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 593, Tue Apr 4 07:26:33 2000 UTC revision 838, Tue Jun 5 19:10:21 2001 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: 2001/06/05 15:10:00 EDT
18    Tag: blume-20010605-cm-index
19    Description:
20    
21    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
22    
23    1. CM can now generate "index files".  These are human-readable files
24       that list on a per-.cm-file basis each toplevel symbol defined or
25       imported.  The location of the index file for
26       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
27       To enable index-file generation, set CM.Control.generate_index to true
28       or export an environment-symbol: export CM_GENERATE_INDEX=true.
29    
30       The CM manual has been updated accordingly.
31    
32    2. I made some slight modifications to the c-calls API in MLRISC.
33    
34         a) There is now a callback to support saving/restoring of
35            dedicated but caller-save registers around the actual call
36            instruction.
37         b) One can optionally specify a comment-annotation for the
38            call instruction.
39    
40    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
41       (For example, the comment annotation shows the C prototype of
42        the function being called.)
43    
44    ----------------------------------------------------------------------
45    Name: Matthias Blume
46    Date: 2001/06/01 13:30:00 EDT
47    Tag: blume-20010601-nlffi-cleanup
48    Description:
49    
50    This is mostly a cleanup of MLFFI stuff:
51    
52       - some signature files have been put into a more exposed place
53       - the ugly 'f type parameter is gone (simplifies types tremendously!)
54       - ml-nlffigen changed accordingly
55       - tutorial updated
56    
57    Other changes:
58    
59       - author's affiliation in CM manual(s) updated
60       - some more recognized keywords added to Allen's sml.sty
61    
62    ----------------------------------------------------------------------
63    Name: Matthias Blume
64    Date: 2001/05/25 15:30:00 EDT
65    Tag: blume-20010525-iptr
66    Description:
67    
68      - put the official 110.33-README (as it appears on the ftp server) under
69        CVS
70      - fixed a small bug related to incomplete pointer types in
71        ml-nlffigen
72      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
73        (it does not need the 'f type parameter)
74    
75    ----------------------------------------------------------------------
76    Name: Matthias Blume
77    Date: 2001/05/23 14:30:00 EDT
78    Tag: Release_110_33
79    Description:
80    
81    New version number (110.33).  New bootfiles.
82    
83    ----------------------------------------------------------------------
84    Name: Matthias Blume
85    Date: 2001/05/22 18:06:00 EDT
86    Tag: blume-20010522-targets
87    Description:
88    
89    Made install.sh use file config/targets.customized if it exists, falling
90    back to config/targets if it doesn't.  This way one can have a customized
91    version of the targets file without touching the "real thing", thus
92    eliminating the constant fear of accidentally checking something bogus
93    back into the CVS repository...  (File config/targets.customized must
94    not be added to the repository!)
95    
96    ----------------------------------------------------------------------
97    Name: Matthias Blume
98    Date: 2001/05/22 16:30:00 EDT
99    Tag: blume-20010522-minitut
100    Description:
101    
102    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
103       struct returns.
104    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
105       incomplete, preliminary documentation for NLFFI.
106    
107    ----------------------------------------------------------------------
108    Name: Matthias Blume
109    Date: 2001/05/14 11:30:00 EDT
110    Tag: blume-20010514-script
111    Description:
112    
113    Some bugs in install script fixed.
114    
115    In addition to that I also made a slight change to the NLFFI API:
116    Functors generated by ml-nlffigen now take the dynamic library as a
117    straight functor argument, not as a suspended one.  (The original
118    functor code used to force the suspension right away anyway, so there
119    was nothing gained by this complication of the interface.)
120    
121    ----------------------------------------------------------------------
122    Name: Matthias Blume
123    Date: 2001/05/11 14:35:00 EDT
124    Tag: blume-20010511-ml-nlffi
125    Description:
126    
127    I finally took the plunge and added my new FFI code to the main
128    repository.  For x86-linux it is now ready for prime-time.
129    
130    There are two new subdirectories of "src":
131    
132      - ml-nlffi-lib:
133           The utility library for programs using the FFI interface.
134           Here is the implementation of $/c.cm and its associated low-level
135           partners $/c-int.cm and $/memory.cm.
136      - ml-nlffigen:
137           A stand-alone program for generating ML glue code from C source
138           code.
139    
140    Building ml-nlffigen requires $/ckit-lib.cm.
141    
142    The config/install.sh script has been updates to do the Right Thing
143    (hopefully).
144    
145    Notice that the source tree for the C-Kit will not be put under "src"
146    but directly under the installation root directory.  (This is the
147    structure that currently exists on the CVS server when you check out
148    module "sml".)  Fortunately, config/install.sh knows about this oddity.
149    
150    Bugs: No documentation yet.
151    
152    ----------------------------------------------------------------------
153    Name: Matthias Blume
154    Date: 2001/05/09 16:35:00 EDT
155    Tag: blume-20010509-cpscontract
156    Description:
157    
158    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
159    wrapper/unwrapper elimination did not decrement usage counts and some
160    dead variables got overlooked by the dead-up logic.)
161    
162    ----------------------------------------------------------------------
163    Name: Lal George
164    Date: 2001/05/08  17:26:09 EDT
165    Tag: george-20010508-omit-frameptr
166    Description:
167    
168    Changes to implement the omit-frame-pointer optimization to support
169    raw C calls. For now, there is only support on the Intel x86, but
170    other architectures will follow as more experience is gained with this.
171    
172    
173    ----------------------------------------------------------------------
174    Name: Matthias Blume
175    Date: 2001/05/07 14:40:00 EDT
176    Tag: blume-20010507-proxies
177    Description:
178    
179    I made into "proxy libraries" all libraries that qualify for such a
180    change.  (A qualifying library is a library that has another library or
181    groups as its sole member and repeats that member's export list
182    verbatim.  A proxy library avoids this repetition by omitting its export
183    list, effectively inheriting the list that its (only) member exports.
184    See the CM manual for more explanation.)
185    The main effect is that explicit export lists for these libraries
186    do not have to be kepts in sync, making maintenance a bit easier.
187    
188    I also added copyright notices to many .cm-files.
189    
190    Last but not least, I made a new set of bootfiles.
191    
192    ----------------------------------------------------------------------
193    Name: Matthias Blume
194    Date: 2001/05/04 17:00:00 EDT
195    Tag: blume-20010504-cm-lsplit
196    Description:
197    
198    0. John merged pending changes to $/smlnj-lib.cm
199    
200    1. Allen's previous change accidentally backed out of one of Lal's
201       earlier changes.  I undid this mistake (re-introducing Lal's change).
202    
203    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
204       within the compiler where applicable.  There is some code simplification
205       because of that.
206    
207    3. The "split" phase (in FLINT) is now part of the default list of phases.
208       Compiler.Control.LambdaSplitting.* can be used to globally control the
209       lambda-splitting (cross-module-inlining) engine.  In addition to that,
210       it can now also be controlled on a per-source basis: CM has been taught
211       a new tool parameter applicable to ML source files.
212    
213       - To turn lambda-splitting off completely:
214            local open Compiler.Control.LambdaSplitting in
215                val _ = set Off
216            end
217       - To make "no lambda-splitting" the global default (but allow per-source
218         overriding); this is the initial setting:
219            local open Compiler.Control.LambdaSplitting in
220                val _ = set (Default NONE)
221            end
222       - To make "lambda-splitting with aggressiveness a" the global default
223         (and allow per-source overriding):
224            local open Compiler.Control.LambdaSplitting in
225                val _ = set (Default (SOME a))
226            end
227    
228       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
229         write (in the respective .cm-file):
230            a.sml (lambdasplitting:off)
231       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
232            a.sml (lambdasplitting:on)
233       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
234         <a> is a decimal non-negative integer):
235            a.sml (lambdasplitting:<a>)
236       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
237            a.sml (lambdasplitting:infinity)
238       - To use the global default for a.sml:
239            a.sml (lambdasplitting:default)
240         or simply
241            a.sml
242    
243    ----------------------------------------------------------------------
244    Name: Allen Leung
245    Date: 2001/05/04 01:57:00 EDT
246    Tag: leunga-20010504-sync
247    Description:
248    
249      MLRISC features.
250    
251      1. Fix to CMPXCHG instructions.
252      2. Changed RA interface to allow annotations in callbacks.
253      3. Added a new method to the stream interface to allow annotations updates.
254    
255    ----------------------------------------------------------------------
256    Name: Matthias Blume
257    Date: 2001/05/01 11:45:00 EDT
258    Tag: blume-20010501-pcedittmp
259    Description:
260    
261    Changed install.sh to use the current working directory instead of
262    /usr/tmp for a temporary file (pcedittmp).  The previous choice
263    of /usr/tmp caused trouble with MacOS X because of file premission
264    problems.
265    
266    ----------------------------------------------------------------------
267    Name: Matthias Blume
268    Date: 2001/04/20 11:10:00 EDT
269    Tag: blume-20010420-inMLflag
270    Description:
271    
272     - added vp_limitPtrMask to vproc-state.h
273       (for use by the raw-C-calls mechanism to implement proper interrupt
274        handling)
275     - made the ML compiler aware of various data-structure offsets so it
276       can generate code for accessing the vp_inML flag and vp_limitPtrMask
277     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
278       raw C-calls
279    
280    ----------------------------------------------------------------------
281    Name: Lal George
282    Date: 2001/04/20 09:15:28 EDT
283    Tag: george-20010420-macosX
284    Description:
285    
286     - Changes to port to Mac OS X; Darwin.
