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 605, Fri Apr 7 14:06:42 2000 UTC revision 847, Mon Jun 25 16:11:22 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: 2001/06/25 12:05:00 EDT
18    Tag: blume-20010625-nlffigen
19    Description:
20    
21    This update fixes a number of minor bugs in ml-nlffigen as reported by
22    Nick Carter <nbc@andrew.cmu.edu>.
23    
24      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
25      2. Default names for generated files are now derived from the name of
26         the C file *without its directory*.  In particular, this causes generated
27         files to be placed locally even if the C file is in some system directory.
28      3. Default names for generated signatures and structures are also derived
29         from the C file name without its directory.  This avoids silly things
30         like "structure GL/GL".
31         (Other silly names are still possible because ml-nlffigen does not do
32          a thorough check of whether generated names are legal ML identifiers.
33          When in doubt, use command line arguments to force particular names.)
34    
35    ----------------------------------------------------------------------
36    Name: Matthias Blume
37    Date: 2001/06/21 12:25:00 EDT
38    Tag: blume-20010621-eXene
39    Description:
40    
41    eXene now compiles and (sort of) works again.
42    
43    The library name (for version > 110.33) is $/eXene.cm.
44    
45    I also added an new example in src/eXene/examples/nbody.  See the
46    README file there for details.
47    
48    ----------------------------------------------------------------------
49    Name: Matthias Blume
50    Date: 2001/06/20 16:40:00 EDT
51    Tag: blume-20010620-cml
52    Description:
53    
54    CML now compiles and works again.
55    
56    Libraries (for version > 110.33):
57    
58      $cml/cml.cm            Main CML library.
59      $cml/basis.cm          CML's version of $/basis.cm.
60      $cml/cml-internal.cm   Internal helper library.
61      $cml/core-cml.cm       Internal helper library.
62      $cml-lib/trace-cml.cm  Tracing facility.
63      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
64    
65    The installer (config/install.sh) has been taught how to properly
66    install this stuff.
67    
68    ----------------------------------------------------------------------
69    Name: Matthias Blume
70    Date: 2001/06/19 17:55:00 EDT
71    Tag: blume-20010619-instantiate
72    Description:
73    
74    This un-breaks the fix for bug 1432.
75    (The bug was originally fixed in 110.9 but I broke it again some
76    time after that.)
77    
78    ----------------------------------------------------------------------
79    Name: Matthias Blume
80    Date: 2001/06/19 17:25:00 EDT
81    Tag: blume-20010619-signals
82    Description:
83    
84    This should (hopefully) fix the long-standing signal handling bug.
85    (The runtime system was constructing a continuation record with an
86    incorrect descriptor which would cause the GC to drop data on the floor...)
87    
88    ----------------------------------------------------------------------
89    Name: Matthias Blume
90    Date: 2001/06/15 15:05:00 EDT
91    Tag: blume-20010615-moresparc
92    Description:
93    
94    Here is a short late-hour update related to Sparc c-calls:
95    
96     -- made handling of double-word arguments a bit smarter
97    
98     -- instruction selection phase tries to collapse certain clumsily
99        constructed ML-Trees; typical example:
100    
101            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
102    
103        This currently has no further impact on SML/NJ since mlriscGen does
104        not seem to generate such patterns in the first place, and c-calls
105        (which did generate them in the beginning) has meanwhile been fixed
106        so as to avoid them as well.
107    
108    ----------------------------------------------------------------------
109    Name: Matthias Blume
110    Date: 2001/06/15 15:05:00 EDT
111    Tag: blume-20010615-sparc
112    Description:
113    
114    The purpose of this update is to provide an implementation of NLFFI
115    on Sparc machines.
116    
117    Here are the changes in detail:
118    
119       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
120       the Sparc implementation of the c-calls API.
121       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
122       for accessing the ML frame.  Thus, we have a real frame pointer and
123       can freely modify %sp without need for an omit-frame-ptr phase.
124       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
125       this case.
126       * ml-nlffigen has been taught to produce code for different architectures
127       and calling conventions.
128       * In a way similar to what was done in the x86 case, the Sparc
129       backend uses its own specific extension to mltree.  (For example,
130       it needs to be able to generate UNIMP instructions which are part
131       of the calling convention.)
132       * ml-nlffi-lib was reorganized to make it more modular (in particular,
133       to make it easier to plug in new machine- and os-dependent parts).
134    
135    There are some other fairly unrelated bug fixes and cleanups as well:
136    
137       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
138       that they properly share their libraries with existing SML/NJ libraries.
139       * I fixed a minor cosmetic bug in CM, supressing certain spurious
140       follow-up error messages.
141       * Updates to CM/CMB documentation.
142    
143    TODO items:
144    
145       * MLRISC should use a different register as its asmTemp on the Sparc.
146         (The current %o2 is a really bad choice because it is part of the
147         calling conventions, so things might interfere in unexpected ways.)
