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 634, Mon May 1 10:03:50 2000 UTC revision 710, Tue Oct 17 08:21:09 2000 UTC
# Line 11  Line 11 
11  Date:  Date:
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15    ----------------------------------------------------------------------
16    Name: Matthias Blume
17    Date: 2000/08/31 17:30:00 JST
18    Tag: blume-20001017-v110p29p1
19    Description:
20    
21    I made a version 110.29.1 with new bootfiles.
22    
23    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
24              CM documentation changes and a small bugfix in CM's error reporting.
25    
26    ----------------------------------------------------------------------
27    Name: Lal George
28    Date: 2000/09/27 14:42:35 EDT
29    Tag: george-20000927-nodestatus
30    Description:
31    
32    Changed the type of the nodestatus, so that:
33    
34            SPILLED(~1)             is now SPILLED
35            SPILLED(m) where m>=0   is now MEMREG(m)
36            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
37    
38    ----------------------------------------------------------------------
39    Name: Matthias Blume
40    Date: 2000/09/07 14:45:00 JST
41    Tag: blume-20000907-cmerrmsg
42    Description:
43    
44    Small tweak to CM to avoid getting ML syntax error messages twice.
45    
46    ----------------------------------------------------------------------
47    Name: Matthias Blume
48    Date: 2000/08/31 18:00:00 JST
49    Tag: blume-20000831-cvsbootfiles
50    Description:
51    
52    New URL for boot files (because the 110.29 files on the BL server do
53    now work correctly with my updated install scripts for yacc and lex).
54    
55    ----------------------------------------------------------------------
56    Name: Matthias Blume
57    Date: 2000/08/08 12:33:00 JST
58    Tag: blume-20000808-manual
59    Description:
60    
61    Tiny update to CM manual.
62    
63    ----------------------------------------------------------------------
64    Name: Allen Leung
65    Date: 2000/08/7 19:31:00
66    Tag: leunga-20000807-a-whole-bunch-of-stuff
67    
68      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
69    and Mobdy.
70    
71    1.  x86
72    
73       a.  Fixes to peephole module by John and Dan.
74       b.  Assembly fix to SETcc by Allen.
75       c.  Fix to c-call by John.
76       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
77       e.  Instruction selection optimization to SETcc as suggested by John.
78    
79           For example,
80    
81            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
82    
83           should generate:
84    
85            MOVL a, x
86            SUBL b, x
87            SHRL 31, x
88    
89    2.  IR stuff
90    
91         A bunch of new DJ-graph related algorithms added.  These
92         speed up SSA construction.
93    
94    3.  SSA + Scheduling
95    
96         Added code for SSA and scheduling to the repository
97    
98    ----------------------------------------------------------------------
99    Name: Lal George
100    Date: 2000/07/27 11:53:14 EDT
101    
102    Tag: lal-20000727-linux-ppc
103    Description:
104    
105     Made changes to support Linux PPC.
106     p.s. I have confirmation that the 110.29 boot files work fine.
107    
108    ----------------------------------------------------------------------
109    Name: Matthias Blume
110    Date: 2000/07/27 17:40:00 JST
111    Tag: blume-20000727-scripts
112    Description:
113    
114    !!!! WARNING !!!!
115    You must recompile the runtime system!
116    !!!! WARNING !!!!
117    
118    This is basically another round of script-enhancements:
119    
120    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
121       and undefine CM preprocessor symbols.
122    
123    2. ml-build avoids generating a new heap image if it finds that the
124       existing one is still ok.  (The condition is that no ML file had to
125       be recompiled and all ML files are found to be older that the heap
126       file.)
127    
128       To make this work smoothly, I also hacked the runtime system as
129       well as SMLofNJ.SysInfo to get access to the heap image suffix
130       (.sparc-solaris, ...) that is currently being used.
131    
132       Moreover, the signature of CM.mk_standalone has changed.  See the
133       CM manual.
134    
135    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
136       CM manual for details.)
137    
138    4. More CM manual updates:
139        - all of the above has been documented.
140        - there is now a section describing the (CM-related) command line
141          arguments that are accepted by the "sml" command
142    
143    ----------------------------------------------------------------------
144    Name: Matthias Blume
145    Date: 2000/07/25 16:20:00 JST
146    Tag: blume-20000725-makedepend
147    Description:
148    
149    Added a script called ml-makedepend.  This can be used in makefiles
150    for Unix' make in a way very similar to the "makedepend" command for
151    C.
152    
153    The script internally uses function CM.sources.
154    
155    Synopsis:
156    
157        ml-makedepend [-f makefile] cmfile targetname
158    
159    The default for the makefile is "makefile" (or "Makefile" should
160    "makefile" not exist).