287    
288     - In the process I found that sqrt was broken on the PPC, because the
289       fsqrt instruction is not implemented.
290    
291    ----------------------------------------------------------------------
292    Name: Matthias Blume
293    Date: 2001/04/18 12:45:00 EDT
294    Tag: blume-20010418-ccalls
295    Description:
296    
297     - fixed two off-by-4 errors in the x86-specific c-calls implementation
298       (this bug prevented structure arguments containing pointers from being
299        passed correctly)
300     - changed the raw-C-call code in mlriscGen.sml in such a way that
301       structure arguments are represented as a pointer to the beginning
302       of the structure (instead of having a series of synthesized arguments,
303       one for each structure member)
304    
305     - made makeml script's verbosity level configurable via environment
306       variable (MAKEML_VERBOSITY)
307    
308     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
309       in rawmem-x86.sml; we are now using the real thing
310    
311    ----------------------------------------------------------------------
312    Name: Matthias Blume
313    Date: 2001/03/22 16:25:00 EST
314    Tag: blume-20010322-bootfiles
315    Description:
316    
317    Created a new set of bootfiles (for your automatic installation convenience).
318    
319    ----------------------------------------------------------------------
320    Name: Matthias Blume
321    Date: 2001/03/22 15:10:00 EST
322    Tag: blume-20010322-rawmem-parcm
323    Description:
324    
325    1. All "raw memory access" primitives for the new FFI are implemented now
326       (at least on the x86).
327    2. Some further cleanup of CM's parallel make mechanism.
328    
329    ----------------------------------------------------------------------
330    Name: Matthias Blume
331    Date: 2001/03/19 17:53:00 EST
332    Tag: blume-20010319-parallel
333    Description:
334    
335    Parallel make (using compile servers) now works again.
336    
337    To this end, CM.stabilize and CMB.make have been modified to work in
338    two passes when compile servers are attached:
339       1. Compile everything, do not perform stabilization; this pass
340          uses compile servers
341       2. Stabilize everything; this pass does not use compile servers
342    If there are no compile servers, the two passes are combined into one
343    (as before).  Splitting the passes increases the inherent parallelism
344    in the dependency graph because the entire graph including all
345    libraries is available at the same time.  This, in turn, improves
346    server utilization.  The downside is that the master process will
347    have to do some extra work after compilation is done (because for
348    technical reasons it must re-read all the binfiles during stabilization).
349    
350    ----------------------------------------------------------------------
351    Name: Matthias Blume
352    Date: 2001/03/16 12:22:00 EST
353    Tag: blume-20010316-bootfiles
354    Description:
355    
356    Created a new set of bootfiles (for your automatic installation convenience).
357    
358    ----------------------------------------------------------------------
359    Name: Matthias Blume
360    Date: 2001/03/16 11:00:00 EST
361    Tag: blume-20010316-MLTREE-fixup
362    Description:
363    
364    This is a minor fixup for an (untagged) earlier commit by Allen.
365    (A file was missing).
366    
367    ----------------------------------------------------------------------
368    Name: Allen Leung
369    Date: Mon Mar  5 18:54:57 EST 2001
370    Tag: leunga-20010305-cut-support
371    
372    1. New support for alternative control-flow in MLTREE.
373       Currently we support
374    
375          FLOW_TO(CALL ...., [k1,...,kn])
376    
377       This is needed for 'cuts to' in C-- and try/handle-like constructs
378       in Moby
379    
380       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
381    
382    2. Register Allocator
383    
384       Changes in interface [from Fermin, John]
385    
386    3. Alpha 8-bit SLL support [Fermin]
387    
388    4. All architectures
389    
390       A new module (ClusterExpandCopies) for expanding parallel copies.
391    
392    ----------------------------------------------------------------------
393    Name: Allen Leung
394    Date: 2001/02/27 23:07:00 EST
395    Tag: leunga-20010227-minor-stuff
396    
397    1. Alpha bug fix for CMOVNE
398    2. Handle mltree COND(..,FCMP ...,...)
399    3. Bug fix in simplifier
400    
401    ----------------------------------------------------------------------
402    Name: Matthias Blume
403    Date: 2001/01/30 17:50:00 EST
404    Tag: blume-20010130-sync
405    Description:
406    
407    This is just a minor update to sync my devel branch with the main brach.
408    The only visible change is the addition of some README files.
409    
410    ----------------------------------------------------------------------
411    Name: Matthias Blume
412    Date: 2001/01/12 23:30:00 JST
413    Tag: blume-20010112-bootfiles
414    Description:
415    
416    Made a new set of bootfiles that goes with the current state of the
417    repository.
418    
419    ----------------------------------------------------------------------
420    Name: Matthias Blume
421    Date: 2001/01/12 21:20:00 JST
422    Tag: blume-20010112-sync
423    Description:
424    
425    I am just flushing out some minor changes that had accumulated in
426    my private branch in order to sync with the main tree.  (This is
427    mainly because I had CVS trouble when trying to merge _into_ my
428    private branch.)
429    
430    Most people should be completely unaffected by this.
431    
432    ----------------------------------------------------------------------
433    Name: Allen Leung
434    Date: Thu Jan 11 21:03:00 EST 2001
435    Tag: leunga-20010111-labexp=mltree
436    Description:
437    
438    1.  Removed the type LabelExp and replace it by MLTree.
439    2.  Rewritten mltree-simplify with the pattern matcher tool.
440    3.  There were some bugs in alpha code generator which would break
441        64-bit code generation.
442    4.  Redo the tools to generate code with the
443    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
444        are now generated by perl scripts.
445    
446    ----------------------------------------------------------------------
447    Name: Matthias Blume
448    Date: 2001/01/10 21:55:00 JST
449    Tag: blume-20010110-rcc
450    Description:
451    
452    The RCC stuff now seems to work (but only on the x86).
453    This required hacking of the c-calls interface (and -implementation) in
454    MLRISC.
455    
456    Normal compiler users should be unaffected.
457    
458    ----------------------------------------------------------------------
459    Name: Matthias Blume
460    Date: 2001/01/09 01:20:00 JST
461    Tag: blume-20010109-rcc
462    Description:
463    
464    This is a fairly big patch, flushing out a large number of pending
465    changes that I made to my development copy over the last couple of days.
466    
467    Of practical relevance at this moment is a workaround for a pickling
468    bug that Allen ran into the other day.  The cause of the bug itself is
469    still unknown and it might be hard to fix it properly, but the
470    workaround has some merits of its own (namely somewhat reducing pickling
471    overhead for certain libraries).  Therefore, I think this solution should
472    be satisfactory at this time.
473    
474    The rest of the changes (i.e., the vast majority) has to do with my
475    ongoing efforts of providing direct support for C function calls from
476    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
477    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
478    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
479    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
480    handle RCC.
481    
482    The last part (the changes to mlriscGen.sml) are still known to be
483    wrong on the x86 and not implemented on all other architectures.  But
484    the infrastructure is in place. I had to change a few functor
485    signatures in the backend to be able to route the CCalls interface
486    from MLRISC there, and I had to specialize the mltree type (on the
487    x86) to include the necessary extensions. (The extensions themselves
488    were already there and redy to go in MLRISC/x86).
489    
490    Everything should be very happy as soon as someone helps me with
491    mlriscGen.sml...
492    
493    In any case, nothing of this should matter to anyone as long as the
494    new primop is not being used (which is going to be the case unless you
495    find it where I hid it :). The rest of the compiler is completely
496    unaffected.
497    
498    ----------------------------------------------------------------------
499    Name: Matthias Blume
500    Date: 2001/01/05 00:30:00 JST
501    Tag: blume-20010105-primops
502    Description:
503    
504    Added some experimental support for work that I am doing right now.
505    These changes mostly concern added primops, but there is also a new
506    experimental C library in the runtime system (but currently not enabled
507    anywhere except on Linux/X86).
508    
509    In the course of adding primops (and playing with them), I discovered that
510    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
511    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
512    To recover, I made type information in INL_PRIM mandatory and changed
513    prim.sml as well as built-in.sml accordingly.  The InLine structure now
514    has complete, correct type information (i.e., no bottom types).
515    
516    Since all these changes mean that we need new binfiles, I also bumped the
517    version number to 110.32.1.
518    
519    ----------------------------------------------------------------------
520    Name: Matthias Blume
521    Date: 2000/12/30 22:10:00 JST
522    Tag: blume-20001230-various
523    Description:
524    
525    Added proxy libraries for MLRISC and let MLRISC libraries refer
526    to each other using path anchors.  (See CM manual for explanation.)
527    
528    Updated CM documentation.
529    
530    Fixed some bugs in CM.
531    
532    Implemented "proxy" libraries (= syntactic sugar for CM).
533    
534    Added "-quiet" option to makeml and changed runtime system accordingly.
535    
536    Added cleanup handler for exportML to reset timers and compiler stats.
537    
538    ----------------------------------------------------------------------
539    Name: Lal George
540    Date: 2000/12/22 22:22:58 EST 2000
541    Tag: Release_110_32
542    Description:
543    
544            Infinite precision used throughout MLRISC.
545            see MLRISC/mltree/machine-int.sig
546    
547    ----------------------------------------------------------------------
548    Name: Matthias Blume
549    Date: 2000/12/22 23:16:00 JST
550    Tag: blume-20001222-warn
551    Description:
552    
553    Corrected wording and formatting of some CM warning message which I
554    broke in my previous patch.
