, there is only one CM (i.e., the "full" version:
758    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
759    interactive top level. ($smlnj/cm/minimal.cm is gone.)
760    
761    To make the life of compiler-hackers easier, "makeml" now also
762    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
763    words, after you bootstrap a new sml for the first time, you will not
764    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
765    around you will still have to do it, though.)
766    
767    * A second change consists of major updates to the CM manual.  There
768    are now several appendices with summary information and also a full
769    specification of the CM description file syntax.
770    
771    * In directory src/system I added the script "allcross".  This script
772    invokes sml and cross-compiles the compiler for all supported
773    architectures.  (Useful when providing a new set of boot files.)
774    
775    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
776    added a small tweak to pickle-util.sml to work around this problem,
777    but it is not a proper fix yet.  I will investigate further.  (The
778    effect of the bug was an inflation of library pickle size.)
779    
780    * Version number increased to 110.28.1 (to avoid compatibility problems).
781    
782    ----------------------------------------------------------------------
783    Name: Allen Leung
784    Date: 2000/05/25 17:28 EDT
785    Tag: leunga-20000525-ra
786    Description:
787    
788      Fixed a bug in freezing phase of the register allocator.
789    
790    ----------------------------------------------------------------------
791    Name: Allen Leung
792    Date: 2000/05/15 22:53 EDT
793    Tag: leunga-20000515-alpha-x86-ra
794    Description:
795    
796      1. Alpha
797    
798          Slight cleanup.  Removed the instruction SGNXL
799    
800      2. X86
801    
802          Added the following instructions to the instruction set:
803    
804            ROLx, RORx,
805            BTx, BTSx, BTLx, BTRx,
806            XCHGx, and variants with the LOCK prefix
807    
808      3. Register Allocation
809    
810          The module ra-rewrite-with-renaming has been improved.
811    
812      These have no effect on SML/NJ.
813    
814    ----------------------------------------------------------------------
815    Name: Matthias Blume
816    Date: 2000/05/15 16:20:00 JST
817    Tag: blume-20000515-lightrebuild
818    Description:
819    
820    1. I added an alternative to "-rebuild" to "makeml".  The difference is
821       that prior to calling CMB.make' the CM-variable "LIGHT" will be
822       defined.  In effect, the command will not build any cross-compiler
823       backends and therefore finish more quickly.
824    
825       The "fixpt" script also takes a "-light" switch to be able to use
826       this new facility while compiling for a fixpoint.
827    
828    2. I replaced all mentions of anchored paths in group owner specifications
829       with simple relative paths (usually starting with "..").
830       The rationale is that a library's internal workings should not be
831       compromised by the lack of some anchor.  (An anchor is necessary
832       for someone who wants to refer to the library by an anchored path,
833       but it should not be necessary to build the same library in the first
834       place.)
835    
836    3. I changed the way CM's tool mechanism determines the shell command
837       string used for things like ml-yacc etc. so that it does not break
838       when CM.Control.implicit_anchors is turned off.
839    
840    ----------------------------------------------------------------------
841    Name: Matthias Blume
842    Date: 2000/05/12 18:20:00 JST
843    Tag: blume-20000512-ml-build
844    Description:
845    
846    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
847    getting installed properly (by config/install.sh).
848    
849    ----------------------------------------------------------------------
850    Name: Matthias Blume
851    Date: 2000/05/12 17:30:00 JST
852    Tag: blume-20000512-anchors
853    Description:
854    
855    !!! NEW BOOT FILES !!!
856    
857    This change is in preparation of fading out support for "implicitly
858    anchored path names".  I went through all sources and used the
859    explicit (and relatively new) $-notation.  See system/README and the
860    CM manual for more info on this.
861    
862    I also modified the anchoring scheme for some things such as "smlnj",
863    "MLRISC", "cm", etc. to take advantage of the fact that explicit
864    anchors are more expressive: anchor name and first arc do not have to
865    coincide.  This entails the following user-visible change:
866    
867    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
868    particular, when you fire up sml with a command-line argument, say,
869    e.g.:
870    
871       sml '$smlnj/cmb.cm'
872    
873    At the ML toplevel prompt:
874    
875       CM.autoload "$smlnj/cmb.cm";
876    
877    There is also a new controller in CM.Control that can be used to turn
878    off all remaining support for implicit anchors by saying:
879    
880        CM.autoload "$smlnj/
881        #set CM.Control.implicit_anchors false;
882    
883    This causes CM to reject implicitly anchored paths.  This is (for the
884    time being) less permissive than the "final" version where there will
885    be no more such implicit anchors and relative paths will be just that:
886    relative.