161    
162    ml-makedepend adds a cmfile/targetname-specific section to this
163    makefile (after removing the previous version of this section).  The
164    section contains a single dependency specification with targetname on
165    the LHS (targetname is an arbitrary name), and a list of files derived
166    from the cmfile on the RHS.  Some of the files on the RHS are
167    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
168    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
169    path names.  The makefile writer is responsible for making sure that
170    these variables have correct at the time "make" is invoked.
171    
172    ----------------------------------------------------------------------
173    Name: Matthias Blume
174    Date: 2000/07/22 23:30:00 JST
175    Tag: blume-20000722-urlupdate
176    Description:
177    
178    Changed BOOT and config/srcarchiveurl to point to BL server:
179    
180        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
181    
182    ----------------------------------------------------------------------
183    Name: Matthias Blume
184    Date: 2000/07/18 18:00:00 JST
185    Tag: blume-20000718-Version_110_29
186    Description:
187    
188    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
189    
190    2. Updated config/version to 110.29
191    
192    3. Updated config/srcarchiveurl
193    
194    3. New boot files!
195       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
196    
197    ----------------------------------------------------------------------
198    Name: Matthias Blume
199    Date: 2000/07/11 13:58:00 JST
200    Tag: blume-20000711-doctypo
201    Description:
202    
203    Fixed a few typos in CM manual.
204    
205    ----------------------------------------------------------------------
206    Name: Allen Leung
207    Date: 2000/06/15 00:38:00
208    Tag: leunga-20000704-sparc-x86
209    
210    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
211    2. fix to x86 RET bug [found by Dan Grossman]
212    3. sparc assembly bug fix for ticc instructions [found by Fermin]
213    
214       Affects c-- and moby only
215    
216    ----------------------------------------------------------------------
217    Name: Matthias Blume
218    Date: 2000/07/04 15:26:00
219    Tag: blume-20000704-trigger
220    Description:
221    
222    1. Improvements to CM manual.
223    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
224       of getting a back-trace.  The function, when called, raises an
225       internal exception which explicitly carries the full back-trace history,
226       so it is unaffected by any intervening handle-raise pairs ("trivial"
227       or not).  The interactive loop will print that history once it arrives
228       at top level.
229       Short of having all exceptions implicitly carry the full history, the
230       recommended way of using this facility is:
231         - compile your program with instrumentation "on"
232         - run it, when it raises an exception, look at the history
233         - if the history is "cut off" because of some handler, go and modify
234           your program so that it explicitly calls BTrace.trigger
235         - recompile (still instrumented), and rerun; look at the full history
236    
237    ----------------------------------------------------------------------
238    Name: Matthias Blume
239    Date: 2000/07/03 15:36:00 JST
240    Tag: blume-20000702-manual
241    Description:
242    
243    Small corrections and updates to CM manual.
244    
245    ----------------------------------------------------------------------
246    Name: Matthias Blume
247    Date: 2000/06/29 16:04:00 JST
248    Tag: blume-20000629-yacctool
249    Description:
250    
251    Changes:
252    
253    1. Class "mlyacc" now takes separate arguments to pass options to
254       generated .sml- and .sig-files independently.
255    2. Corresponding CM manual updates.
256    3. BTrace module now also reports call sites.  (However, for loop clusters
257       it only shows from where the cluster was entered.)  There are associated
258       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
259    
260    ----------------------------------------------------------------------
261    Name: Matthias Blume
262    Date: 2000/06/27 16:51:00 JST
263    Tag: blume-20000627-noweb
264    Description:
265    
266    Changes:
267    
268     1. Implemented "subdir" and "witness" options for noweb tool.
269        This caused some slight internal changes in CM's tool implementation.
270     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
271        some remaining issues from earlier path anchor changes.
272     3. Updated CM manual accordingly.
273    
274     4. Changed implementation of back-tracing so that I now consider it
275        ready for prime-time.
276    
277        In particular, you don't have to explicitly trigger the back-trace
278        anymore.  Instead, if you are running BTrace-instrumented code and
279        there is an uncaught exception (regardless of whether or not it was
280        raised in instrumented code), the top-level evalloop will print
281        the back-trace.
282    
283        Features:
284    
285          - Instrumented and uninstrumented code work together seemlessly.
286            (Of course, uninstrumented code is never mentioned in actual
287             back-traces.)
288    
289          - Asymptotic time- and space-complexity of instrumented code is
290            equal to that of uninstrumented code.  (This means that
291            tail-recursion is preserved by the instrumentation phase.)
292    
293          - Modules whose code has been instrumented in different sessions
294            work together without problem.
295    
296          - There is no penalty whatsoever on uninstrumented code.