555    
556    ----------------------------------------------------------------------
557    Name: Matthias Blume
558    Date: 2000/12/22 21:20:00 JST
559    Tag: blume-20001222-anchorenv
560    Description:
561    
562    Fixed CM's handling of anchor environments in connection with CMB.make.
563    
564    ----------------------------------------------------------------------
565    Name: Matthias Blume
566    Date: 2000/12/22 13:15:00 JST
567    Tag: blume-20001222-cleanup
568    Description:
569    
570    Removed src/cm/ffi which does not (and did not) belong here.
571    
572    ----------------------------------------------------------------------
573    Name: Matthias Blume
574    Date: 2000/12/21 23:55:00 JST
575    Tag: blume-20001221-exn
576    Description:
577    
578    Probably most important: CM no longer silently swallows all exceptions
579    in the compiler.
580    Plus: some other minor CM changes.  For example, CM now reports some
581    sizes for generated binfiles (code, data, envpickle, lambdapickle).
582    
583    ----------------------------------------------------------------------
584    Name: Matthias Blume
585    Date: 2000/12/15 00:01:05 JST
586    Tag: blume-20001215-dirtool
587    Description:
588    
589    - "dir" tool added.
590    - improvements and cleanup to Tools structure
591    - documentation updates
592    
593    ----------------------------------------------------------------------
594    Name: Allen Leung
595    Date: Thu Dec 14 03:45:24 EST 2000
596    Description:
597    Tag:  leunga-20001214-int-inf
598    Description:
599    
600       In IntInf, added these standard functions, which are missing from our
601    implementation:
602    
603        andb : int * int -> int
604        xorb : int * int -> int
605        orb  : int * int -> int
606        notb : int -> int
607         <<   : int * word -> int
608        ~>>  : int * word -> int
609    
610       Not tested, I hope they are correct.
611    
612    ----------------------------------------------------------------------
613    Name: Allen Leung
614    Date: Fri Dec  8 19:23:26 EST 2000
615    Description:
616    Tag:  leunga-20001208-nowhere
617    Description:
618    
619      Slight improvements to the 'nowhere' tool to handle OR-patterns,
620    to generate better error messages etc.  Plus a brief manual.
621    
622    ----------------------------------------------------------------------
623    Name: Lal George
624    Date: 2000/12/08 09:54:02 EST 2000
625    Tag: Release_110_31
626    Description:
627    
628    - Version 110.31
629    ----------------------------------------------------------------------
630    Name: Allen Leung
631    Date: Thu Dec  7 22:01:04 EST 2000
632    Tag:  leunga-20001207-cell-monster-hack
633    Description:
634    
635    Major MLRISC internal changes.  Affect all clients.
636    Summary:
637    
638    1.  Type CELLS.cell = int is now replaced by a datatype.
639        As a result, the old regmap is now gone.  Almost all interfaces
640        in MLRISC change as a consequence.
641    
642    2.  A new brand version of machine description tool (v3.0) that generates
643        modules expecting the new interface.  The old version is removed.
644    
645    3.  The RA interface has been further abstracted into two new functors.
646        RISC_RA and X86RA.  These functors have much simpler interfaces.
647        [See also directory MLRISC/demo.]
648    
649    4.  Some other new source->source code generation tools are available:
650    
651        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
652        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
653           I use this tool to generate the peephole optimizers---with the new
654           cell type changes, peephole rules are becoming difficult to write
655           without conditional pattern matching.
656    
657    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
658        cover the entire MLRISC source tree so many things broke.
659    
660    6.  CM files have been moved to the subdirectory MLRISC/cm.
661        They are moved because there are a lot of them and they clutter up the
662        root dir.
663    
664    7.  More detailed documentation to come...
665    
666        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
667        a makeml -rebuild first.  This is because of other other
668        changes that Matthias has made (see below).
669    
670    
671    ----------------------------------------------------------------------
672    Name: Matthias Blume
673    Date: 2000/11/30 23:12:00 JST
674    Tag: blume-20001130-filereorg
675    Description:
676    
677    Some manual updates and some file reorganizations in CM.
678    
679    ----------------------------------------------------------------------
680    Name: Matthias Blume
681    Date: 2000/11/24 17:45:00 JST
682    Tag: blume-20001124-link
683    Description:
684    
685    Drastically improved link traversal code for the case that the dynamic
686    value was already loaded at bootstrap time.  As a result, CM and CMB
687    now both load blazingly fast -- even on a very slow machine.  Also,
688    memory consumption has been further reduced by this.
689    
690    Warning: The format of the PIDMAP file has changed.  THerefore, to
691    bootstrap you have to do this:
692    
693    1. Run CMB.make
694    2. Make a symbolic link for the boot directory:
695         ln -s sml.boot.ARCH-OS xxx
696    3. "Rebuild" the boot directory:
697         ./makeml -boot xxx -rebuild sml ; rm xxx
698    4. Boot normally:
699          ./makeml
700    
701    ----------------------------------------------------------------------
702    Name: Matthias Blume
703    Date: 2000/11/21 21:20:00 JST
704    Tag: blume-20001121-tools
705    Description:
706    
707    Continued hacking on autoloading problem -- with success this time.
708    Also changed tool-plugin mechanism.  See new CM manual.
709    
710    ----------------------------------------------------------------------
711    Name: Matthias Blume
712    Date: 2000/11/19 14:30:00 JST
713    Tag:  blume-20001119-autoload
714    Description:
715    
716    Some hacking to make autoloading faster.  Success for CMB, no success
717    so far for CM.  There is a reduced structure CM' that autoloads faster.
718    (This is a temporary, non-documented hack to be eliminated again when
719    the general problem is solved.)
720    
721    ----------------------------------------------------------------------
722    Name: Matthias Blume
723    Date: 2000/11/17 14:10:00 JST
724    Tag: blume-20001117-pickle-lib
725    Description:
726    
727    1. Eliminated comp-lib.cm
728    2. Made pickle-lib.cm
729    3. Eliminated all uses of intset.sml (from comp-lib.cm)
730    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
731       equivalent constructs from smlnj-lib.cm (INtHashTable).
732    5. Point 4. also goes for those uses of intmap.* in MLRISC.
733       Duplicated intmap modules thrown out.
734    6. Hunted down all duplicated SCC code and replaced it with
735       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
736    7. Rewrote Feedback module.
737    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
738       should be thrown out and equivalent modules from smlnj-lib.cm
739       should be used (IntRedBlackSet, IntListSet, ...).
740    
741    Confirmed that compiler compiles to fixpoint.
742    
743    ----------------------------------------------------------------------
744    Name: Allen Leung
745    Date: 2000/11/10 18:00:00
746    Tag: leunga-20001110-new-x86-fp
747    
748    A new x86 floating point code generator has been added.
749    By default this is turned off.  To turn this on, do:
750    
751        CM.autoload "$smlnj/compiler.cm";
752        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
753    
754    Changes:
755    
756    1.  Changed FTAN to FPTAN so that the assembly output is correct.
757    2.  Changed the extension callback for FTANGENT to generate:
758    
759              fptan
760              fstp  %st(0)
761        instead of
762              fptan
763              fstpl ftempmem
764    
765    3.  Numerous assembly fixes for x86.
766    
767    5.  Cleaned up the machine code output module x86/x86MC.sml and added
768        support for a whole bunch of instructions and addressing modes:
769    
770          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
771          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
772          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
773          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
774          fxch %st(n)
775          fld %st(n)
776          fst %st(n)
777          fst mem
778          fstp %st(n)
779          fucom %st(n)
780          fucomp %st(n)
781    
782        All these are now generated when the fast fp mode is turned on.
783    
784    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
785    
786    ----------------------------------------------------------------------
787    Name: Matthias Blume
788    Date: 2000/11/09 11:20:00 JST
789    Tag: blume-20001109-scc
790    Description:
791    
792    Eliminated some code duplication:
793    
794    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
795       (Otherwise the functor is useless.)
796    2. Used GraphSCCFn where SCCUtilFun was used previously.
797    3. Got rid of SCCUtilFun (in comp-lib.cm).
798    
799    ----------------------------------------------------------------------
800    Name: Lal George
801    Date: 2000/11/06 09:02:21 EST 2000
802    Tag: Release_110_30
803    Description:
804    
805    - Version 110.30
806    ----------------------------------------------------------------------
807    Name: Matthias Blume
808    Date: 2000/11/04 14:45:00
809    Tag: blume-20001104-mlbuild
810    Description:
811    
812    - Made ml-build faster on startup.
813    - Documentation fixes.
814    
815    ----------------------------------------------------------------------
816    Name: Matthias Blume
817    Date: 2000/11/02 17:00:00 JST
818    Tag: blume-20001102-condcomp
819    Description:
820    
821    - Small tweaks to pickler -- new BOOTFILES!
822    - Version bumped to 110.29.2.
823    - Added conditional compilation facility to init.cmi (see comment there).
824    ----------------------------------------------------------------------
825    Name: Allen Leung
826    Date: 2000/10/23 19:31:00
827    Tag: leunga-20001023-demo-ra
828    
829    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
830    2. Test programs for the graph library updated
831    3. Some new MLRISC demo programs added
832    
833    ----------------------------------------------------------------------
834    Name: Matthias Blume
835    Date: 2000/08/31 22:15:00 JST
836    Tag: blume-20001017-errmsg
837    Description:
838    
839    More error message grief: Where there used to be no messages, there
840    now were some that had bogus error regions.  Fixed.
