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 587, Thu Mar 30 09:01:52 2000 UTC revision 703, Thu Sep 7 05:43:23 2000 UTC
# Line 11  Line 11 
11  Date:  Date:
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    ----------------------------------------------------------------------
15    Name: Matthias Blume
16    Date: 2000/09/07 14:45:00 JST
17    Tag: blume-20000907-cmerrmsg
18    Description:
19    
20    Small tweak to CM to avoid getting ML syntax error messages twice.
21    
22    ----------------------------------------------------------------------
23    Name: Matthias Blume
24    Date: 2000/08/31 18:00:00 JST
25    Tag: blume-20000831-cvsbootfiles
26    Description:
27    
28    New URL for boot files (because the 110.29 files on the BL server do
29    now work correctly with my updated install scripts for yacc and lex).
30    
31    ----------------------------------------------------------------------
32    Name: Matthias Blume
33    Date: 2000/08/08 12:33:00 JST
34    Tag: blume-20000808-manual
35    Description:
36    
37    Tiny update to CM manual.
38    
39    ----------------------------------------------------------------------
40    Name: Allen Leung
41    Date: 2000/08/7 19:31:00
42    Tag: leunga-20000807-a-whole-bunch-of-stuff
43    
44      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
45    and Mobdy.
46    
47    1.  x86
48    
49       a.  Fixes to peephole module by John and Dan.
50       b.  Assembly fix to SETcc by Allen.
51       c.  Fix to c-call by John.
52       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
53       e.  Instruction selection optimization to SETcc as suggested by John.
54    
55           For example,
56    
57            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
58    
59           should generate:
60    
61            MOVL a, x
62            SUBL b, x
63            SHRL 31, x
64    
65    2.  IR stuff
66    
67         A bunch of new DJ-graph related algorithms added.  These
68         speed up SSA construction.
69    
70    3.  SSA + Scheduling
71    
72         Added code for SSA and scheduling to the repository
73    
74    ----------------------------------------------------------------------
75    Name: Lal George
76    Date: 2000/07/27 11:53:14 EDT
77    
78    Tag: lal-20000727-linux-ppc
79    Description:
80    
81     Made changes to support Linux PPC.
82     p.s. I have confirmation that the 110.29 boot files work fine.
83    
84    ----------------------------------------------------------------------
85    Name: Matthias Blume
86    Date: 2000/07/27 17:40:00 JST
87    Tag: blume-20000727-scripts
88    Description:
89    
90    !!!! WARNING !!!!
91    You must recompile the runtime system!
92    !!!! WARNING !!!!
93    
94    This is basically another round of script-enhancements:
95    
96    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
97       and undefine CM preprocessor symbols.
98    
99    2. ml-build avoids generating a new heap image if it finds that the
100       existing one is still ok.  (The condition is that no ML file had to
101       be recompiled and all ML files are found to be older that the heap
102       file.)
103    
104       To make this work smoothly, I also hacked the runtime system as
105       well as SMLofNJ.SysInfo to get access to the heap image suffix
106       (.sparc-solaris, ...) that is currently being used.
107    
108       Moreover, the signature of CM.mk_standalone has changed.  See the
109       CM manual.
110    
111    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
112       CM manual for details.)
113    
114    4. More CM manual updates:
115        - all of the above has been documented.
116        - there is now a section describing the (CM-related) command line
117          arguments that are accepted by the "sml" command
118    
119    ----------------------------------------------------------------------
120    Name: Matthias Blume
121    Date: 2000/07/25 16:20:00 JST
122    Tag: blume-20000725-makedepend
123    Description:
124    
125    Added a script called ml-makedepend.  This can be used in makefiles
126    for Unix' make in a way very similar to the "makedepend" command for
127    C.
128    
129    The script internally uses function CM.sources.
130    
131    Synopsis:
132    
133        ml-makedepend [-f makefile] cmfile targetname
134    
135    The default for the makefile is "makefile" (or "Makefile" should
136    "makefile" not exist).
137    
138    ml-makedepend adds a cmfile/targetname-specific section to this
139    makefile (after removing the previous version of this section).  The
140    section contains a single dependency specification with targetname on
141    the LHS (targetname is an arbitrary name), and a list of files derived
142    from the cmfile on the RHS.  Some of the files on the RHS are
143    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
144    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
145    path names.  The makefile writer is responsible for making sure that
146    these variables have correct at the time "make" is invoked.
147    
148    ----------------------------------------------------------------------
149    Name: Matthias Blume
150    Date: 2000/07/22 23:30:00 JST
151    Tag: blume-20000722-urlupdate
152    Description:
153    
154    Changed BOOT and config/srcarchiveurl to point to BL server:
155    
156        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
157    
158    ----------------------------------------------------------------------
159    Name: Matthias Blume
160    Date: 2000/07/18 18:00:00 JST
161    Tag: blume-20000718-Version_110_29
162    Description:
163    
164    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
165    
166    2. Updated config/version to 110.29
167    
168    3. Updated config/srcarchiveurl
169    
170    3. New boot files!
171       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
172    
173    ----------------------------------------------------------------------
174    Name: Matthias Blume
175    Date: 2000/07/11 13:58:00 JST
176    Tag: blume-20000711-doctypo
177    Description:
178    
179    Fixed a few typos in CM manual.
180    
181    ----------------------------------------------------------------------
182    Name: Allen Leung
183    Date: 2000/06/15 00:38:00
184    Tag: leunga-20000704-sparc-x86
185    
186    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
187    2. fix to x86 RET bug [found by Dan Grossman]
188    3. sparc assembly bug fix for ticc instructions [found by Fermin]
189    
190       Affects c-- and moby only
191    
192    ----------------------------------------------------------------------
193    Name: Matthias Blume
194    Date: 2000/07/04 15:26:00
195    Tag: blume-20000704-trigger
196    Description:
197    
198    1. Improvements to CM manual.
199    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
200       of getting a back-trace.  The function, when called, raises an
201       internal exception which explicitly carries the full back-trace history,
202       so it is unaffected by any intervening handle-raise pairs ("trivial"
203       or not).  The interactive loop will print that history once it arrives
204       at top level.