297    
298          - There is no penalty on "raise" expressions, even in
299            instrumented code.
300    
301        A potential bug (or perhaps it is a feature, too):
302    
303          A back-trace reaches no further than the outermost instrumented
304          non-trivial "raise".  Here, a "trivial" raise is one that is the
305          sole RHS of a "handle" rule.  Thus, back-traces reach trough
306    
307               <exp> handle e => raise e
308    
309          and even
310    
311               <exp> handle Foo => raise Bar
312    
313          and, of course, through
314    
315               <exp> handle Foo => ...
316    
317         if the exception was not Foo.
318    
319         Back-traces always reach right through any un-instrumented code
320         including any of its "handle" expressions, trivial or not.
321    
322       To try this out, do the following:
323    
324         - Erase all existing binfiles for your program.
325           (You may keep binfiles for those modules where you think you
326            definitely don't need back-tracing.)
327         - Turn on back-trace instrumentation:
328              SMLofNJ.Internals.BTrace.mode (SOME true);
329         - Recompile your program.  (I.e., run "CM.make" or "use".)
330         - You may now turn instrumentation off again (if you want):
331              SMLofNJ.Internals.BTrace.mode (SOME false);
332         - Run your program as usual.  If it raises an exception that
333           reaches the interactive toplevel, then a back-trace will
334           automatically be printed.  After that, the toplevel loop
335           will print the exception history as usual.
336    
337    ----------------------------------------------------------------------
338    Name: Matthias Blume
339    Date: 2000/06/26 09:56:46 JST
340    Tag: blume-20000626-setup
341    Description:
342    
343    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
344          ML code before and after compiling a file (e.g., to set compiler
345          flags)
346    
347    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
348              - associated changes to btrace.sml (BTrace instrumentation pass)
349              - cleaner implementation of btimp.sml (BTrace tracing and report
350                module)
351    
352    CM manual: * new path encoding documented
353               * description of setup-parameter to "sml" added
354    
355    The biggest user-visible change to back-tracing is that it is no
356    longer necessary to compile all traced modules within the same
357    session.  (This was a real limitation.)
358    
359    ----------------------------------------------------------------------
360    Name: Matthias Blume
361    Date: 2000/06/24 12:40:00 JST
362    Tag: blume-20000624-startup
363    Description:
364    
365    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
366    bit too often -- to put it mildly. :)
367    
368    ----------------------------------------------------------------------
369    Name: Matthias Blume
370    Date: 2000/06/23 18:20:00 JST
371    Tag: blume-20000623-btrace
372    Description:
373    
374    This updates adds a backtrace facility to aid programmers in debugging
375    their programs.  This involves the following changes:
376    
377    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
378       keeping track of the current call stack.  When programs are compiled
379       in a special mode, the compiler will insert calls to these hooks
380       into the user program.
381       "Hook" means that it is possible for different implementations of
382       back-tracing to register themselves (at different times).
383    
384    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
385       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
386       It can be turned on using this call:
387         SMLofNJ.Internals.BTrace.mode (SOME true);
388       Turning it off again:
389         SMLofNJ.Internals.BTrace.mode (SOME false);
390       Querying the current status:
391         SMLofNJ.Internals.BTrace.mode NONE;
392       Annotated programs are about twice as big as normal ones, and they
393       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
394       where all hooks do nothing).  The slowdown with a plugin that is
395       actually useful (such as the one supplied by default) is even greater,
396       but in the case of the default plugin it is still only an constant
397       factor (amortized).
398    
399    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
400       with a sub-structure BTrace for controlling back-tracing.  In particular,
401       the above-mentioned function "mode" controls whether the annotation
402       phase is invoked by the compiler.  Another important function is
403       "trigger": when called it aborts the current execution and causes
404       the top-level loop to print a full back-trace.
405    
406    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
407       for back-tracing.  It keeps track of the dynamic call stack and in
408       addition to that it keeps a partial history at each "level" of that
409       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
410       d and b (at separate times, dynamically), then the report will show:
411    
412       GOTO   d
413             /c
414       GOTO  \b
415       CALL   a
416    
417       This shows that there was an initial non-tail call of a, then a
418       tail-call to b or c, looping behavior in a cluster of functions that
419       consist of b and c, and then a goto from that cluster (i.e., either from
420       b or from c) to d.
421    
422       Note that (depending on the user program) the amount of information
423       that the back-trace module has to keep track of at each level is bounded
424       by a constant.  Thus, the whole implementation has the same asymptotical
425       complexity as the original program (both in space and in time).
426    
427    5. compiler/TopLevel/interact/evalloop.sml has been modified to
428       handle the special exception SMLofNJ.Internals.BTrace.BTrace
429       which is raised by the "trigger" function mentioned above.