841    
842    ----------------------------------------------------------------------
843    Name: Matthias Blume
844    Date: 2000/08/31 17:30:00 JST
845    Tag: blume-20001017-v110p29p1
846    Description:
847    
848    I made a version 110.29.1 with new bootfiles.
849    
850    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
851              CM documentation changes and a small bugfix in CM's error reporting.
852    
853    ----------------------------------------------------------------------
854    Name: Lal George
855    Date: 2000/09/27 14:42:35 EDT
856    Tag: george-20000927-nodestatus
857    Description:
858    
859    Changed the type of the nodestatus, so that:
860    
861            SPILLED(~1)             is now SPILLED
862            SPILLED(m) where m>=0   is now MEMREG(m)
863            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
864    
865    ----------------------------------------------------------------------
866    Name: Matthias Blume
867    Date: 2000/09/07 14:45:00 JST
868    Tag: blume-20000907-cmerrmsg
869    Description:
870    
871    Small tweak to CM to avoid getting ML syntax error messages twice.
872    
873    ----------------------------------------------------------------------
874    Name: Matthias Blume
875    Date: 2000/08/31 18:00:00 JST
876    Tag: blume-20000831-cvsbootfiles
877    Description:
878    
879    New URL for boot files (because the 110.29 files on the BL server do
880    now work correctly with my updated install scripts for yacc and lex).
881    
882    ----------------------------------------------------------------------
883    Name: Matthias Blume
884    Date: 2000/08/08 12:33:00 JST
885    Tag: blume-20000808-manual
886    Description:
887    
888    Tiny update to CM manual.
889    
890    ----------------------------------------------------------------------
891    Name: Allen Leung
892    Date: 2000/08/7 19:31:00
893    Tag: leunga-20000807-a-whole-bunch-of-stuff
894    
895      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
896    and Mobdy.
897    
898    1.  x86
899    
900       a.  Fixes to peephole module by John and Dan.
901       b.  Assembly fix to SETcc by Allen.
902       c.  Fix to c-call by John.
903       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
904       e.  Instruction selection optimization to SETcc as suggested by John.
905    
906           For example,
907    
908            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
909    
910           should generate:
911    
912            MOVL a, x
913            SUBL b, x
914            SHRL 31, x
915    
916    2.  IR stuff
917    
918         A bunch of new DJ-graph related algorithms added.  These
919         speed up SSA construction.
920    
921    3.  SSA + Scheduling
922    
923         Added code for SSA and scheduling to the repository
924    
925    ----------------------------------------------------------------------
926    Name: Lal George
927    Date: 2000/07/27 11:53:14 EDT
928    
929    Tag: lal-20000727-linux-ppc
930    Description:
931    
932     Made changes to support Linux PPC.
933     p.s. I have confirmation that the 110.29 boot files work fine.
934    
935    ----------------------------------------------------------------------
936    Name: Matthias Blume
937    Date: 2000/07/27 17:40:00 JST
938    Tag: blume-20000727-scripts
939    Description:
940    
941    !!!! WARNING !!!!
942    You must recompile the runtime system!
943    !!!! WARNING !!!!
944    
945    This is basically another round of script-enhancements:
946    
947    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
948       and undefine CM preprocessor symbols.
949    
950    2. ml-build avoids generating a new heap image if it finds that the
951       existing one is still ok.  (The condition is that no ML file had to
952       be recompiled and all ML files are found to be older that the heap
953       file.)
954    
955       To make this work smoothly, I also hacked the runtime system as
956       well as SMLofNJ.SysInfo to get access to the heap image suffix
957       (.sparc-solaris, ...) that is currently being used.
958    
959       Moreover, the signature of CM.mk_standalone has changed.  See the
960       CM manual.
961    
962    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
963       CM manual for details.)
964    
965    4. More CM manual updates:
966        - all of the above has been documented.
967        - there is now a section describing the (CM-related) command line
968          arguments that are accepted by the "sml" command
969    
970    ----------------------------------------------------------------------
971    Name: Matthias Blume
972    Date: 2000/07/25 16:20:00 JST
973    Tag: blume-20000725-makedepend
974    Description:
975    
976    Added a script called ml-makedepend.  This can be used in makefiles
977    for Unix' make in a way very similar to the "makedepend" command for
978    C.
979    
980    The script internally uses function CM.sources.
981    
982    Synopsis:
983    
984        ml-makedepend [-f makefile] cmfile targetname
985    
986    The default for the makefile is "makefile" (or "Makefile" should
987    "makefile" not exist).
988    
989    ml-makedepend adds a cmfile/targetname-specific section to this
990    makefile (after removing the previous version of this section).  The
991    section contains a single dependency specification with targetname on
992    the LHS (targetname is an arbitrary name), and a list of files derived
993    from the cmfile on the RHS.  Some of the files on the RHS are
994    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
995    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
996    path names.  The makefile writer is responsible for making sure that
997    these variables have correct at the time "make" is invoked.
998    
999    ----------------------------------------------------------------------
1000    Name: Matthias Blume
1001    Date: 2000/07/22 23:30:00 JST
1002    Tag: blume-20000722-urlupdate
1003    Description:
1004    
1005    Changed BOOT and config/srcarchiveurl to point to BL server:
1006    
1007        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
1008    
1009    ----------------------------------------------------------------------
1010    Name: Matthias Blume
1011    Date: 2000/07/18 18:00:00 JST
1012    Tag: blume-20000718-Version_110_29
1013    Description:
1014    
1015    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
1016    
1017    2. Updated config/version to 110.29
1018    
1019    3. Updated config/srcarchiveurl
1020    
1021    3. New boot files!
1022       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
1023    
1024    ----------------------------------------------------------------------
1025    Name: Matthias Blume
1026    Date: 2000/07/11 13:58:00 JST
1027    Tag: blume-20000711-doctypo
1028    Description:
1029    
1030    Fixed a few typos in CM manual.
1031    
1032    ----------------------------------------------------------------------
1033    Name: Allen Leung
1034    Date: 2000/06/15 00:38:00
1035    Tag: leunga-20000704-sparc-x86
1036    
1037    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
1038    2. fix to x86 RET bug [found by Dan Grossman]
1039    3. sparc assembly bug fix for ticc instructions [found by Fermin]
1040    
1041       Affects c-- and moby only
1042    
1043    ----------------------------------------------------------------------
1044    Name: Matthias Blume
1045    Date: 2000/07/04 15:26:00
1046    Tag: blume-20000704-trigger
1047    Description:
1048    
1049    1. Improvements to CM manual.
1050    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
1051       of getting a back-trace.  The function, when called, raises an
1052       internal exception which explicitly carries the full back-trace history,
1053       so it is unaffected by any intervening handle-raise pairs ("trivial"
1054       or not).  The interactive loop will print that history once it arrives
1055       at top level.
1056       Short of having all exceptions implicitly carry the full history, the
1057       recommended way of using this facility is:
1058         - compile your program with instrumentation "on"
1059         - run it, when it raises an exception, look at the history
1060         - if the history is "cut off" because of some handler, go and modify
1061           your program so that it explicitly calls BTrace.trigger
1062         - recompile (still instrumented), and rerun; look at the full history
1063    
1064    ----------------------------------------------------------------------
1065    Name: Matthias Blume
1066    Date: 2000/07/03 15:36:00 JST
1067    Tag: blume-20000702-manual
1068    Description:
1069    
1070    Small corrections and updates to CM manual.
1071    
1072    ----------------------------------------------------------------------
1073    Name: Matthias Blume
1074    Date: 2000/06/29 16:04:00 JST
1075    Tag: blume-20000629-yacctool
1076    Description:
1077    
1078    Changes:
1079    
1080    1. Class "mlyacc" now takes separate arguments to pass options to
1081       generated .sml- and .sig-files independently.
1082    2. Corresponding CM manual updates.
1083    3. BTrace module now also reports call sites.  (However, for loop clusters
1084       it only shows from where the cluster was entered.)  There are associated
1085       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1086    
1087    ----------------------------------------------------------------------
1088    Name: Matthias Blume
1089    Date: 2000/06/27 16:51:00 JST
1090    Tag: blume-20000627-noweb
1091    Description:
1092    
1093    Changes:
1094    
1095     1. Implemented "subdir" and "witness" options for noweb tool.
1096        This caused some slight internal changes in CM's tool implementation.
1097     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
1098        some remaining issues from earlier path anchor changes.
1099     3. Updated CM manual accordingly.
1100    
1101     4. Changed implementation of back-tracing so that I now consider it
1102        ready for prime-time.
1103    
1104        In particular, you don't have to explicitly trigger the back-trace
1105        anymore.  Instead, if you are running BTrace-instrumented code and
1106        there is an uncaught exception (regardless of whether or not it was
1107        raised in instrumented code), the top-level evalloop will print
1108        the back-trace.
1109    
1110        Features:
1111    
1112          - Instrumented and uninstrumented code work together seemlessly.
1113            (Of course, uninstrumented code is never mentioned in actual
1114             back-traces.)
1115    
1116          - Asymptotic time- and space-complexity of instrumented code is
1117            equal to that of uninstrumented code.  (This means that
1118            tail-recursion is preserved by the instrumentation phase.)
1119    
1120          - Modules whose code has been instrumented in different sessions
1121            work together without problem.
1122    
1123          - There is no penalty whatsoever on uninstrumented code.
1124    
1125          - There is no penalty on "raise" expressions, even in
1126            instrumented code.