205       Short of having all exceptions implicitly carry the full history, the
206       recommended way of using this facility is:
207         - compile your program with instrumentation "on"
208         - run it, when it raises an exception, look at the history
209         - if the history is "cut off" because of some handler, go and modify
210           your program so that it explicitly calls BTrace.trigger
211         - recompile (still instrumented), and rerun; look at the full history
212    
213    ----------------------------------------------------------------------
214    Name: Matthias Blume
215    Date: 2000/07/03 15:36:00 JST
216    Tag: blume-20000702-manual
217    Description:
218    
219    Small corrections and updates to CM manual.
220    
221    ----------------------------------------------------------------------
222    Name: Matthias Blume
223    Date: 2000/06/29 16:04:00 JST
224    Tag: blume-20000629-yacctool
225    Description:
226    
227    Changes:
228    
229    1. Class "mlyacc" now takes separate arguments to pass options to
230       generated .sml- and .sig-files independently.
231    2. Corresponding CM manual updates.
232    3. BTrace module now also reports call sites.  (However, for loop clusters
233       it only shows from where the cluster was entered.)  There are associated
234       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
235    
236    ----------------------------------------------------------------------
237    Name: Matthias Blume
238    Date: 2000/06/27 16:51:00 JST
239    Tag: blume-20000627-noweb
240    Description:
241    
242    Changes:
243    
244     1. Implemented "subdir" and "witness" options for noweb tool.
245        This caused some slight internal changes in CM's tool implementation.
246     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
247        some remaining issues from earlier path anchor changes.
248     3. Updated CM manual accordingly.
249    
250     4. Changed implementation of back-tracing so that I now consider it
251        ready for prime-time.
252    
253        In particular, you don't have to explicitly trigger the back-trace
254        anymore.  Instead, if you are running BTrace-instrumented code and
255        there is an uncaught exception (regardless of whether or not it was
256        raised in instrumented code), the top-level evalloop will print
257        the back-trace.
258    
259        Features:
260    
261          - Instrumented and uninstrumented code work together seemlessly.
262            (Of course, uninstrumented code is never mentioned in actual
263             back-traces.)
264    
265          - Asymptotic time- and space-complexity of instrumented code is
266            equal to that of uninstrumented code.  (This means that
267            tail-recursion is preserved by the instrumentation phase.)
268    
269          - Modules whose code has been instrumented in different sessions
270            work together without problem.
271    
272          - There is no penalty whatsoever on uninstrumented code.
273    
274          - There is no penalty on "raise" expressions, even in
275            instrumented code.
276    
277        A potential bug (or perhaps it is a feature, too):
278    
279          A back-trace reaches no further than the outermost instrumented
280          non-trivial "raise".  Here, a "trivial" raise is one that is the
281          sole RHS of a "handle" rule.  Thus, back-traces reach trough
282    
283               <exp> handle e => raise e
284    
285          and even
286    
287               <exp> handle Foo => raise Bar
288    
289          and, of course, through
290    
291               <exp> handle Foo => ...
292    
293         if the exception was not Foo.
294    
295         Back-traces always reach right through any un-instrumented code
296         including any of its "handle" expressions, trivial or not.
297    
298       To try this out, do the following:
299    
300         - Erase all existing binfiles for your program.
301           (You may keep binfiles for those modules where you think you
302            definitely don't need back-tracing.)
303         - Turn on back-trace instrumentation:
304              SMLofNJ.Internals.BTrace.mode (SOME true);
305         - Recompile your program.  (I.e., run "CM.make" or "use".)
306         - You may now turn instrumentation off again (if you want):
307              SMLofNJ.Internals.BTrace.mode (SOME false);
308         - Run your program as usual.  If it raises an exception that
309           reaches the interactive toplevel, then a back-trace will
310           automatically be printed.  After that, the toplevel loop
311           will print the exception history as usual.
312    
313    ----------------------------------------------------------------------
314    Name: Matthias Blume
315    Date: 2000/06/26 09:56:46 JST
316    Tag: blume-20000626-setup
317    Description:
318    
319    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
320          ML code before and after compiling a file (e.g., to set compiler
321          flags)
322    
323    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
324              - associated changes to btrace.sml (BTrace instrumentation pass)
325              - cleaner implementation of btimp.sml (BTrace tracing and report
326                module)
327    
328    CM manual: * new path encoding documented
329               * description of setup-parameter to "sml" added
330    
331    The biggest user-visible change to back-tracing is that it is no
332    longer necessary to compile all traced modules within the same
333    session.  (This was a real limitation.)
334    
335    ----------------------------------------------------------------------
336    Name: Matthias Blume
337    Date: 2000/06/24 12:40:00 JST
338    Tag: blume-20000624-startup
339    Description:
340    
341    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
342    bit too often -- to put it mildly. :)
343    
344    ----------------------------------------------------------------------
345    Name: Matthias Blume
346    Date: 2000/06/23 18:20:00 JST
347    Tag: blume-20000623-btrace
348    Description:
349    
350    This updates adds a backtrace facility to aid programmers in debugging
351    their programs.  This involves the following changes:
352    
353    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
354       keeping track of the current call stack.  When programs are compiled
355       in a special mode, the compiler will insert calls to these hooks
356       into the user program.
357       "Hook" means that it is possible for different implementations of
358       back-tracing to register themselves (at different times).
359    
360    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
361       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
362       It can be turned on using this call:
363         SMLofNJ.Internals.BTrace.mode (SOME true);
364       Turning it off again:
365         SMLofNJ.Internals.BTrace.mode (SOME false);
366       Querying the current status:
367         SMLofNJ.Internals.BTrace.mode NONE;
368       Annotated programs are about twice as big as normal ones, and they
369       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
370       where all hooks do nothing).  The slowdown with a plugin that is
371       actually useful (such as the one supplied by default) is even greater,
372       but in the case of the default plugin it is still only an constant
373       factor (amortized).
374    
375    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
376       with a sub-structure BTrace for controlling back-tracing.  In particular,
377       the above-mentioned function "mode" controls whether the annotation
378       phase is invoked by the compiler.  Another important function is
379       "trigger": when called it aborts the current execution and causes
380       the top-level loop to print a full back-trace.
381    
382    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
383       for back-tracing.  It keeps track of the dynamic call stack and in
384       addition to that it keeps a partial history at each "level" of that
385       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
386       d and b (at separate times, dynamically), then the report will show:
387    
388       GOTO   d
389             /c
390       GOTO  \b
391       CALL   a
392    
393       This shows that there was an initial non-tail call of a, then a
394       tail-call to b or c, looping behavior in a cluster of functions that
395       consist of b and c, and then a goto from that cluster (i.e., either from
396       b or from c) to d.