430    
431    Notes on usage:
432    
433    - Annotated code works well together with unannotated code:
434    Unannotated calls simply do not show up at all in the backtrace.
435    
436    - It is not a good idea to let modules that were annotated during
437    different sessions run at the same time.  This is because the compiler
438    chooses small integers to identify individual functions, and there
439    will be clashes if different modules were compiled in separate sessions.
440    (Nothing will crash, and you will even be told about the clashes, but
441    back-trace information will in general not be useful.)
442    
443    - Back-tracing can be confused by callcc and capture.
444    
445    - The only way of getting a back-trace right now is to explicitly
446    invoke the "trigger" function from your user program.  Eventually, we
447    should make every exception carry back-trace information (if
448    available).  But since this creates more overhead at "raise"-time
449    (similar to the current exnHistory overhead), I have not yet
450    implemented this.  (The implementation will be rather easy.)  With
451    exceptions carrying back-trace information, this facility will be even
452    more useful because users don't need to modify their programs...
453    
454    - While it is possible to compile the compiler with back-trace
455    annotations turned on (I did it to get some confidence in
456    correctness), you must make absolutely sure that core.sml and
457    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
458    be compiled with annotation because there is no core access yet, but
459    if you compile btimp.sml with annotation, then the system will go into
460    an infinite recursion and crash.)
461    Since CM currently does not know about BTrace, the only way to turn
462    annotations on and off for different modules of the compiler is to
463    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
464    this is awkward and clumsy.
465    
466    Sample sessions:
467    
468    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
469    - SMLofNJ.Internals.BTrace.mode (SOME true);
470    [autoloading]
471    [autoloading done]
472    val it = false : bool
473    - structure X = struct
474    -     fun main n = let
475    -         fun a (x, 0) = d x
476    -           | a (x, n) = b (x, n - 1)
477    -         and b (x, n) = c (x, n)
478    -         and c (x, n) = a (x, n)
479    -         and d x = e (x, 3)
480    -         and e (x, 0) = f x
481    -           | e (x, n) = e (x, n - 1)
482    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
483    -           | f n = n * g (n - 1)
484    -         and g n = a (n, 3)
485    -     in
486    -         f n
487    -     end
488    - end;
489    structure X : sig val main : int -> int end
490    - X.main 3;
491    *** BACK-TRACE ***
492    GOTO   stdIn:4.2-13.20: X.main[2].f
493    GOTO-( stdIn:4.2-13.20: X.main[2].e
494    GOTO   stdIn:4.2-13.20: X.main[2].d
495         / stdIn:4.2-13.20: X.main[2].a
496         | stdIn:4.2-13.20: X.main[2].b
497    GOTO-\ stdIn:4.2-13.20: X.main[2].c
498    CALL   stdIn:4.2-13.20: X.main[2].g
499    GOTO   stdIn:4.2-13.20: X.main[2].f
500    GOTO-( stdIn:4.2-13.20: X.main[2].e
501    GOTO   stdIn:4.2-13.20: X.main[2].d
502         / stdIn:4.2-13.20: X.main[2].a
503         | stdIn:4.2-13.20: X.main[2].b
504    GOTO-\ stdIn:4.2-13.20: X.main[2].c
505    CALL   stdIn:4.2-13.20: X.main[2].g
506    GOTO   stdIn:4.2-13.20: X.main[2].f
507    GOTO-( stdIn:4.2-13.20: X.main[2].e
508    GOTO   stdIn:4.2-13.20: X.main[2].d
509         / stdIn:4.2-13.20: X.main[2].a
510         | stdIn:4.2-13.20: X.main[2].b
511    GOTO-\ stdIn:4.2-13.20: X.main[2].c
512    CALL   stdIn:4.2-13.20: X.main[2].g
513    GOTO   stdIn:4.2-13.20: X.main[2].f
514    CALL   stdIn:2.15-17.4: X.main[2]
515    -
516    
517    (Note that because of a FLINt bug the above code currently does not
518    compile without BTrace turned on.)
519    
520    Here is another example, using my modified Tiger compiler:
521    
522    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
523    - SMLofNJ.Internals.BTrace.mode (SOME true);
524    [autoloading]
525    [autoloading done]
526    val it = false : bool
527    - CM.make "sources.cm";
528    [autoloading]
529    ...
530    [autoloading done]
531    [scanning sources.cm]
532    [parsing (sources.cm):parse.sml]
533    [creating directory CM/SKEL ...]
534    [parsing (sources.cm):tiger.lex.sml]
535    ...
536    [wrote CM/sparc-unix/semant.sml]
537    [compiling (sources.cm):main.sml]
538    [wrote CM/sparc-unix/main.sml]
539    [New bindings added.]