1127    
1128        A potential bug (or perhaps it is a feature, too):
1129    
1130          A back-trace reaches no further than the outermost instrumented
1131          non-trivial "raise".  Here, a "trivial" raise is one that is the
1132          sole RHS of a "handle" rule.  Thus, back-traces reach trough
1133    
1134               <exp> handle e => raise e
1135    
1136          and even
1137    
1138               <exp> handle Foo => raise Bar
1139    
1140          and, of course, through
1141    
1142               <exp> handle Foo => ...
1143    
1144         if the exception was not Foo.
1145    
1146         Back-traces always reach right through any un-instrumented code
1147         including any of its "handle" expressions, trivial or not.
1148    
1149       To try this out, do the following:
1150    
1151         - Erase all existing binfiles for your program.
1152           (You may keep binfiles for those modules where you think you
1153            definitely don't need back-tracing.)
1154         - Turn on back-trace instrumentation:
1155              SMLofNJ.Internals.BTrace.mode (SOME true);
1156         - Recompile your program.  (I.e., run "CM.make" or "use".)
1157         - You may now turn instrumentation off again (if you want):
1158              SMLofNJ.Internals.BTrace.mode (SOME false);
1159         - Run your program as usual.  If it raises an exception that
1160           reaches the interactive toplevel, then a back-trace will
1161           automatically be printed.  After that, the toplevel loop
1162           will print the exception history as usual.
1163    
1164    ----------------------------------------------------------------------
1165    Name: Matthias Blume
1166    Date: 2000/06/26 09:56:46 JST
1167    Tag: blume-20000626-setup
1168    Description:
1169    
1170    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1171          ML code before and after compiling a file (e.g., to set compiler
1172          flags)
1173    
1174    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1175              - associated changes to btrace.sml (BTrace instrumentation pass)
1176              - cleaner implementation of btimp.sml (BTrace tracing and report
1177                module)
1178    
1179    CM manual: * new path encoding documented
1180               * description of setup-parameter to "sml" added
1181    
1182    The biggest user-visible change to back-tracing is that it is no
1183    longer necessary to compile all traced modules within the same
1184    session.  (This was a real limitation.)
1185    
1186    ----------------------------------------------------------------------
1187    Name: Matthias Blume
1188    Date: 2000/06/24 12:40:00 JST
1189    Tag: blume-20000624-startup
1190    Description:
1191    
1192    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1193    bit too often -- to put it mildly. :)
1194    
1195    ----------------------------------------------------------------------
1196    Name: Matthias Blume
1197    Date: 2000/06/23 18:20:00 JST
1198    Tag: blume-20000623-btrace
1199    Description:
1200    
1201    This updates adds a backtrace facility to aid programmers in debugging
1202    their programs.  This involves the following changes:
1203    
1204    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1205       keeping track of the current call stack.  When programs are compiled
1206       in a special mode, the compiler will insert calls to these hooks
1207       into the user program.
1208       "Hook" means that it is possible for different implementations of
1209       back-tracing to register themselves (at different times).
1210    
1211    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1212       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1213       It can be turned on using this call:
1214         SMLofNJ.Internals.BTrace.mode (SOME true);
1215       Turning it off again:
1216         SMLofNJ.Internals.BTrace.mode (SOME false);
1217       Querying the current status:
1218         SMLofNJ.Internals.BTrace.mode NONE;
1219       Annotated programs are about twice as big as normal ones, and they
1220       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1221       where all hooks do nothing).  The slowdown with a plugin that is
1222       actually useful (such as the one supplied by default) is even greater,
1223       but in the case of the default plugin it is still only an constant
1224       factor (amortized).
1225    
1226    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1227       with a sub-structure BTrace for controlling back-tracing.  In particular,
1228       the above-mentioned function "mode" controls whether the annotation
1229       phase is invoked by the compiler.  Another important function is
1230       "trigger": when called it aborts the current execution and causes
1231       the top-level loop to print a full back-trace.
1232    
1233    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1234       for back-tracing.  It keeps track of the dynamic call stack and in
1235       addition to that it keeps a partial history at each "level" of that
1236       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1237       d and b (at separate times, dynamically), then the report will show:
1238    
1239       GOTO   d
1240             /c
1241       GOTO  \b
1242       CALL   a
1243    
1244       This shows that there was an initial non-tail call of a, then a
1245       tail-call to b or c, looping behavior in a cluster of functions that
1246       consist of b and c, and then a goto from that cluster (i.e., either from
1247       b or from c) to d.
1248    
1249       Note that (depending on the user program) the amount of information
1250       that the back-trace module has to keep track of at each level is bounded
1251       by a constant.  Thus, the whole implementation has the same asymptotical
1252       complexity as the original program (both in space and in time).
1253    
1254    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1255       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1256       which is raised by the "trigger" function mentioned above.
1257    
1258    Notes on usage:
1259    
1260    - Annotated code works well together with unannotated code:
1261    Unannotated calls simply do not show up at all in the backtrace.
1262    
1263    - It is not a good idea to let modules that were annotated during
1264    different sessions run at the same time.  This is because the compiler
1265    chooses small integers to identify individual functions, and there
1266    will be clashes if different modules were compiled in separate sessions.
1267    (Nothing will crash, and you will even be told about the clashes, but
1268    back-trace information will in general not be useful.)
1269    
1270    - Back-tracing can be confused by callcc and capture.
1271    
1272    - The only way of getting a back-trace right now is to explicitly
1273    invoke the "trigger" function from your user program.  Eventually, we
1274    should make every exception carry back-trace information (if
1275    available).  But since this creates more overhead at "raise"-time
1276    (similar to the current exnHistory overhead), I have not yet
1277    implemented this.  (The implementation will be rather easy.)  With
1278    exceptions carrying back-trace information, this facility will be even
1279    more useful because users don't need to modify their programs...
1280    
1281    - While it is possible to compile the compiler with back-trace
1282    annotations turned on (I did it to get some confidence in
1283    correctness), you must make absolutely sure that core.sml and
1284    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1285    be compiled with annotation because there is no core access yet, but
1286    if you compile btimp.sml with annotation, then the system will go into
1287    an infinite recursion and crash.)
1288    Since CM currently does not know about BTrace, the only way to turn
1289    annotations on and off for different modules of the compiler is to
1290    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1291    this is awkward and clumsy.
1292    
1293    Sample sessions:
1294    
1295    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1296    - SMLofNJ.Internals.BTrace.mode (SOME true);
1297    [autoloading]
1298    [autoloading done]
1299    val it = false : bool
1300    - structure X = struct
1301    -     fun main n = let
1302    -         fun a (x, 0) = d x
1303    -           | a (x, n) = b (x, n - 1)
1304    -         and b (x, n) = c (x, n)
1305    -         and c (x, n) = a (x, n)
1306    -         and d x = e (x, 3)
1307    -         and e (x, 0) = f x
1308    -           | e (x, n) = e (x, n - 1)
1309    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1310    -           | f n = n * g (n - 1)
1311    -         and g n = a (n, 3)
1312    -     in
1313    -         f n
1314    -     end
1315    - end;
1316    structure X : sig val main : int -> int end
1317    - X.main 3;
1318    *** BACK-TRACE ***
1319    GOTO   stdIn:4.2-13.20: X.main[2].f
1320    GOTO-( stdIn:4.2-13.20: X.main[2].e
1321    GOTO   stdIn:4.2-13.20: X.main[2].d
1322         / stdIn:4.2-13.20: X.main[2].a
1323         | stdIn:4.2-13.20: X.main[2].b
1324    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1325    CALL   stdIn:4.2-13.20: X.main[2].g
1326    GOTO   stdIn:4.2-13.20: X.main[2].f
1327    GOTO-( stdIn:4.2-13.20: X.main[2].e
1328    GOTO   stdIn:4.2-13.20: X.main[2].d
1329         / stdIn:4.2-13.20: X.main[2].a
1330         | stdIn:4.2-13.20: X.main[2].b
1331    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1332    CALL   stdIn:4.2-13.20: X.main[2].g
1333    GOTO   stdIn:4.2-13.20: X.main[2].f
1334    GOTO-( stdIn:4.2-13.20: X.main[2].e
1335    GOTO   stdIn:4.2-13.20: X.main[2].d
1336         / stdIn:4.2-13.20: X.main[2].a
1337         | stdIn:4.2-13.20: X.main[2].b
1338    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1339    CALL   stdIn:4.2-13.20: X.main[2].g
1340    GOTO   stdIn:4.2-13.20: X.main[2].f
1341    CALL   stdIn:2.15-17.4: X.main[2]
1342    -
1343    
1344    (Note that because of a FLINt bug the above code currently does not
1345    compile without BTrace turned on.)
1346    
1347    Here is another example, using my modified Tiger compiler:
1348    
1349    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1350    - SMLofNJ.Internals.BTrace.mode (SOME true);
1351    [autoloading]
1352    [autoloading done]
1353    val it = false : bool
1354    - CM.make "sources.cm";
1355    [autoloading]
1356    ...
1357    [autoloading done]
1358    [scanning sources.cm]
1359    [parsing (sources.cm):parse.sml]
1360    [creating directory CM/SKEL ...]
1361    [parsing (sources.cm):tiger.lex.sml]
1362    ...
1363    [wrote CM/sparc-unix/semant.sml]
1364    [compiling (sources.cm):main.sml]
1365    [wrote CM/sparc-unix/main.sml]
1366    [New bindings added.]