887    
888    The next step (version after next version?) will be to make the
889    default for CM.Control.implicit_anchors false.  After the dust has
890    settled, I can then produce the "final" version of this...
891    
892    Note: Since bootstrapping is a bit tricky, I provided new boot files.
893    
894    ----------------------------------------------------------------------
895    Name: Matthias Blume
896    Date: 2000/05/11 16:30:00 JST
897    Tag: blume-20000511-sources
898    Description:
899    
900    The main change is that I added function CM.sources as a generalized
901    version of the earlier CM.makedepend.  This entails the following
902    additional changes:
903    
904      - CM.makedepend has been dropped.
905    
906      - CM manual has been updated.
907    
908      - TOOLS signature and API have been changed.
909    
910    ----------------------------------------------------------------------
911    Name: Allen Leung
912    Date: 2000/05/10 21:17 EDT
913    Tag: leunga-20000510-moby-c--ssa
914    Description:
915    
916      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
917    None of these affect SML/NJ.
918    
919    1. Register Allocation
920    
921        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
922           This module tries to remove local (i.e. basic block level) redundancies
923           during spilling.
924    
925        b. A new framework for performing region based register allocation.
926           Not yet entirely functional.
927    
928    2. X86
929    
930       a. DefUse for POP was missing the stack pointer [found by Lal]
931       b. Reload for CALL was incorrect in X86Spill [found by John]
932       c. Various fixes in X86Spill so that it can be used correctly for
933          the new spilling module.
934    
935    3. SSA/IR
936    
937       a. New module ir/dj-dataflow.sml implements elimination based
938          data flow analysis.
939    
940    4. MLRiscGen
941    
942       a. Fix for gc type annotation
943    
944    5. MDGen
945    
946       Various fixes for machine description -> ml code translation.  For ssa
947       only.
948    
949    ----------------------------------------------------------------------
950    Name: Allen Leung
951    Date: 2000/05/08 22:17 EDT
952    Tag: leunga-20000508-labexp
953    Description:
954    
955      Fermin has found a few assembly problems with constant expressions
956      generated in LabelExp.  Mostly, the problems involve extra parentheses,
957      which choke on dumb assemblers.  This is his fix.
958    
959    ----------------------------------------------------------------------
960  Name: Dave MacQueen  Name: Dave MacQueen
961  Date: 2000/04/09 14:00 EDT  Date: 2000/04/09 14:00 EDT
962  Tag: dbm-20000502-Version_110_28  Tag: dbm-20000502-Version_110_28
# Line 325  Line 1270 
1270    
1271  3.  Assembly  3.  Assembly
1272    
1273      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
1274      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
1275      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
1276    
# Line 348  Line 1293 
1293    
1294  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
1295  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
1296  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
1297  legal surface-syntax symbol).  legal surface-syntax symbol).
1298    
1299  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 1303 
1303  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
1304    
1305  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1306  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
1307  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
1308  more details.  more details.
1309    
# Line 473  Line 1418 
1418     (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
1419     used).     used).
1420    
1421     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
1422     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
1423     Makefiles or Shell scripts.     Makefiles or Shell scripts.
1424    
1425     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 486  Line 1431 
1431    
1432     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
1433    
1434     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
1435     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
1436    
1437  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 1565 
1565    
1566       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
1567       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
1568       testl is perferable to andl.       testl is preferable to andl.
1569    
1570  7. RA (x86 only)  7. RA (x86 only)
1571    
# Line 802  Line 1747 
1747    
1748  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
1749  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
1750  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
1751  explanation.  explanation.
1752    
1753  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
1754  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
1755  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
1756  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 1794 
1794  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
1795  <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,
1796  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
1797  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
1798  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
1799  and, thus, saves a lot of time.  and, thus, saves a lot of time.
1800    
# Line 899  Line 1844 
1844    it from that remote directory.    it from that remote directory.
1845    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
1846    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
1847    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
1848    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
1849    
1850    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.635  
changed lines
  Added in v.695

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