540    val it = true : bool
541    - Main.compile ("../testcases/merge.tig", "foo.out");
542    *** BACK-TRACE ***
543    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
544    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
545    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
546    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
547    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
548    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
549    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
550         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
551    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
552    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
553    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
554    CALL   main.sml:1.18-118.4: Main.compile[2]
555    -
556    
557    ----------------------------------------------------------------------
558    Name: Matthias Blumen
559    Date: 2000/06/21 18:00:00 JST
560    Tag: blume-20000621-manual
561    Description:
562    
563    CM manual update: Path environments documented.
564    
565    ----------------------------------------------------------------------
566    Name: Matthias Blume
567    Date: 2000/06/19 13:40:00
568    Tag: blume-20000619-manual
569    Description:
570    
571    CM manual and system/README update.  This only covers the fact that
572    there are no more implicit anchors.  (Path environments and the "bind"
573    option to "cm" have yet to be documented.)
574    
575    ----------------------------------------------------------------------
576    Name: Matthias Blume
577    Date: 2000/06/19 11:05:00 JST
578    Tag: blume-20000619-chdir-bugfix
579    Description:
580    
581    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
582    
583    ----------------------------------------------------------------------
584    Name: Matthias Blume
585    Date: 2000/06/18 22:00:10 JST
586    Tag: blume-20000618-implicit-anchors-really-gone
587    Description:
588    
589    I updates the previous HISTORY entry where I forgot to mention that
590    implicit anchors are no longer with us.
591    
592    The current update also gets rid of the (now useless) controller
593    CM.Control.implicit_anchors.
594    
595    ----------------------------------------------------------------------
596    Name: Matthias Blume
597    Date: 2000/06/16 17:30:00 JST
598    Tag: blume-20000616-anchorenv
599    Description:
600    
601    This patch implements the long anticipated (just kidding :) "anchor
602    environment" mechanism.  In the course of doing this, I also
603    re-implemented CM's internal "SrcPath" module from scratch.  The new
604    one should be more robust in certain boundary cases.  In any case, it
605    is a lot cleaner than its predecessor (IMHO).
606    
607    This time, although there is yet another boot file format change, I
608    kept the unpickler backward-compatible.  As a result, no new bootfiles
609    are necessary and bootstrapping is straightforward.  (You cannot read
610    new bootfiles into an old system, but the other way around is no
611    problem.)
612    
613    Visible changes:
614    
615    ** 0. Implicit path anchors (without the leading $-symbol) are no
616    longer recognized at all. This means that such path names are not
617    illegal either.  For example, the name basis.cm simply refers to a
618    local file called "basis.cm" (i.e, the name is an ordinary path
619    relative to .cm-files directory).  Or, to put it differently, only
620    names that start with $ are anchored paths.
621    
622    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
623    vanished.
624    
625    John (Reppy) had critizised this as soon as I originally proposed and
626    implemented it, but at that time I did not really deeply believe
627    him. :) Now I came full-circle because I need the $<singlearc> syntax
628    in another place where it cannot be seen as an abbreviation for
629    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
630    seems to mean (i.e., it "expands" into the corresponding anchor
631    value).
632    
633    However, when paths are used as members in CM description files, it
634    continues to be true that there must be at least another arc after the
635    anchor.  This is now enforced separately during semantic analysis
636    (i.e., from a lexical/syntactical point of view, the notation is ok.)
637    
638    ** 2. The "cm" class now accepts an option "bind".  The option's value
639    is a sub-option list of precisely two items -- one labeled "anchor"
640    and the other one labeled "value".  As you might expect, "anchor" is
641    used to specify an anchor name to be bound, and "value" specifies what
642    the anchor is being bound to.
643    
644    The value must be a directory name and can be given in either standard
645    syntax (including the possibility that it is itself an anchored path)
646    or native syntax.
647    
648    Examples:
649    
650       foo.cm (bind:(anchor:bar value:$mystuff/bar))
651       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
652    
653    and so on.
654    
655    The meaning of this is that the .cm-file will be processed with an
656    augmented anchor environment where the given anchor(s) is/are bound to
657    the given values(s).
658    
659    The rationale for having this feature is this: Suppose you are trying
660    to use two different (already stable) libraries a.cm and b.cm (that
661    you perhaps didn't write yourself).  Further, suppose each of these
662    two libraries internally uses its own auxiliary library $aux/lib.cm.