1367    val it = true : bool
1368    - Main.compile ("../testcases/merge.tig", "foo.out");
1369    *** BACK-TRACE ***
1370    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
1371    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1372    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1373    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1374    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1375    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1376    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
1377         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
1378    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
1379    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1380    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
1381    CALL   main.sml:1.18-118.4: Main.compile[2]
1382    -
1383    
1384    ----------------------------------------------------------------------
1385    Name: Matthias Blumen
1386    Date: 2000/06/21 18:00:00 JST
1387    Tag: blume-20000621-manual
1388    Description:
1389    
1390    CM manual update: Path environments documented.
1391    
1392    ----------------------------------------------------------------------
1393    Name: Matthias Blume
1394    Date: 2000/06/19 13:40:00
1395    Tag: blume-20000619-manual
1396    Description:
1397    
1398    CM manual and system/README update.  This only covers the fact that
1399    there are no more implicit anchors.  (Path environments and the "bind"
1400    option to "cm" have yet to be documented.)
1401    
1402    ----------------------------------------------------------------------
1403    Name: Matthias Blume
1404    Date: 2000/06/19 11:05:00 JST
1405    Tag: blume-20000619-chdir-bugfix
1406    Description:
1407    
1408    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1409    
1410    ----------------------------------------------------------------------
1411    Name: Matthias Blume
1412    Date: 2000/06/18 22:00:10 JST
1413    Tag: blume-20000618-implicit-anchors-really-gone
1414    Description:
1415    
1416    I updates the previous HISTORY entry where I forgot to mention that
1417    implicit anchors are no longer with us.
1418    
1419    The current update also gets rid of the (now useless) controller
1420    CM.Control.implicit_anchors.
1421    
1422    ----------------------------------------------------------------------
1423    Name: Matthias Blume
1424    Date: 2000/06/16 17:30:00 JST
1425    Tag: blume-20000616-anchorenv
1426    Description:
1427    
1428    This patch implements the long anticipated (just kidding :) "anchor
1429    environment" mechanism.  In the course of doing this, I also
1430    re-implemented CM's internal "SrcPath" module from scratch.  The new
1431    one should be more robust in certain boundary cases.  In any case, it
1432    is a lot cleaner than its predecessor (IMHO).
1433    
1434    This time, although there is yet another boot file format change, I
1435    kept the unpickler backward-compatible.  As a result, no new bootfiles
1436    are necessary and bootstrapping is straightforward.  (You cannot read
1437    new bootfiles into an old system, but the other way around is no
1438    problem.)
1439    
1440    Visible changes:
1441    
1442    ** 0. Implicit path anchors (without the leading $-symbol) are no
1443    longer recognized at all. This means that such path names are not
1444    illegal either.  For example, the name basis.cm simply refers to a
1445    local file called "basis.cm" (i.e, the name is an ordinary path
1446    relative to .cm-files directory).  Or, to put it differently, only
1447    names that start with $ are anchored paths.
1448    
1449    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1450    vanished.
1451    
1452    John (Reppy) had critizised this as soon as I originally proposed and
1453    implemented it, but at that time I did not really deeply believe
1454    him. :) Now I came full-circle because I need the $<singlearc> syntax
1455    in another place where it cannot be seen as an abbreviation for
1456    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1457    seems to mean (i.e., it "expands" into the corresponding anchor
1458    value).
1459    
1460    However, when paths are used as members in CM description files, it
1461    continues to be true that there must be at least another arc after the
1462    anchor.  This is now enforced separately during semantic analysis
1463    (i.e., from a lexical/syntactical point of view, the notation is ok.)
1464    
1465    ** 2. The "cm" class now accepts an option "bind".  The option's value
1466    is a sub-option list of precisely two items -- one labeled "anchor"
1467    and the other one labeled "value".  As you might expect, "anchor" is
1468    used to specify an anchor name to be bound, and "value" specifies what
1469    the anchor is being bound to.
1470    
1471    The value must be a directory name and can be given in either standard
1472    syntax (including the possibility that it is itself an anchored path)
1473    or native syntax.
1474    
1475    Examples:
1476    
1477       foo.cm (bind:(anchor:bar value:$mystuff/bar))
1478       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1479    
1480    and so on.
1481    
1482    The meaning of this is that the .cm-file will be processed with an
1483    augmented anchor environment where the given anchor(s) is/are bound to
1484    the given values(s).
1485    
1486    The rationale for having this feature is this: Suppose you are trying
1487    to use two different (already stable) libraries a.cm and b.cm (that
1488    you perhaps didn't write yourself).  Further, suppose each of these
1489    two libraries internally uses its own auxiliary library $aux/lib.cm.
1490    Normally you would now have a problem because the anchor "lib" can not
1491    be bound to more than one value globally.  Therefore, the project that
1492    uses both a.cm and b.cm must locally redirect the anchor to some other
1493    place:
1494    
1495       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1496       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1497    
1498    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1499    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
1500    
1501    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1502    you when you do so at the time of CM.stabilize).  Therefore, you can
1503    also use an anchored path as the value:
1504    
1505      a.cm (bind:(anchor:lib value:$a-lib))
1506      b.cm (bind:(anchor:lib value:$b-lib))
1507    
1508    Now you can globally configure (using the usual CM.Anchor.anchor or
1509    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1510    itself is always locally bound, setting it globally is no longer
1511    meaningful or necessary (but it does not hurt either).  In fact, "lib"
1512    can still be used as a global anchor for separate purposes.  As a
1513    matter of fact, one can locally define "lib" in terms of a global
1514    "lib":
1515    
1516      a.cm (bind:(anchor:lib value:$lib/a))
1517      b.cm (bind:(anchor:lib value:$lib/b))
1518    
1519    ** 3: The encoding of path names has changed.  This affects the way
1520    path names are shown in CM's progress report and also the internal
1521    protocol encoding used for parallel make.
1522    
1523    The encoding now uses one or more ':'-separated segments.  Each
1524    segments corresponds to a file that has been specified relative to the
1525    file given by its preceding segment.  The first segment is either
1526    relative to the CWD, absolute, or anchored.  Each segment itself is
1527    basically a Unix pathname; all segments but the first are relative.
1528    
1529    Example:
1530    
1531       $foo/bar/baz.cm:a/b/c.sml
1532    
1533    This path denotes the file bar/a/b/c.sml relative to the directory
1534    denoted by anchor "foo".  Notice that the encoding also includes
1535    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1536    paths are resolved relative to the .cm-files directory, so baz.cm must
1537    be ignored to get the "real" pathname.
1538    
1539    To make this fact more obvious, CM puts the names of such "virtual
1540    arcs" into parentheses when they appear in progress reports. (No
1541    parentheses will appear in the internal protocol encoding.)  Thus,
1542    what you really see is:
1543    
1544      $foo/bar/(baz.cm):a/b/c.sml
1545    
1546    I find this notation to be much more informative than before.
1547    
1548    Another new feature of the encoding is that special characters
1549    including parentheses, colons, (back)slashes, and white space are
1550    written as \ddd (where ddd is the decimal encoding of the character).
1551    
1552    *** The CM manual still needs to be updated.
1553    
1554    ----------------------------------------------------------------------
1555    Name: Allen Leung
1556    Date: 2000/06/15 00:38:00
1557    Tag: leunga-20000615-x86-peephole
1558    
1559    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1560    
1561    ----------------------------------------------------------------------
1562    Name: Matthias Blume
1563    Date: 2000/06/12 11:40:00
1564    Tag: blume-20000612-parmakefix
1565    Description:
1566    
1567    More cleanup after changing the file naming scheme: This time I
1568    repaired the parallel make mechanism for CMB.make which I broke earlier.
1569    
1570    ----------------------------------------------------------------------
1571    Name: Allen Leung
1572    Date: 2000/06/09 01:25:00
1573    Tag: leunga-20000609-various
1574    
1575    None of these things should affect normal SML/NJ operations
1576    
1577    1. Peephole improvements provided by Fermin (c--)
1578    2. New annotation DEFUSE for adding extra dependence (moby)
1579    3. New X86 LOCK instructions (moby)
1580    4. New machine description language for reservation tables (scheduling)
1581    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1582       trees etc.)
1583    6. I've changed the CM files so that they can work with versions
1584       110.0.6, 110.25 and 110.28
1585    
1586    ----------------------------------------------------------------------
1587    Name: Matthias Blume
1588    Date: 2000/06/09 12:40:00
1589    Tag: blume-20000609-log
1590    Description:
1591    
1592    - Removed all(?) remaining RCS Log entries from sources.
1593    
1594    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1595      anchored paths).
1596    
1597    ----------------------------------------------------------------------
1598    Name: Matthias Blume
1599    Date: 2000/06/07 17:00:00 JST
1600    Tag: blume-20000607-no-implicit-anchors
1601    Description:
1602    
1603    1. This update changes the default setting for
1604    CM.Control.implicit_anchors from true to false.  This means that
1605    implicit anchors are no longer permitted by default.  I also tried to
1606    make sure that nothing else still relies on implicit anchors.
1607    (This is the next step on the schedule towards a CM that does not even
1608    have the notion of implicit anchors anymore.)
1609    
1610    2. More CM manual updates.
1611    
1612    3. I managed to track down and fix the pickling bug I mentioned last
1613    time.  Because of the previously existing workaround, this entails no
1614    immediate practical changes.
1615    
1616    ----------------------------------------------------------------------
1617    Name: Matthias Blume
1618    Date: 2000/06/06 11:15:00 JST
1619    Tag: blume-20000606-lazierpickle
1620    Description:
1621    
1622    !!!! NEW BOOT FILES !!!!