148    
149    ----------------------------------------------------------------------
150    Name: Matthias Blume
151    Date: 2001/06/07
152    Tag: blume-20010607-calls
153    Description:
154    
155    A number of internal changes related to C calls and calling conventions:
156    
157    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
158       number of bytes popped implicitly (by the callee).  In most cases
159       this field is 0 but on x86/win32 it is some non-zero value.  This
160       is information provided for the benefit of the "omit-frameptr" pass.
161    2. The CALL instruction on the x86 carries a similar "pops" field.
162       The instruction selection phase copies its value from the ML-Tree
163       CALL statement.
164    3. On all other architectures, the instruction selection phase checks
165       whether "pops=0" and complains if not.
166    4. The c-calls implementation for x86 now accepts two calling conventions:
167       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
168       up after the call and pops is set to 0.  For "stdcall", the caller
169       does nothing, leaving the cleanup to the callee; pops is set to
170       the number of bytes that were pushed onto the stack.
171    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
172       distinguish between "ccall" and "stdcall".
173    6. The UNIMP instruction has been added to the supported Sparc instruction
174       set. (This is needed for implementing the official C calling convention
175       on this architecture.)
176    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
177       work with the latest CM.
178    
179    ----------------------------------------------------------------------
180    Name: Matthias Blume
181    Date: 2001/06/05 15:10:00 EDT
182    Tag: blume-20010605-cm-index
183    Description:
184    
185    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
186    
187    1. CM can now generate "index files".  These are human-readable files
188       that list on a per-.cm-file basis each toplevel symbol defined or
189       imported.  The location of the index file for
190       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
191       To enable index-file generation, set CM.Control.generate_index to true
192       or export an environment-symbol: export CM_GENERATE_INDEX=true.
193    
194       The CM manual has been updated accordingly.
195    
196    2. I made some slight modifications to the c-calls API in MLRISC.
197    
198         a) There is now a callback to support saving/restoring of
199            dedicated but caller-save registers around the actual call
200            instruction.
201         b) One can optionally specify a comment-annotation for the
202            call instruction.
203    
204    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
205       (For example, the comment annotation shows the C prototype of
206        the function being called.)
207    
208    ----------------------------------------------------------------------
209    Name: Matthias Blume
210    Date: 2001/06/01 13:30:00 EDT
211    Tag: blume-20010601-nlffi-cleanup
212    Description:
213    
214    This is mostly a cleanup of MLFFI stuff:
215    
216       - some signature files have been put into a more exposed place
217       - the ugly 'f type parameter is gone (simplifies types tremendously!)
218       - ml-nlffigen changed accordingly
219       - tutorial updated
220    
221    Other changes:
222    
223       - author's affiliation in CM manual(s) updated
224       - some more recognized keywords added to Allen's sml.sty
225    
226    ----------------------------------------------------------------------
227    Name: Matthias Blume
228    Date: 2001/05/25 15:30:00 EDT
229    Tag: blume-20010525-iptr
230    Description:
231    
232      - put the official 110.33-README (as it appears on the ftp server) under
233        CVS
234      - fixed a small bug related to incomplete pointer types in
235        ml-nlffigen
236      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
237        (it does not need the 'f type parameter)
238    
239    ----------------------------------------------------------------------
240    Name: Matthias Blume
241    Date: 2001/05/23 14:30:00 EDT
242    Tag: Release_110_33
243    Description:
244    
245    New version number (110.33).  New bootfiles.
246    
247    ----------------------------------------------------------------------
248    Name: Matthias Blume
249    Date: 2001/05/22 18:06:00 EDT
250    Tag: blume-20010522-targets
251    Description:
252    
253    Made install.sh use file config/targets.customized if it exists, falling
254    back to config/targets if it doesn't.  This way one can have a customized
255    version of the targets file without touching the "real thing", thus
256    eliminating the constant fear of accidentally checking something bogus
257    back into the CVS repository...  (File config/targets.customized must
258    not be added to the repository!)
259    
260    ----------------------------------------------------------------------
261    Name: Matthias Blume
262    Date: 2001/05/22 16:30:00 EDT
263    Tag: blume-20010522-minitut
264    Description:
265    
266    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
267       struct returns.
268    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
269       incomplete, preliminary documentation for NLFFI.
270    
271    ----------------------------------------------------------------------
272    Name: Matthias Blume
273    Date: 2001/05/14 11:30:00 EDT
274    Tag: blume-20010514-script
275    Description:
276    
277    Some bugs in install script fixed.
278    
279    In addition to that I also made a slight change to the NLFFI API:
280    Functors generated by ml-nlffigen now take the dynamic library as a
281    straight functor argument, not as a suspended one.  (The original
282    functor code used to force the suspension right away anyway, so there
283    was nothing gained by this complication of the interface.)
284    
285    ----------------------------------------------------------------------
286    Name: Matthias Blume
287    Date: 2001/05/11 14:35:00 EDT
288    Tag: blume-20010511-ml-nlffi
289    Description:
290    
291    I finally took the plunge and added my new FFI code to the main
292    repository.  For x86-linux it is now ready for prime-time.