663    Normally you would now have a problem because the anchor "lib" can not
664    be bound to more than one value globally.  Therefore, the project that
665    uses both a.cm and b.cm must locally redirect the anchor to some other
666    place:
667    
668       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
669       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
670    
671    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
672    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
673    
674    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
675    you when you do so at the time of CM.stabilize).  Therefore, you can
676    also use an anchored path as the value:
677    
678      a.cm (bind:(anchor:lib value:$a-lib))
679      b.cm (bind:(anchor:lib value:$b-lib))
680    
681    Now you can globally configure (using the usual CM.Anchor.anchor or
682    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
683    itself is always locally bound, setting it globally is no longer
684    meaningful or necessary (but it does not hurt either).  In fact, "lib"
685    can still be used as a global anchor for separate purposes.  As a
686    matter of fact, one can locally define "lib" in terms of a global
687    "lib":
688    
689      a.cm (bind:(anchor:lib value:$lib/a))
690      b.cm (bind:(anchor:lib value:$lib/b))
691    
692    ** 3: The encoding of path names has changed.  This affects the way
693    path names are shown in CM's progress report and also the internal
694    protocol encoding used for parallel make.
695    
696    The encoding now uses one or more ':'-separated segments.  Each
697    segments corresponds to a file that has been specified relative to the
698    file given by its preceding segment.  The first segment is either
699    relative to the CWD, absolute, or anchored.  Each segment itself is
700    basically a Unix pathname; all segments but the first are relative.
701    
702    Example:
703    
704       $foo/bar/baz.cm:a/b/c.sml
705    
706    This path denotes the file bar/a/b/c.sml relative to the directory
707    denoted by anchor "foo".  Notice that the encoding also includes
708    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
709    paths are resolved relative to the .cm-files directory, so baz.cm must
710    be ignored to get the "real" pathname.
711    
712    To make this fact more obvious, CM puts the names of such "virtual
713    arcs" into parentheses when they appear in progress reports. (No
714    parentheses will appear in the internal protocol encoding.)  Thus,
715    what you really see is:
716    
717      $foo/bar/(baz.cm):a/b/c.sml
718    
719    I find this notation to be much more informative than before.
720    
721    Another new feature of the encoding is that special characters
722    including parentheses, colons, (back)slashes, and white space are
723    written as \ddd (where ddd is the decimal encoding of the character).
724    
725    *** The CM manual still needs to be updated.
726    
727    ----------------------------------------------------------------------
728    Name: Allen Leung
729    Date: 2000/06/15 00:38:00
730    Tag: leunga-20000615-x86-peephole
731    
732    x86 Peephole fix by Fermin.  Affects c-- and moby only.
733    
734    ----------------------------------------------------------------------
735    Name: Matthias Blume
736    Date: 2000/06/12 11:40:00
737    Tag: blume-20000612-parmakefix
738    Description:
739    
740    More cleanup after changing the file naming scheme: This time I
741    repaired the parallel make mechanism for CMB.make which I broke earlier.
742    
743    ----------------------------------------------------------------------
744    Name: Allen Leung
745    Date: 2000/06/09 01:25:00
746    Tag: leunga-20000609-various
747    
748    None of these things should affect normal SML/NJ operations
749    
750    1. Peephole improvements provided by Fermin (c--)
751    2. New annotation DEFUSE for adding extra dependence (moby)
752    3. New X86 LOCK instructions (moby)
753    4. New machine description language for reservation tables (scheduling)
754    5. Fixes to various optimization/analysis modules (branch chaining, dominator
755       trees etc.)
756    6. I've changed the CM files so that they can work with versions
757       110.0.6, 110.25 and 110.28
758    
759    ----------------------------------------------------------------------
760    Name: Matthias Blume
761    Date: 2000/06/09 12:40:00
762    Tag: blume-20000609-log
763    Description:
764    
765    - Removed all(?) remaining RCS Log entries from sources.
766    
767    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
768      anchored paths).
769    
770    ----------------------------------------------------------------------
771    Name: Matthias Blume
772    Date: 2000/06/07 17:00:00 JST
773    Tag: blume-20000607-no-implicit-anchors
774    Description:
775    
776    1. This update changes the default setting for
777    CM.Control.implicit_anchors from true to false.  This means that
778    implicit anchors are no longer permitted by default.  I also tried to
779    make sure that nothing else still relies on implicit anchors.
780    (This is the next step on the schedule towards a CM that does not even
781    have the notion of implicit anchors anymore.)
782    
783    2. More CM manual updates.
784    
785    3. I managed to track down and fix the pickling bug I mentioned last
786    time.  Because of the previously existing workaround, this entails no
787    immediate practical changes.
788    
789    ----------------------------------------------------------------------
790    Name: Matthias Blume
791    Date: 2000/06/06 11:15:00 JST
792    Tag: blume-20000606-lazierpickle
793    Description:
794    
795    !!!! NEW BOOT FILES !!!!