1623    
1624    * The main purpose of this update is to make library pickles lazier in
1625    order to reduce the initial space penalty for autoloading a library.
1626    As a result, it is now possible to have $smlnj/compiler.cm
1627    pre-registered.  This should take care of the many complaints or
1628    inquiries about missing structure Compiler.  This required changes to
1629    CM's internal data structures and small tweaks to some algorithms.
1630    
1631    As a neat additional effect, it is no longer necessary (for the sake
1632    of lean heap image files) to distinguish between a "minimal" CM and a
1633    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1634    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1635    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1636    
1637    To make the life of compiler-hackers easier, "makeml" now also
1638    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1639    words, after you bootstrap a new sml for the first time, you will not
1640    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1641    around you will still have to do it, though.)
1642    
1643    * A second change consists of major updates to the CM manual.  There
1644    are now several appendices with summary information and also a full
1645    specification of the CM description file syntax.
1646    
1647    * In directory src/system I added the script "allcross".  This script
1648    invokes sml and cross-compiles the compiler for all supported
1649    architectures.  (Useful when providing a new set of boot files.)
1650    
1651    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1652    added a small tweak to pickle-util.sml to work around this problem,
1653    but it is not a proper fix yet.  I will investigate further.  (The
1654    effect of the bug was an inflation of library pickle size.)
1655    
1656    * Version number increased to 110.28.1 (to avoid compatibility problems).
1657    
1658    ----------------------------------------------------------------------
1659    Name: Allen Leung
1660    Date: 2000/05/25 17:28 EDT
1661    Tag: leunga-20000525-ra
1662    Description:
1663    
1664      Fixed a bug in freezing phase of the register allocator.
1665    
1666    ----------------------------------------------------------------------
1667    Name: Allen Leung
1668    Date: 2000/05/15 22:53 EDT
1669    Tag: leunga-20000515-alpha-x86-ra
1670    Description:
1671    
1672      1. Alpha
1673    
1674          Slight cleanup.  Removed the instruction SGNXL
1675    
1676      2. X86
1677    
1678          Added the following instructions to the instruction set:
1679    
1680            ROLx, RORx,
1681            BTx, BTSx, BTLx, BTRx,
1682            XCHGx, and variants with the LOCK prefix
1683    
1684      3. Register Allocation
1685    
1686          The module ra-rewrite-with-renaming has been improved.
1687    
1688      These have no effect on SML/NJ.
1689    
1690    ----------------------------------------------------------------------
1691    Name: Matthias Blume
1692    Date: 2000/05/15 16:20:00 JST
1693    Tag: blume-20000515-lightrebuild
1694    Description:
1695    
1696    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1697       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1698       defined.  In effect, the command will not build any cross-compiler
1699       backends and therefore finish more quickly.
1700    
1701       The "fixpt" script also takes a "-light" switch to be able to use
1702       this new facility while compiling for a fixpoint.
1703    
1704    2. I replaced all mentions of anchored paths in group owner specifications
1705       with simple relative paths (usually starting with "..").
1706       The rationale is that a library's internal workings should not be
1707       compromised by the lack of some anchor.  (An anchor is necessary
1708       for someone who wants to refer to the library by an anchored path,
1709       but it should not be necessary to build the same library in the first
1710       place.)
1711    
1712    3. I changed the way CM's tool mechanism determines the shell command
1713       string used for things like ml-yacc etc. so that it does not break
1714       when CM.Control.implicit_anchors is turned off.
1715    
1716    ----------------------------------------------------------------------
1717    Name: Matthias Blume
1718    Date: 2000/05/12 18:20:00 JST
1719    Tag: blume-20000512-ml-build
1720    Description:
1721    
1722    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1723    getting installed properly (by config/install.sh).
1724    
1725    ----------------------------------------------------------------------
1726    Name: Matthias Blume
1727    Date: 2000/05/12 17:30:00 JST
1728    Tag: blume-20000512-anchors
1729    Description:
1730    
1731    !!! NEW BOOT FILES !!!
1732    
1733    This change is in preparation of fading out support for "implicitly
1734    anchored path names".  I went through all sources and used the
1735    explicit (and relatively new) $-notation.  See system/README and the
1736    CM manual for more info on this.
1737    
1738    I also modified the anchoring scheme for some things such as "smlnj",
1739    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1740    anchors are more expressive: anchor name and first arc do not have to
1741    coincide.  This entails the following user-visible change:
1742    
1743    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1744    particular, when you fire up sml with a command-line argument, say,
1745    e.g.:
1746    
1747       sml '$smlnj/cmb.cm'
1748    
1749    At the ML toplevel prompt:
1750    
1751       CM.autoload "$smlnj/cmb.cm";
1752    
1753    There is also a new controller in CM.Control that can be used to turn
1754    off all remaining support for implicit anchors by saying:
1755    
1756        CM.autoload "$smlnj/
1757        #set CM.Control.implicit_anchors false;
1758    
1759    This causes CM to reject implicitly anchored paths.  This is (for the
1760    time being) less permissive than the "final" version where there will
1761    be no more such implicit anchors and relative paths will be just that:
1762    relative.
1763    
1764    The next step (version after next version?) will be to make the
1765    default for CM.Control.implicit_anchors false.  After the dust has
1766    settled, I can then produce the "final" version of this...
1767    
1768    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1769    
1770    ----------------------------------------------------------------------
1771    Name: Matthias Blume
1772    Date: 2000/05/11 16:30:00 JST
1773    Tag: blume-20000511-sources
1774    Description:
1775    
1776    The main change is that I added function CM.sources as a generalized
1777    version of the earlier CM.makedepend.  This entails the following
1778    additional changes:
1779    
1780      - CM.makedepend has been dropped.
1781    
1782      - CM manual has been updated.
1783    
1784      - TOOLS signature and API have been changed.
1785    
1786    ----------------------------------------------------------------------
1787    Name: Allen Leung
1788    Date: 2000/05/10 21:17 EDT
1789    Tag: leunga-20000510-moby-c--ssa
1790    Description:
1791    
1792      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1793    None of these affect SML/NJ.
1794    
1795    1. Register Allocation
1796    
1797        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1798           This module tries to remove local (i.e. basic block level) redundancies
1799           during spilling.
1800    
1801        b. A new framework for performing region based register allocation.
1802           Not yet entirely functional.
1803    
1804    2. X86
1805    
1806       a. DefUse for POP was missing the stack pointer [found by Lal]
1807       b. Reload for CALL was incorrect in X86Spill [found by John]
1808       c. Various fixes in X86Spill so that it can be used correctly for
1809          the new spilling module.
1810    
1811    3. SSA/IR
1812    
1813       a. New module ir/dj-dataflow.sml implements elimination based
1814          data flow analysis.
1815    
1816    4. MLRiscGen
1817    
1818       a. Fix for gc type annotation
1819    
1820    5. MDGen
1821    
1822       Various fixes for machine description -> ml code translation.  For ssa
1823       only.
1824    
1825    ----------------------------------------------------------------------
1826    Name: Allen Leung
1827    Date: 2000/05/08 22:17 EDT
1828    Tag: leunga-20000508-labexp
1829    Description:
1830    
1831      Fermin has found a few assembly problems with constant expressions
1832      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1833      which choke on dumb assemblers.  This is his fix.
1834    
1835    ----------------------------------------------------------------------
1836    Name: Dave MacQueen
1837    Date: 2000/04/09 14:00 EDT
1838    Tag: dbm-20000502-Version_110_28
1839    Description:
1840    
1841    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1842    
1843    2. Updated config/version to 110.28
1844    
1845    3. Updated config/srcarchiveurl
1846    
1847    3. New boot files!
1848       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1849    
1850    ----------------------------------------------------------------------
1851    Name: Matthias Blume
1852    Date: 2000/05/01 19:05:00 JST
1853    Tag: blume-20000501-noweb
1854    Description:
1855    
1856    A new noweb tool has been added.  The existing system is entirely
1857    unaffected by this, but some CM users have asked for renewed noweb
1858    support.  Everything is documented in the CM manual.
1859    
1860    New (plugin) libraries:
1861    
1862       noweb-tool.cm
1863       nw-ext.cm
1864    
1865    ----------------------------------------------------------------------
1866    Name: Dave MacQueen
1867    Date: 2000/04/30 12:40PM EDT
1868    Tag: dbm-20000430-bug_fixes
1869    Description:
1870    
1871    1. Fix for bug 1498
1872       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1873       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1874         added toRealArray function
1875       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1876         added check for tag Obj.RealArray to array printing case in ppObj
1877    
1878    2. Fix for bug 1510
1879       smlnj/src/compiler/Semant/types/typesutil.sml
1880         fixed definition of dummyargs (used by equalTycon) so that
1881         dummy args are distinct types
1882    
1883    ----------------------------------------------------------------------
1884    Name: Matthias Blume
1885    Date: 2000/04/30 01:00:00 JST
1886    Tag: blume-20000430-versions
1887    Description:
1888    
1889    1. CM version numbering added.  This is an implementation of Lal's
1890       proposal for adding version numbers and version checking to .cm
1891       files.  Lal said that his proposal was just that -- a proposal.
1892       For the time being I went ahead and implemented it so that people
1893       can comment on it.  Everything is completely backward-compatible
1894       (except for the stable library format, i.e., new bootfiles!).
1895    
1896       As usual, see the CM manual for details.
1897    
1898    2. An alternative syntax for anchored paths has been implemented.