397    
398       Note that (depending on the user program) the amount of information
399       that the back-trace module has to keep track of at each level is bounded
400       by a constant.  Thus, the whole implementation has the same asymptotical
401       complexity as the original program (both in space and in time).
402    
403    5. compiler/TopLevel/interact/evalloop.sml has been modified to
404       handle the special exception SMLofNJ.Internals.BTrace.BTrace
405       which is raised by the "trigger" function mentioned above.
406    
407    Notes on usage:
408    
409    - Annotated code works well together with unannotated code:
410    Unannotated calls simply do not show up at all in the backtrace.
411    
412    - It is not a good idea to let modules that were annotated during
413    different sessions run at the same time.  This is because the compiler
414    chooses small integers to identify individual functions, and there
415    will be clashes if different modules were compiled in separate sessions.
416    (Nothing will crash, and you will even be told about the clashes, but
417    back-trace information will in general not be useful.)
418    
419    - Back-tracing can be confused by callcc and capture.
420    
421    - The only way of getting a back-trace right now is to explicitly
422    invoke the "trigger" function from your user program.  Eventually, we
423    should make every exception carry back-trace information (if
424    available).  But since this creates more overhead at "raise"-time
425    (similar to the current exnHistory overhead), I have not yet
426    implemented this.  (The implementation will be rather easy.)  With
427    exceptions carrying back-trace information, this facility will be even
428    more useful because users don't need to modify their programs...
429    
430    - While it is possible to compile the compiler with back-trace
431    annotations turned on (I did it to get some confidence in
432    correctness), you must make absolutely sure that core.sml and
433    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
434    be compiled with annotation because there is no core access yet, but
435    if you compile btimp.sml with annotation, then the system will go into
436    an infinite recursion and crash.)
437    Since CM currently does not know about BTrace, the only way to turn
438    annotations on and off for different modules of the compiler is to
439    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
440    this is awkward and clumsy.
441    
442    Sample sessions:
443    
444    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
445    - SMLofNJ.Internals.BTrace.mode (SOME true);
446    [autoloading]
447    [autoloading done]
448    val it = false : bool
449    - structure X = struct
450    -     fun main n = let
451    -         fun a (x, 0) = d x
452    -           | a (x, n) = b (x, n - 1)
453    -         and b (x, n) = c (x, n)
454    -         and c (x, n) = a (x, n)
455    -         and d x = e (x, 3)
456    -         and e (x, 0) = f x
457    -           | e (x, n) = e (x, n - 1)
458    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
459    -           | f n = n * g (n - 1)
460    -         and g n = a (n, 3)
461    -     in
462    -         f n
463    -     end
464    - end;
465    structure X : sig val main : int -> int end
466    - X.main 3;
467    *** BACK-TRACE ***
468    GOTO   stdIn:4.2-13.20: X.main[2].f
469    GOTO-( stdIn:4.2-13.20: X.main[2].e
470    GOTO   stdIn:4.2-13.20: X.main[2].d
471         / stdIn:4.2-13.20: X.main[2].a
472         | stdIn:4.2-13.20: X.main[2].b
473    GOTO-\ stdIn:4.2-13.20: X.main[2].c
474    CALL   stdIn:4.2-13.20: X.main[2].g
475    GOTO   stdIn:4.2-13.20: X.main[2].f
476    GOTO-( stdIn:4.2-13.20: X.main[2].e
477    GOTO   stdIn:4.2-13.20: X.main[2].d
478         / stdIn:4.2-13.20: X.main[2].a
479         | stdIn:4.2-13.20: X.main[2].b
480    GOTO-\ stdIn:4.2-13.20: X.main[2].c
481    CALL   stdIn:4.2-13.20: X.main[2].g
482    GOTO   stdIn:4.2-13.20: X.main[2].f
483    GOTO-( stdIn:4.2-13.20: X.main[2].e
484    GOTO   stdIn:4.2-13.20: X.main[2].d
485         / stdIn:4.2-13.20: X.main[2].a
486         | stdIn:4.2-13.20: X.main[2].b
487    GOTO-\ stdIn:4.2-13.20: X.main[2].c
488    CALL   stdIn:4.2-13.20: X.main[2].g
489    GOTO   stdIn:4.2-13.20: X.main[2].f
490    CALL   stdIn:2.15-17.4: X.main[2]
491    -
492    
493    (Note that because of a FLINt bug the above code currently does not
494    compile without BTrace turned on.)
495    
496    Here is another example, using my modified Tiger compiler:
497    
498    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
499    - SMLofNJ.Internals.BTrace.mode (SOME true);
500    [autoloading]
501    [autoloading done]
502    val it = false : bool
503    - CM.make "sources.cm";
504    [autoloading]
505    ...
506    [autoloading done]
507    [scanning sources.cm]
508    [parsing (sources.cm):parse.sml]
509    [creating directory CM/SKEL ...]
510    [parsing (sources.cm):tiger.lex.sml]
511    ...
512    [wrote CM/sparc-unix/semant.sml]
513    [compiling (sources.cm):main.sml]
514    [wrote CM/sparc-unix/main.sml]
515    [New bindings added.]
516    val it = true : bool
517    - Main.compile ("../testcases/merge.tig", "foo.out");
518    *** BACK-TRACE ***
519    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
520    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
521    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
522    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
523    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
524    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
525    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
526         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
527    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
528    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
529    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
530    CALL   main.sml:1.18-118.4: Main.compile[2]
531    -
532    
533    ----------------------------------------------------------------------
534    Name: Matthias Blumen
535    Date: 2000/06/21 18:00:00 JST
536    Tag: blume-20000621-manual
537    Description:
538    
539    CM manual update: Path environments documented.
540    
541    ----------------------------------------------------------------------
542    Name: Matthias Blume
543    Date: 2000/06/19 13:40:00
544    Tag: blume-20000619-manual
545    Description:
546    
547    CM manual and system/README update.  This only covers the fact that
548    there are no more implicit anchors.  (Path environments and the "bind"
549    option to "cm" have yet to be documented.)