796    
797    * The main purpose of this update is to make library pickles lazier in
798    order to reduce the initial space penalty for autoloading a library.
799    As a result, it is now possible to have $smlnj/compiler.cm
800    pre-registered.  This should take care of the many complaints or
801    inquiries about missing structure Compiler.  This required changes to
802    CM's internal data structures and small tweaks to some algorithms.
803    
804    As a neat additional effect, it is no longer necessary (for the sake
805    of lean heap image files) to distinguish between a "minimal" CM and a
806    "full" CM.  Now, there is only one CM (i.e., the "full" version:
807    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
808    interactive top level. ($smlnj/cm/minimal.cm is gone.)
809    
810    To make the life of compiler-hackers easier, "makeml" now also
811    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
812    words, after you bootstrap a new sml for the first time, you will not
813    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
814    around you will still have to do it, though.)
815    
816    * A second change consists of major updates to the CM manual.  There
817    are now several appendices with summary information and also a full
818    specification of the CM description file syntax.
819    
820    * In directory src/system I added the script "allcross".  This script
821    invokes sml and cross-compiles the compiler for all supported
822    architectures.  (Useful when providing a new set of boot files.)
823    
824    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
825    added a small tweak to pickle-util.sml to work around this problem,
826    but it is not a proper fix yet.  I will investigate further.  (The
827    effect of the bug was an inflation of library pickle size.)
828    
829    * Version number increased to 110.28.1 (to avoid compatibility problems).
830    
831    ----------------------------------------------------------------------
832    Name: Allen Leung
833    Date: 2000/05/25 17:28 EDT
834    Tag: leunga-20000525-ra
835    Description:
836    
837      Fixed a bug in freezing phase of the register allocator.
838    
839    ----------------------------------------------------------------------
840    Name: Allen Leung
841    Date: 2000/05/15 22:53 EDT
842    Tag: leunga-20000515-alpha-x86-ra
843    Description:
844    
845      1. Alpha
846    
847          Slight cleanup.  Removed the instruction SGNXL
848    
849      2. X86
850    
851          Added the following instructions to the instruction set:
852    
853            ROLx, RORx,
854            BTx, BTSx, BTLx, BTRx,
855            XCHGx, and variants with the LOCK prefix
856    
857      3. Register Allocation
858    
859          The module ra-rewrite-with-renaming has been improved.
860    
861      These have no effect on SML/NJ.
862    
863    ----------------------------------------------------------------------
864    Name: Matthias Blume
865    Date: 2000/05/15 16:20:00 JST
866    Tag: blume-20000515-lightrebuild
867    Description:
868    
869    1. I added an alternative to "-rebuild" to "makeml".  The difference is
870       that prior to calling CMB.make' the CM-variable "LIGHT" will be
871       defined.  In effect, the command will not build any cross-compiler
872       backends and therefore finish more quickly.
873    
874       The "fixpt" script also takes a "-light" switch to be able to use
875       this new facility while compiling for a fixpoint.
876    
877    2. I replaced all mentions of anchored paths in group owner specifications
878       with simple relative paths (usually starting with "..").
879       The rationale is that a library's internal workings should not be
880       compromised by the lack of some anchor.  (An anchor is necessary
881       for someone who wants to refer to the library by an anchored path,
882       but it should not be necessary to build the same library in the first
883       place.)
884    
885    3. I changed the way CM's tool mechanism determines the shell command
886       string used for things like ml-yacc etc. so that it does not break
887       when CM.Control.implicit_anchors is turned off.
888    
889    ----------------------------------------------------------------------
890    Name: Matthias Blume
891    Date: 2000/05/12 18:20:00 JST
892    Tag: blume-20000512-ml-build
893    Description:
894    
895    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
896    getting installed properly (by config/install.sh).
897    
898    ----------------------------------------------------------------------
899    Name: Matthias Blume
900    Date: 2000/05/12 17:30:00 JST
901    Tag: blume-20000512-anchors
902    Description:
903    
904    !!! NEW BOOT FILES !!!
905    
906    This change is in preparation of fading out support for "implicitly
907    anchored path names".  I went through all sources and used the
908    explicit (and relatively new) $-notation.  See system/README and the
909    CM manual for more info on this.
910    
911    I also modified the anchoring scheme for some things such as "smlnj",
912    "MLRISC", "cm", etc. to take advantage of the fact that explicit
913    anchors are more expressive: anchor name and first arc do not have to
914    coincide.  This entails the following user-visible change:
915    
916    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
917    particular, when you fire up sml with a command-line argument, say,
918    e.g.:
919    
920       sml '$smlnj/cmb.cm'
921    
922    At the ML toplevel prompt:
923    
924       CM.autoload "$smlnj/cmb.cm";
925    
926    There is also a new controller in CM.Control that can be used to turn
927    off all remaining support for implicit anchors by saying:
928    
929        CM.autoload "$smlnj/
930        #set CM.Control.implicit_anchors false;
931    
932    This causes CM to reject implicitly anchored paths.  This is (for the
933    time being) less permissive than the "final" version where there will
934    be no more such implicit anchors and relative paths will be just that:
935    relative.