1899       Dave has recently voiced the same concerns that I had when I did
1900       this, so there should be some support.  My take is that eventually
1901       I will let support for the current syntax (where anchors are
1902       "implicit") fade out in favor of the new, explicit syntax.
1903       In order to be backward-compatible, both old and new syntax are
1904       currently supported.
1905    
1906       Again, see the CM manual for details.
1907    
1908    3. Parallel make is trying to be slightly smarter:  When the master
1909       process finds a "bottleneck", i.e., when there is only one
1910       compilation unit that can be compiled and everybody else is
1911       waiting on it, then it will simply compile it directly instead
1912       of clumsily telling one of the slaves to do it.
1913    
1914    4. Support for "unsharing" added.  This is necessary in order to be
1915       able to have two different versions of the same library running
1916       at the same time (e.g., for trying out a new MLRISC while still
1917       having the old MLRISC linked into the current compiler, etc.)
1918       See the CM manual.
1919    
1920    5. Simple "makedepend" functionality added for generating Makefile
1921       dependency information.  (This is rather crude at the moment.
1922       Expect some changes here in the future.)
1923    
1924    6. ".fun" added as a recognized suffix for ML files. Also documented
1925       explicitly in the manual that the fallback behavior (unknown suffix
1926       -> ML file) is not an official feature!
1927    
1928    7. Small changes to the pickler for stable libraries.
1929    
1930    8. Several internal changes to CM (for cleanup/improvement).
1931    
1932    
1933    !!!! NEW BINFILES !!!!
1934    
1935    ----------------------------------------------------------------------
1936    Name: Matthias Blume
1937    Date: 2000/04/28 17:30:00 JST
1938    Tag: blume-20000428-pathconfig
1939    Description:
1940    
1941    1. I changed config/install.sh to remove duplicate entries from the
1942       lib/pathconfig file at the end.  Moreover, the final version of
1943       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1944       in src/system/installml.
1945    
1946    2. The config/install.sh script now consistently uses relative
1947       pathnames in lib/pathconfig whenever the anchor is in the lib
1948       directory.  (So far this was true for the libraries that come
1949       pre-compiled and bundled as part of the bootfiles but not for
1950       libraries that are compiled by the script itself.)
1951    
1952    ----------------------------------------------------------------------
1953    Name: Matthias Blume
1954    Date: 2000/04/26 13:10:00 JST
1955    Tag: blume-20000426-fun_suffix
1956    Description:
1957    
1958    Added ".fun" as a recognized file name suffix (for ML code).
1959    
1960    ----------------------------------------------------------------------
1961    Name: Allen Leung
1962    Date: 2000/04/25 17:00:00 EST
1963    Tag: leunga-20000425-alpha-ra
1964    Description:
1965    
1966    1. Alpha
1967    
1968        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1969    in C--.
1970    
1971    2. RA
1972    
1973       Added a flag "ra-dump-size" to print out the size of the flowgraph
1974       and the interference graph.
1975    
1976    ----------------------------------------------------------------------
1977    Name: Dave MacQueen
1978    Date: 2000/04/25/
1979    Tag: dbm-20000425-mlyacc_doc_examples
1980    Description:
1981      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1982      Updated all three examples in src/ml-yacc/examples to run
1983      under 110.* using CM.make.
1984    
1985    ----------------------------------------------------------------------
1986    Name: Allen Leung
1987    Date: 2000/04/20 23:04:00 EST
1988    Tag: leunga-20000420-ssa-c---stuff
1989    Description:
1990    
1991      This update synchronizes my repository with Yale's.  Most of these
1992    changes, however, do not affect SML/NJ at all (the RA is an exception).
1993    
1994    1. Register Allocator
1995    
1996       a. An improvement in the interference graph construction:
1997          Given a copy
1998    
1999                s <- t
2000    
2001          no interference edge between s and t is added for this definition of s.
2002    
2003       b. I've added two new spill heuristic modules that Fermin and I developed
2004          (in the new library RA.cm). These are unused in SML/NJ but maybe
2005          useful for others (Moby?)
2006    
2007    2. X86
2008    
2009       a. Various fixes in the backend provided by Fermin [C--] and Lal.
2010    
2011    3. Alpha
2012    
2013       a. Added the BSR instruction and code generation that goes with it [C--]
2014       b. Other fixes too numerous to recount provided by Fermin [C--]
2015    
2016    4. Regmaps
2017    
2018       a. The regmaps are not initialized with the identity physical bindings
2019          at creation time.  This is unneeded.
2020    
2021    5. MLRISC Optimizations
2022    
2023       a. The DJ-Graph module can now compute the iterated dominance frontiers
2024          intersects with liveness incrementally in linear time! Woohoo!
2025          This is now used in my new SSA construction algorithm.
2026    
2027       b. THe branch reorganization module is now smarter about linear chains of
2028          basic blocks.
2029    
2030    
2031    ----------------------------------------------------------------------
2032    Name: Matthias Blume
2033    Date: 2000/04/12 13:52:00 JST
2034    Tag: blume_main_v110p27_1
2035    Description:
2036    
2037    Changed install.sh script to handle archive files without version number
2038    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
2039    name of the boot file archive.
2040    
2041    ----------------------------------------------------------------------
2042    Name: Dave MacQueen
2043    Date: 2000/04/09 14:00 EDT
2044    Tag: dbm-20000410-Version_110_27
2045    Description:
2046    
2047    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
2048    
2049    2. Updated src/config/version to 110.27
2050    
2051    3. New boot files!
2052    
2053    ----------------------------------------------------------------------
2054    Name: Allen Leung
2055    Date: 2000/04/09 19:09:00 EST
2056    Tag: leunga-20000409-misc
2057    Description:
2058    
2059    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
2060    
2061    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
2062    
2063    ----------------------------------------------------------------------
2064    Name: Stefan
2065    Date: 2000/04/07 10:00:00 EDT
2066    Tag: monnier-20000406-branch-handling
2067    Description:
2068    
2069    Improved handling of branches (mostly those generated from
2070    polymorphic equality), removed switchoff and changed the
2071    default optimization settings (more cpsopt and less flintopt).
2072    
2073    ----------------------------------------------------------------------
2074    Name: Allen Leung
2075    Date: 2000/04/06 01:30:00 EST
2076    Tag: leunga-20000406-peephole-x86-SSA-2
2077    Description:
2078    
2079       Forgot a few files.
2080    
2081    ----------------------------------------------------------------------
2082    Name: Allen Leung
2083    Date: 2000/04/06 00:36:00 EST
2084    Tag: leunga-20000406-peephole-x86-SSA
2085    Description:
2086    
2087    1.  New Peephole code
2088    
2089    2.  Minor improvement to X86 instruction selection
2090    
2091    3.  Various fixes to SSA and machine description -> code translator
2092    
2093    ----------------------------------------------------------------------
2094    Name: Matthias Blume
2095    Date: 2000/04/05 12:30:00 JST
2096    Tag: blume_main_v110p26p2_3
2097    Description:
2098    
2099    This update just merges three minor cosmetic updates to CM's sources
2100    to get ready for the 110.27 code freeze on Friday.  No functionality
2101    has changed.
2102    
2103    ----------------------------------------------------------------------
2104    Name: Allen Leung
2105    Date: 2000/04/04 19:39:00 EST
2106    Tag: leunga-20000404-x86-asm
2107    Description:
2108    
2109    1.  Fixed a problem in X86 assembly.
2110    
2111        Things like
2112    
2113           jmp %eax
2114           jmp (%eax)
2115    
2116        should be output as
2117    
2118           jmp *%eax
2119           jmp *(%eax)
2120    
2121    2.  Assembly output
2122    
2123          Added a new flag
2124    
2125              "asm-indent-copies" (default to false)
2126    
2127          When this flag is on, parallel copies will be indented an extra level.
2128    
2129  ----------------------------------------------------------------------  ----------------------------------------------------------------------
2130  Name: Allen Leung  Name: Allen Leung
2131  Date: 2000/04/04 03:18:00 EST  Date: 2000/04/04 03:18:00 EST
# Line 31  Line 2146 
2146    
2147  3.  Assembly  3.  Assembly
2148    
2149      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
2150      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
2151      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
2152    
# Line 54  Line 2169 
2169    
2170  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
2171  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
2172  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
2173  legal surface-syntax symbol).  legal surface-syntax symbol).
2174    
2175  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 64  Line 2179 
2179  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
2180    
2181  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2182  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
2183  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
2184  more details.  more details.
2185    
# Line 179  Line 2294 
2294     (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
2295     used).     used).
2296    
2297     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
2298     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
2299     Makefiles or Shell scripts.     Makefiles or Shell scripts.
2300    
2301     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 192  Line 2307 
2307    
2308     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
2309    
2310     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
2311     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
2312    
2313  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 326  Line 2441 
2441    
2442       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
2443       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
2444       testl is perferable to andl.       testl is preferable to andl.
2445    
2446  7. RA (x86 only)  7. RA (x86 only)
2447    
# Line 508  Line 2623 
2623    
2624  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2625  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2626  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
2627  explanation.  explanation.
2628    
2629  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2630  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2631  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2632  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 555  Line 2670 
2670  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
2671  <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,
2672  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
2673  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2674  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2675  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2676    
# Line 605  Line 2720 
2720    it from that remote directory.    it from that remote directory.
2721    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2722    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2723    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2724    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2725    
2726    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.593  
changed lines
  Added in v.838

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