550    
551    ----------------------------------------------------------------------
552    Name: Matthias Blume
553    Date: 2000/06/19 11:05:00 JST
554    Tag: blume-20000619-chdir-bugfix
555    Description:
556    
557    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
558    
559    ----------------------------------------------------------------------
560    Name: Matthias Blume
561    Date: 2000/06/18 22:00:10 JST
562    Tag: blume-20000618-implicit-anchors-really-gone
563    Description:
564    
565    I updates the previous HISTORY entry where I forgot to mention that
566    implicit anchors are no longer with us.
567    
568    The current update also gets rid of the (now useless) controller
569    CM.Control.implicit_anchors.
570    
571    ----------------------------------------------------------------------
572    Name: Matthias Blume
573    Date: 2000/06/16 17:30:00 JST
574    Tag: blume-20000616-anchorenv
575    Description:
576    
577    This patch implements the long anticipated (just kidding :) "anchor
578    environment" mechanism.  In the course of doing this, I also
579    re-implemented CM's internal "SrcPath" module from scratch.  The new
580    one should be more robust in certain boundary cases.  In any case, it
581    is a lot cleaner than its predecessor (IMHO).
582    
583    This time, although there is yet another boot file format change, I
584    kept the unpickler backward-compatible.  As a result, no new bootfiles
585    are necessary and bootstrapping is straightforward.  (You cannot read
586    new bootfiles into an old system, but the other way around is no
587    problem.)
588    
589    Visible changes:
590    
591    ** 0. Implicit path anchors (without the leading $-symbol) are no
592    longer recognized at all. This means that such path names are not
593    illegal either.  For example, the name basis.cm simply refers to a
594    local file called "basis.cm" (i.e, the name is an ordinary path
595    relative to .cm-files directory).  Or, to put it differently, only
596    names that start with $ are anchored paths.
597    
598    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
599    vanished.
600    
601    John (Reppy) had critizised this as soon as I originally proposed and
602    implemented it, but at that time I did not really deeply believe
603    him. :) Now I came full-circle because I need the $<singlearc> syntax
604    in another place where it cannot be seen as an abbreviation for
605    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
606    seems to mean (i.e., it "expands" into the corresponding anchor
607    value).
608    
609    However, when paths are used as members in CM description files, it
610    continues to be true that there must be at least another arc after the
611    anchor.  This is now enforced separately during semantic analysis
612    (i.e., from a lexical/syntactical point of view, the notation is ok.)
613    
614    ** 2. The "cm" class now accepts an option "bind".  The option's value
615    is a sub-option list of precisely two items -- one labeled "anchor"
616    and the other one labeled "value".  As you might expect, "anchor" is
617    used to specify an anchor name to be bound, and "value" specifies what
618    the anchor is being bound to.
619    
620    The value must be a directory name and can be given in either standard
621    syntax (including the possibility that it is itself an anchored path)
622    or native syntax.
623    
624    Examples:
625    
626       foo.cm (bind:(anchor:bar value:$mystuff/bar))
627       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
628    
629    and so on.
630    
631    The meaning of this is that the .cm-file will be processed with an
632    augmented anchor environment where the given anchor(s) is/are bound to
633    the given values(s).
634    
635    The rationale for having this feature is this: Suppose you are trying
636    to use two different (already stable) libraries a.cm and b.cm (that
637    you perhaps didn't write yourself).  Further, suppose each of these
638    two libraries internally uses its own auxiliary library $aux/lib.cm.
639    Normally you would now have a problem because the anchor "lib" can not
640    be bound to more than one value globally.  Therefore, the project that
641    uses both a.cm and b.cm must locally redirect the anchor to some other
642    place:
643    
644       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
645       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
646    
647    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
648    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
649    
650    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
651    you when you do so at the time of CM.stabilize).  Therefore, you can
652    also use an anchored path as the value:
653    
654      a.cm (bind:(anchor:lib value:$a-lib))
655      b.cm (bind:(anchor:lib value:$b-lib))
656    
657    Now you can globally configure (using the usual CM.Anchor.anchor or
658    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
659    itself is always locally bound, setting it globally is no longer
660    meaningful or necessary (but it does not hurt either).  In fact, "lib"
661    can still be used as a global anchor for separate purposes.  As a
662    matter of fact, one can locally define "lib" in terms of a global
663    "lib":
664    
665      a.cm (bind:(anchor:lib value:$lib/a))
666      b.cm (bind:(anchor:lib value:$lib/b))
667    
668    ** 3: The encoding of path names has changed.  This affects the way
669    path names are shown in CM's progress report and also the internal
670    protocol encoding used for parallel make.
671    
672    The encoding now uses one or more ':'-separated segments.  Each
673    segments corresponds to a file that has been specified relative to the
674    file given by its preceding segment.  The first segment is either
675    relative to the CWD, absolute, or anchored.  Each segment itself is
676    basically a Unix pathname; all segments but the first are relative.
677    
678    Example:
679    
680       $foo/bar/baz.cm:a/b/c.sml
681    
682    This path denotes the file bar/a/b/c.sml relative to the directory
683    denoted by anchor "foo".  Notice that the encoding also includes
684    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
685    paths are resolved relative to the .cm-files directory, so baz.cm must
686    be ignored to get the "real" pathname.
687    
688    To make this fact more obvious, CM puts the names of such "virtual
689    arcs" into parentheses when they appear in progress reports. (No
690    parentheses will appear in the internal protocol encoding.)  Thus,
691    what you really see is:
692    
693      $foo/bar/(baz.cm):a/b/c.sml
694    
695    I find this notation to be much more informative than before.
696    
697    Another new feature of the encoding is that special characters
698    including parentheses, colons, (back)slashes, and white space are
699    written as \ddd (where ddd is the decimal encoding of the character).
700    
701    *** The CM manual still needs to be updated.
702    
703    ----------------------------------------------------------------------
704    Name: Allen Leung
705    Date: 2000/06/15 00:38:00
706    Tag: leunga-20000615-x86-peephole
707    
708    x86 Peephole fix by Fermin.  Affects c-- and moby only.
709    
710    ----------------------------------------------------------------------
711    Name: Matthias Blume
712    Date: 2000/06/12 11:40:00
713    Tag: blume-20000612-parmakefix
714    Description:
715    
716    More cleanup after changing the file naming scheme: This time I
717    repaired the parallel make mechanism for CMB.make which I broke earlier.