936    
937    The next step (version after next version?) will be to make the
938    default for CM.Control.implicit_anchors false.  After the dust has
939    settled, I can then produce the "final" version of this...
940    
941    Note: Since bootstrapping is a bit tricky, I provided new boot files.
942    
943    ----------------------------------------------------------------------
944    Name: Matthias Blume
945    Date: 2000/05/11 16:30:00 JST
946    Tag: blume-20000511-sources
947    Description:
948    
949    The main change is that I added function CM.sources as a generalized
950    version of the earlier CM.makedepend.  This entails the following
951    additional changes:
952    
953      - CM.makedepend has been dropped.
954    
955      - CM manual has been updated.
956    
957      - TOOLS signature and API have been changed.
958    
959    ----------------------------------------------------------------------
960    Name: Allen Leung
961    Date: 2000/05/10 21:17 EDT
962    Tag: leunga-20000510-moby-c--ssa
963    Description:
964    
965      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
966    None of these affect SML/NJ.
967    
968    1. Register Allocation
969    
970        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
971           This module tries to remove local (i.e. basic block level) redundancies
972           during spilling.
973    
974        b. A new framework for performing region based register allocation.
975           Not yet entirely functional.
976    
977    2. X86
978    
979       a. DefUse for POP was missing the stack pointer [found by Lal]
980       b. Reload for CALL was incorrect in X86Spill [found by John]
981       c. Various fixes in X86Spill so that it can be used correctly for
982          the new spilling module.
983    
984    3. SSA/IR
985    
986       a. New module ir/dj-dataflow.sml implements elimination based
987          data flow analysis.
988    
989    4. MLRiscGen
990    
991       a. Fix for gc type annotation
992    
993    5. MDGen
994    
995       Various fixes for machine description -> ml code translation.  For ssa
996       only.
997    
998    ----------------------------------------------------------------------
999    Name: Allen Leung
1000    Date: 2000/05/08 22:17 EDT
1001    Tag: leunga-20000508-labexp
1002    Description:
1003    
1004      Fermin has found a few assembly problems with constant expressions
1005      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1006      which choke on dumb assemblers.  This is his fix.
1007    
1008    ----------------------------------------------------------------------
1009    Name: Dave MacQueen
1010    Date: 2000/04/09 14:00 EDT
1011    Tag: dbm-20000502-Version_110_28
1012    Description:
1013    
1014    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1015    
1016    2. Updated config/version to 110.28
1017    
1018    3. Updated config/srcarchiveurl
1019    
1020    3. New boot files!
1021       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1022    
1023  ----------------------------------------------------------------------  ----------------------------------------------------------------------
1024  Name: Matthias Blume  Name: Matthias Blume
1025  Date: 2000/05/01 19:05:00 JST  Date: 2000/05/01 19:05:00 JST
# Line 310  Line 1319 
1319    
1320  3.  Assembly  3.  Assembly
1321    
1322      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
1323      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
1324      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
1325    
# Line 333  Line 1342 
1342    
1343  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
1344  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
1345  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
1346  legal surface-syntax symbol).  legal surface-syntax symbol).
1347    
1348  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 343  Line 1352 
1352  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
1353    
1354  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1355  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
1356  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
1357  more details.  more details.
1358    
# Line 458  Line 1467 
1467     (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
1468     used).     used).
1469    
1470     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
1471     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
1472     Makefiles or Shell scripts.     Makefiles or Shell scripts.
1473    
1474     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 471  Line 1480 
1480    
1481     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
1482    
1483     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
1484     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
1485    
1486  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 605  Line 1614 
1614    
1615       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
1616       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
1617       testl is perferable to andl.       testl is preferable to andl.
1618    
1619  7. RA (x86 only)  7. RA (x86 only)
1620    
# Line 787  Line 1796 
1796    
1797  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
1798  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
1799  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
1800  explanation.  explanation.
1801    
1802  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
1803  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
1804  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
1805  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 834  Line 1843 
1843  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
1844  <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,
1845  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
1846  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
1847  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
1848  and, thus, saves a lot of time.  and, thus, saves a lot of time.
1849    
# Line 884  Line 1893 
1893    it from that remote directory.    it from that remote directory.
1894    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
1895    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
1896    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
1897    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
1898    
1899    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.634  
changed lines
  Added in v.710

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