718    
719    ----------------------------------------------------------------------
720    Name: Allen Leung
721    Date: 2000/06/09 01:25:00
722    Tag: leunga-20000609-various
723    
724    None of these things should affect normal SML/NJ operations
725    
726    1. Peephole improvements provided by Fermin (c--)
727    2. New annotation DEFUSE for adding extra dependence (moby)
728    3. New X86 LOCK instructions (moby)
729    4. New machine description language for reservation tables (scheduling)
730    5. Fixes to various optimization/analysis modules (branch chaining, dominator
731       trees etc.)
732    6. I've changed the CM files so that they can work with versions
733       110.0.6, 110.25 and 110.28
734    
735    ----------------------------------------------------------------------
736    Name: Matthias Blume
737    Date: 2000/06/09 12:40:00
738    Tag: blume-20000609-log
739    Description:
740    
741    - Removed all(?) remaining RCS Log entries from sources.
742    
743    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
744      anchored paths).
745    
746    ----------------------------------------------------------------------
747    Name: Matthias Blume
748    Date: 2000/06/07 17:00:00 JST
749    Tag: blume-20000607-no-implicit-anchors
750    Description:
751    
752    1. This update changes the default setting for
753    CM.Control.implicit_anchors from true to false.  This means that
754    implicit anchors are no longer permitted by default.  I also tried to
755    make sure that nothing else still relies on implicit anchors.
756    (This is the next step on the schedule towards a CM that does not even
757    have the notion of implicit anchors anymore.)
758    
759    2. More CM manual updates.
760    
761    3. I managed to track down and fix the pickling bug I mentioned last
762    time.  Because of the previously existing workaround, this entails no
763    immediate practical changes.
764    
765    ----------------------------------------------------------------------
766    Name: Matthias Blume
767    Date: 2000/06/06 11:15:00 JST
768    Tag: blume-20000606-lazierpickle
769    Description:
770    
771    !!!! NEW BOOT FILES !!!!
772    
773    * The main purpose of this update is to make library pickles lazier in
774    order to reduce the initial space penalty for autoloading a library.
775    As a result, it is now possible to have $smlnj/compiler.cm
776    pre-registered.  This should take care of the many complaints or
777    inquiries about missing structure Compiler.  This required changes to
778    CM's internal data structures and small tweaks to some algorithms.
779    
780    As a neat additional effect, it is no longer necessary (for the sake
781    of lean heap image files) to distinguish between a "minimal" CM and a
782    "full" CM.  Now, there is only one CM (i.e., the "full" version:
783    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
784    interactive top level. ($smlnj/cm/minimal.cm is gone.)
785    
786    To make the life of compiler-hackers easier, "makeml" now also
787    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
788    words, after you bootstrap a new sml for the first time, you will not
789    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
790    around you will still have to do it, though.)
791    
792    * A second change consists of major updates to the CM manual.  There
793    are now several appendices with summary information and also a full
794    specification of the CM description file syntax.
795    
796    * In directory src/system I added the script "allcross".  This script
797    invokes sml and cross-compiles the compiler for all supported
798    architectures.  (Useful when providing a new set of boot files.)
799    
800    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
801    added a small tweak to pickle-util.sml to work around this problem,
802    but it is not a proper fix yet.  I will investigate further.  (The
803    effect of the bug was an inflation of library pickle size.)
804    
805    * Version number increased to 110.28.1 (to avoid compatibility problems).
806    
807    ----------------------------------------------------------------------
808    Name: Allen Leung
809    Date: 2000/05/25 17:28 EDT
810    Tag: leunga-20000525-ra
811    Description:
812    
813      Fixed a bug in freezing phase of the register allocator.
814    
815    ----------------------------------------------------------------------
816    Name: Allen Leung
817    Date: 2000/05/15 22:53 EDT
818    Tag: leunga-20000515-alpha-x86-ra
819    Description:
820    
821      1. Alpha
822    
823          Slight cleanup.  Removed the instruction SGNXL
824    
825      2. X86
826    
827          Added the following instructions to the instruction set:
828    
829            ROLx, RORx,
830            BTx, BTSx, BTLx, BTRx,
831            XCHGx, and variants with the LOCK prefix
832    
833      3. Register Allocation
834    
835          The module ra-rewrite-with-renaming has been improved.
836    
837      These have no effect on SML/NJ.
838    
839    ----------------------------------------------------------------------
840    Name: Matthias Blume
841    Date: 2000/05/15 16:20:00 JST
842    Tag: blume-20000515-lightrebuild
843    Description:
844    
845    1. I added an alternative to "-rebuild" to "makeml".  The difference is
846       that prior to calling CMB.make' the CM-variable "LIGHT" will be
847       defined.  In effect, the command will not build any cross-compiler
848       backends and therefore finish more quickly.
849    
850       The "fixpt" script also takes a "-light" switch to be able to use
851       this new facility while compiling for a fixpoint.
852    
853    2. I replaced all mentions of anchored paths in group owner specifications
854       with simple relative paths (usually starting with "..").
855       The rationale is that a library's internal workings should not be
856       compromised by the lack of some anchor.  (An anchor is necessary
857       for someone who wants to refer to the library by an anchored path,
858       but it should not be necessary to build the same library in the first
859       place.)
860    
861    3. I changed the way CM's tool mechanism determines the shell command
862       string used for things like ml-yacc etc. so that it does not break
863       when CM.Control.implicit_anchors is turned off.
864    
865    ----------------------------------------------------------------------
866    Name: Matthias Blume
867    Date: 2000/05/12 18:20:00 JST
868    Tag: blume-20000512-ml-build
869    Description:
870    
871    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
872    getting installed properly (by config/install.sh).
873    
874    ----------------------------------------------------------------------
875    Name: Matthias Blume
876    Date: 2000/05/12 17:30:00 JST
877    Tag: blume-20000512-anchors
878    Description:
879    
880    !!! NEW BOOT FILES !!!
881    
882    This change is in preparation of fading out support for "implicitly
883    anchored path names".  I went through all sources and used the
884    explicit (and relatively new) $-notation.  See system/README and the
885    CM manual for more info on this.
886    
887    I also modified the anchoring scheme for some things such as "smlnj",
888    "MLRISC", "cm", etc. to take advantage of the fact that explicit
889    anchors are more expressive: anchor name and first arc do not have to
890    coincide.  This entails the following user-visible change:
891    
892    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
893    particular, when you fire up sml with a command-line argument, say,
894    e.g.:
895    
896       sml '$smlnj/cmb.cm'
897    
898    At the ML toplevel prompt:
899    
900       CM.autoload "$smlnj/cmb.cm";
901    
902    There is also a new controller in CM.Control that can be used to turn
903    off all remaining support for implicit anchors by saying:
904    
905        CM.autoload "$smlnj/
906        #set CM.Control.implicit_anchors false;
907    
908    This causes CM to reject implicitly anchored paths.  This is (for the
909    time being) less permissive than the "final" version where there will
910    be no more such implicit anchors and relative paths will be just that:
911    relative.
912    
913    The next step (version after next version?) will be to make the
914    default for CM.Control.implicit_anchors false.  After the dust has
915    settled, I can then produce the "final" version of this...
916    
917    Note: Since bootstrapping is a bit tricky, I provided new boot files.
918    
919    ----------------------------------------------------------------------
920    Name: Matthias Blume
921    Date: 2000/05/11 16:30:00 JST
922    Tag: blume-20000511-sources
923    Description:
924    
925    The main change is that I added function CM.sources as a generalized
926    version of the earlier CM.makedepend.  This entails the following
927    additional changes:
928    
929      - CM.makedepend has been dropped.
930    
931      - CM manual has been updated.
932    
933      - TOOLS signature and API have been changed.
934    
935    ----------------------------------------------------------------------
936    Name: Allen Leung
937    Date: 2000/05/10 21:17 EDT
938    Tag: leunga-20000510-moby-c--ssa
939    Description:
940    
941      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
942    None of these affect SML/NJ.
943    
944    1. Register Allocation
945    
946        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
947           This module tries to remove local (i.e. basic block level) redundancies
948           during spilling.
949    
950        b. A new framework for performing region based register allocation.
951           Not yet entirely functional.
952    
953    2. X86
954    
955       a. DefUse for POP was missing the stack pointer [found by Lal]
956       b. Reload for CALL was incorrect in X86Spill [found by John]
957       c. Various fixes in X86Spill so that it can be used correctly for
958          the new spilling module.
959    
960    3. SSA/IR
961    
962       a. New module ir/dj-dataflow.sml implements elimination based
963          data flow analysis.
964    
965    4. MLRiscGen
966    
967       a. Fix for gc type annotation
968    
969    5. MDGen
970    
971       Various fixes for machine description -> ml code translation.  For ssa
972       only.
973    
974    ----------------------------------------------------------------------
975    Name: Allen Leung
976    Date: 2000/05/08 22:17 EDT
977    Tag: leunga-20000508-labexp
978    Description:
979    
980      Fermin has found a few assembly problems with constant expressions
981      generated in LabelExp.  Mostly, the problems involve extra parentheses,
982      which choke on dumb assemblers.  This is his fix.
983    
984    ----------------------------------------------------------------------
985    Name: Dave MacQueen
986    Date: 2000/04/09 14:00 EDT
987    Tag: dbm-20000502-Version_110_28
988    Description:
989    
990    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
991    
992    2. Updated config/version to 110.28
993    
994    3. Updated config/srcarchiveurl
995    
996    3. New boot files!
997       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
998    
999    ----------------------------------------------------------------------
1000    Name: Matthias Blume
1001    Date: 2000/05/01 19:05:00 JST
1002    Tag: blume-20000501-noweb
1003    Description:
1004    
1005    A new noweb tool has been added.  The existing system is entirely
1006    unaffected by this, but some CM users have asked for renewed noweb
1007    support.  Everything is documented in the CM manual.
1008    
1009    New (plugin) libraries:
1010    
1011       noweb-tool.cm
1012       nw-ext.cm
1013    
1014    ----------------------------------------------------------------------
1015    Name: Dave MacQueen
1016    Date: 2000/04/30 12:40PM EDT
1017    Tag: dbm-20000430-bug_fixes
1018    Description:
1019    
1020    1. Fix for bug 1498
1021       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1022       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1023         added toRealArray function
1024       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1025         added check for tag Obj.RealArray to array printing case in ppObj
1026    
1027    2. Fix for bug 1510
1028       smlnj/src/compiler/Semant/types/typesutil.sml
1029         fixed definition of dummyargs (used by equalTycon) so that
1030         dummy args are distinct types
1031    
1032    ----------------------------------------------------------------------
1033    Name: Matthias Blume
1034    Date: 2000/04/30 01:00:00 JST
1035    Tag: blume-20000430-versions
1036    Description:
1037    
1038    1. CM version numbering added.  This is an implementation of Lal's
1039       proposal for adding version numbers and version checking to .cm
1040       files.  Lal said that his proposal was just that -- a proposal.
1041       For the time being I went ahead and implemented it so that people
1042       can comment on it.  Everything is completely backward-compatible
1043       (except for the stable library format, i.e., new bootfiles!).
1044    
1045       As usual, see the CM manual for details.
1046    
1047    2. An alternative syntax for anchored paths has been implemented.
1048       Dave has recently voiced the same concerns that I had when I did
1049       this, so there should be some support.  My take is that eventually
1050       I will let support for the current syntax (where anchors are
1051       "implicit") fade out in favor of the new, explicit syntax.
1052       In order to be backward-compatible, both old and new syntax are
1053       currently supported.
1054    
1055       Again, see the CM manual for details.
1056    
1057    3. Parallel make is trying to be slightly smarter:  When the master
1058       process finds a "bottleneck", i.e., when there is only one
1059       compilation unit that can be compiled and everybody else is
1060       waiting on it, then it will simply compile it directly instead
1061       of clumsily telling one of the slaves to do it.
1062    
1063    4. Support for "unsharing" added.  This is necessary in order to be
1064       able to have two different versions of the same library running
1065       at the same time (e.g., for trying out a new MLRISC while still
1066       having the old MLRISC linked into the current compiler, etc.)
1067       See the CM manual.
1068    
1069    5. Simple "makedepend" functionality added for generating Makefile
1070       dependency information.  (This is rather crude at the moment.
1071       Expect some changes here in the future.)
1072    
1073    6. ".fun" added as a recognized suffix for ML files. Also documented
1074       explicitly in the manual that the fallback behavior (unknown suffix
1075       -> ML file) is not an official feature!
1076    
1077    7. Small changes to the pickler for stable libraries.
1078    
1079    8. Several internal changes to CM (for cleanup/improvement).
1080    
1081    
1082    !!!! NEW BINFILES !!!!
1083    
1084    ----------------------------------------------------------------------
1085    Name: Matthias Blume
1086    Date: 2000/04/28 17:30:00 JST
1087    Tag: blume-20000428-pathconfig
1088    Description:
1089    
1090    1. I changed config/install.sh to remove duplicate entries from the
1091       lib/pathconfig file at the end.  Moreover, the final version of
1092       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1093       in src/system/installml.
1094    
1095    2. The config/install.sh script now consistently uses relative
1096       pathnames in lib/pathconfig whenever the anchor is in the lib
1097       directory.  (So far this was true for the libraries that come
1098       pre-compiled and bundled as part of the bootfiles but not for
1099       libraries that are compiled by the script itself.)
1100    
1101    ----------------------------------------------------------------------
1102    Name: Matthias Blume
1103    Date: 2000/04/26 13:10:00 JST
1104    Tag: blume-20000426-fun_suffix
1105    Description:
1106    
1107    Added ".fun" as a recognized file name suffix (for ML code).
1108    
1109    ----------------------------------------------------------------------
1110    Name: Allen Leung
1111    Date: 2000/04/25 17:00:00 EST
1112    Tag: leunga-20000425-alpha-ra
1113    Description:
1114    
1115    1. Alpha
1116    
1117        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1118    in C--.
1119    
1120    2. RA
1121    
1122       Added a flag "ra-dump-size" to print out the size of the flowgraph
1123       and the interference graph.
1124    
1125    ----------------------------------------------------------------------
1126    Name: Dave MacQueen
1127    Date: 2000/04/25/
1128    Tag: dbm-20000425-mlyacc_doc_examples
1129    Description:
1130      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1131      Updated all three examples in src/ml-yacc/examples to run
1132      under 110.* using CM.make.
1133    
1134    ----------------------------------------------------------------------
1135    Name: Allen Leung
1136    Date: 2000/04/20 23:04:00 EST
1137    Tag: leunga-20000420-ssa-c---stuff
1138    Description:
1139    
1140      This update synchronizes my repository with Yale's.  Most of these
1141    changes, however, do not affect SML/NJ at all (the RA is an exception).
1142    
1143    1. Register Allocator
1144    
1145       a. An improvement in the interference graph construction:
1146          Given a copy
1147    
1148                s <- t
1149    
1150          no interference edge between s and t is added for this definition of s.
1151    
1152       b. I've added two new spill heuristic modules that Fermin and I developed
1153          (in the new library RA.cm). These are unused in SML/NJ but maybe
1154          useful for others (Moby?)
1155    
1156    2. X86
1157    
1158       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1159    
1160    3. Alpha
1161    
1162       a. Added the BSR instruction and code generation that goes with it [C--]
1163       b. Other fixes too numerous to recount provided by Fermin [C--]
1164    
1165    4. Regmaps
1166    
1167       a. The regmaps are not initialized with the identity physical bindings
1168          at creation time.  This is unneeded.
1169    
1170    5. MLRISC Optimizations
1171    
1172       a. The DJ-Graph module can now compute the iterated dominance frontiers
1173          intersects with liveness incrementally in linear time! Woohoo!
1174          This is now used in my new SSA construction algorithm.
1175    
1176       b. THe branch reorganization module is now smarter about linear chains of
1177          basic blocks.
1178    
1179    
1180    ----------------------------------------------------------------------
1181    Name: Matthias Blume
1182    Date: 2000/04/12 13:52:00 JST
1183    Tag: blume_main_v110p27_1
1184    Description:
1185    
1186    Changed install.sh script to handle archive files without version number
1187    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1188    name of the boot file archive.
1189    
1190    ----------------------------------------------------------------------
1191    Name: Dave MacQueen
1192    Date: 2000/04/09 14:00 EDT
1193    Tag: dbm-20000410-Version_110_27
1194    Description:
1195    
1196    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1197    
1198    2. Updated src/config/version to 110.27
1199    
1200    3. New boot files!
1201    
1202    ----------------------------------------------------------------------
1203    Name: Allen Leung
1204    Date: 2000/04/09 19:09:00 EST
1205    Tag: leunga-20000409-misc
1206    Description:
1207    
1208    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1209    
1210    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1211    
1212    ----------------------------------------------------------------------
1213    Name: Stefan
1214    Date: 2000/04/07 10:00:00 EDT
1215    Tag: monnier-20000406-branch-handling
1216    Description:
1217    
1218    Improved handling of branches (mostly those generated from
1219    polymorphic equality), removed switchoff and changed the
1220    default optimization settings (more cpsopt and less flintopt).
1221    
1222    ----------------------------------------------------------------------
1223    Name: Allen Leung
1224    Date: 2000/04/06 01:30:00 EST
1225    Tag: leunga-20000406-peephole-x86-SSA-2
1226    Description:
1227    
1228       Forgot a few files.
1229    
1230    ----------------------------------------------------------------------
1231    Name: Allen Leung
1232    Date: 2000/04/06 00:36:00 EST
1233    Tag: leunga-20000406-peephole-x86-SSA
1234    Description:
1235    
1236    1.  New Peephole code
1237    
1238    2.  Minor improvement to X86 instruction selection
1239    
1240    3.  Various fixes to SSA and machine description -> code translator
1241    
1242    ----------------------------------------------------------------------
1243    Name: Matthias Blume
1244    Date: 2000/04/05 12:30:00 JST
1245    Tag: blume_main_v110p26p2_3
1246    Description:
1247    
1248    This update just merges three minor cosmetic updates to CM's sources
1249    to get ready for the 110.27 code freeze on Friday.  No functionality
1250    has changed.
1251    
1252    ----------------------------------------------------------------------
1253    Name: Allen Leung
1254    Date: 2000/04/04 19:39:00 EST
1255    Tag: leunga-20000404-x86-asm
1256    Description:
1257    
1258    1.  Fixed a problem in X86 assembly.
1259    
1260        Things like
1261    
1262           jmp %eax
1263           jmp (%eax)
1264    
1265        should be output as
1266    
1267           jmp *%eax
1268           jmp *(%eax)
1269    
1270    2.  Assembly output
1271    
1272          Added a new flag
1273    
1274              "asm-indent-copies" (default to false)
1275    
1276          When this flag is on, parallel copies will be indented an extra level.
1277    
1278    ----------------------------------------------------------------------
1279    Name: Allen Leung
1280    Date: 2000/04/04 03:18:00 EST
1281    Tag: leunga-20000404-C--Moby
1282    Description:
1283    
1284        All of these fixes are related to C--, Moby, and my own optimization
1285        stuff; so they shouldn't affect SML/NJ.
1286    
1287    1.  X86
1288    
1289        Various fixes related floating point, and extensions.
1290    
1291    2.  Alpha
1292    
1293        Some extra patterns related to loads with signed/zero extension
1294        provided by Fermin.
1295    
1296    3.  Assembly
1297    
1298        When generating assembly, resolve the value of client defined constants,
1299        instead of generating symbolic values.  This is controlled by the
1300        new flag "asm-resolve-constants", which is default to true.
1301    
1302    4.  Machine Descriptions
1303    
1304        a. The precedence parser was slightly broken when parsing infixr symbols.
1305        b. The type generalizing code had the bound variables reversed, resulting
1306           in a problem during arity raising.
1307        c. Various fixes in machine descriptions.
1308    
1309    ----------------------------------------------------------------------
1310    Name: Matthias Blume
1311    Date: 2000/04/03 16:05:00 JST
1312    Tag: blume_main_v110p26p2_2
1313    Description:
1314    
1315    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
1316    now done via the ordinary static environment that is context to each
1317    compilation unit.
1318    
1319    To this end, I arranged that instead of "structure Core" as "structure
1320    _Core" is bound in the pervasive environment.  Core access is done via
1321    _Core (which can never be accidentally rebound because _Core is not a
1322    legal surface-syntax symbol).
1323    
1324    The current solution is much cleaner because the core environment is
1325    now simply part of the pervasive environment which is part of every
1326    compilation unit's context anyway.  In particular, this eliminates all
1327    special-case handling that was necessary until now in order to deal
1328    with dynamic and symbolic parts of the core environment.
1329    
1330    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1331    compilation manager's bootstrap compiler (actually: in the "init group"
1332    handling).  See the comments in src/system/smlnj/init/init.cmi for
1333    more details.
1334    
1335    I also tried to track down all mentions of "Core" (as string argument
1336    to Symbol.strSymbol) in the compiler and replaced them with a
1337    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
1338    name appears in one place only.
1339    
1340    Binfile and bootfile format have not changed, but the switchover from
1341    the old "init.cmi" to the new one is a bit tricky, so I supplied new
1342    bootfiles anyway.
1343    
1344    ----------------------------------------------------------------------
1345    Name: Allen Leung
1346    Date: 2000/04/02 21:17:00 EST
1347    Tag: leunga-20000402-mltree
1348    Description:
1349    
1350       1. Renamed the constructor CALL in MLTREE by popular demand.
1351       2. Added a bunch of files from my repository.  These are currently
1352          used by other non-SMLNJ backends.
1353    
1354    ----------------------------------------------------------------------
1355    Name: Allen Leung
1356    Date: 2000/03/31 21:15:00 EST
1357    Tag: leunga-20000331-aliasing
1358    Description:
1359    
1360    This update contains a rewritten (and hopefully more correct) module
1361    for extracting aliasing information from CPS.
1362    
1363       To turn on this feature:
1364    
1365            Compiler.Control.CG.memDisambiguate := true
1366    
1367       To pretty print the region information with assembly
1368    
1369           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
1370    
1371       To control how many levels of aliasing information are printed, use:
1372    
1373           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
1374    
1375       The default of n is 3.
1376    
1377    ----------------------------------------------------------------------
1378    Name: David MacQueen
1379    Date: 2000/03/31 11:15:00 EST
1380    Tag: dbm-20000331-runtime_fix
1381    Description:
1382    
1383    This update contains:
1384    
1385    1. runtime/c-lib/c-libraries.c
1386       includes added in revision 1.2 caused compilation errors on hppa-hpux
1387    
1388    2. fix for bug 1556
1389       system/Basis/Implementation/NJ/internal-signals.sml
1390    
1391    ----------------------------------------------------------------------
1392    Name: Matthias Blume
1393    Date: 2000/03/31 18:00:00 JST
1394    Tag: blume_main_v110p26p2_1
1395    Description:
1396    
1397    This update contains:
1398    
1399    1. A small change to CM's handling of stable libraries:
1400       CM now maintains one "global" modmap that is used for all stable
1401       libraries.  The use of such a global modmap maximizes sharing and
1402       minimizes the need for re-traversing parts of environments during
1403       modmap construction.  (However, this has minor impact since modmap
1404       construction seems to account for just one percent or less of total
1405       compile time.)
1406    
1407    2. I added a "genmap" phase to the statistics.  This is where I got the
1408       "one percent" number (see above).
1409    
1410    3. CM's new tool parameter mechanism just became _even_ better. :)
1411       - The parser understands named parameters and recursive options.
1412       - The "make" and "shell" tools use these new features.
1413         (This makes it a lot easier to cascade these tools.)
1414       - There is a small syntax change: named parameters use a
1415    
1416           <name> : ( <option> ... )            or
1417           <name> : <string>
1418    
1419         syntax.  Previously, named parameters were implemented in an
1420         ad-hoc fashion by each tool individually (by parsing strings)
1421         and had the form
1422    
1423           <name>=<string>
1424    
1425       See the CM manual for a full description of these issues.
1426    
1427  ----------------------------------------------------------------------  ----------------------------------------------------------------------
1428  Name: Matthias Blume  Name: Matthias Blume
# Line 31  Line 1443 
1443     (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
1444     used).     used).
1445    
1446     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
1447     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
1448     Makefiles or Shell scripts.     Makefiles or Shell scripts.
1449    
1450     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 44  Line 1456 
1456    
1457     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
1458    
1459     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
1460     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
1461    
1462  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 178  Line 1590 
1590    
1591       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
1592       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
1593       testl is perferable to andl.       testl is preferable to andl.
1594    
1595  7. RA (x86 only)  7. RA (x86 only)
1596    
# Line 360  Line 1772 
1772    
1773  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
1774  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
1775  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
1776  explanation.  explanation.
1777    
1778  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
1779  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
1780  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
1781  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 407  Line 1819 
1819  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
1820  <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,
1821  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
1822  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
1823  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
1824  and, thus, saves a lot of time.  and, thus, saves a lot of time.
1825    
# Line 457  Line 1869 
1869    it from that remote directory.    it from that remote directory.
1870    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
1871    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
1872    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
1873    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
1874    
1875    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.587  
changed lines
  Added in v.703

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