293    
294    There are two new subdirectories of "src":
295    
296      - ml-nlffi-lib:
297           The utility library for programs using the FFI interface.
298           Here is the implementation of $/c.cm and its associated low-level
299           partners $/c-int.cm and $/memory.cm.
300      - ml-nlffigen:
301           A stand-alone program for generating ML glue code from C source
302           code.
303    
304    Building ml-nlffigen requires $/ckit-lib.cm.
305    
306    The config/install.sh script has been updates to do the Right Thing
307    (hopefully).
308    
309    Notice that the source tree for the C-Kit will not be put under "src"
310    but directly under the installation root directory.  (This is the
311    structure that currently exists on the CVS server when you check out
312    module "sml".)  Fortunately, config/install.sh knows about this oddity.
313    
314    Bugs: No documentation yet.
315    
316    ----------------------------------------------------------------------
317    Name: Matthias Blume
318    Date: 2001/05/09 16:35:00 EDT
319    Tag: blume-20010509-cpscontract
320    Description:
321    
322    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
323    wrapper/unwrapper elimination did not decrement usage counts and some
324    dead variables got overlooked by the dead-up logic.)
325    
326    ----------------------------------------------------------------------
327    Name: Lal George
328    Date: 2001/05/08  17:26:09 EDT
329    Tag: george-20010508-omit-frameptr
330    Description:
331    
332    Changes to implement the omit-frame-pointer optimization to support
333    raw C calls. For now, there is only support on the Intel x86, but
334    other architectures will follow as more experience is gained with this.
335    
336    
337    ----------------------------------------------------------------------
338    Name: Matthias Blume
339    Date: 2001/05/07 14:40:00 EDT
340    Tag: blume-20010507-proxies
341    Description:
342    
343    I made into "proxy libraries" all libraries that qualify for such a
344    change.  (A qualifying library is a library that has another library or
345    groups as its sole member and repeats that member's export list
346    verbatim.  A proxy library avoids this repetition by omitting its export
347    list, effectively inheriting the list that its (only) member exports.
348    See the CM manual for more explanation.)
349    The main effect is that explicit export lists for these libraries
350    do not have to be kepts in sync, making maintenance a bit easier.
351    
352    I also added copyright notices to many .cm-files.
353    
354    Last but not least, I made a new set of bootfiles.
355    
356    ----------------------------------------------------------------------
357    Name: Matthias Blume
358    Date: 2001/05/04 17:00:00 EDT
359    Tag: blume-20010504-cm-lsplit
360    Description:
361    
362    0. John merged pending changes to $/smlnj-lib.cm
363    
364    1. Allen's previous change accidentally backed out of one of Lal's
365       earlier changes.  I undid this mistake (re-introducing Lal's change).
366    
367    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
368       within the compiler where applicable.  There is some code simplification
369       because of that.
370    
371    3. The "split" phase (in FLINT) is now part of the default list of phases.
372       Compiler.Control.LambdaSplitting.* can be used to globally control the
373       lambda-splitting (cross-module-inlining) engine.  In addition to that,
374       it can now also be controlled on a per-source basis: CM has been taught
375       a new tool parameter applicable to ML source files.
376    
377       - To turn lambda-splitting off completely:
378            local open Compiler.Control.LambdaSplitting in
379                val _ = set Off
380            end
381       - To make "no lambda-splitting" the global default (but allow per-source
382         overriding); this is the initial setting:
383            local open Compiler.Control.LambdaSplitting in
384                val _ = set (Default NONE)
385            end
386       - To make "lambda-splitting with aggressiveness a" the global default
387         (and allow per-source overriding):
388            local open Compiler.Control.LambdaSplitting in
389                val _ = set (Default (SOME a))
390            end
391    
392       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
393         write (in the respective .cm-file):
394            a.sml (lambdasplitting:off)
395       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
396            a.sml (lambdasplitting:on)
397       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
398         <a> is a decimal non-negative integer):
399            a.sml (lambdasplitting:<a>)
400       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
401            a.sml (lambdasplitting:infinity)
402       - To use the global default for a.sml:
403            a.sml (lambdasplitting:default)
404         or simply
405            a.sml
406    
407    ----------------------------------------------------------------------
408    Name: Allen Leung
409    Date: 2001/05/04 01:57:00 EDT
410    Tag: leunga-20010504-sync
411    Description:
412    
413      MLRISC features.
414    
415      1. Fix to CMPXCHG instructions.
416      2. Changed RA interface to allow annotations in callbacks.
417      3. Added a new method to the stream interface to allow annotations updates.
418    
419    ----------------------------------------------------------------------
420    Name: Matthias Blume
421    Date: 2001/05/01 11:45:00 EDT
422    Tag: blume-20010501-pcedittmp
423    Description:
424    
425    Changed install.sh to use the current working directory instead of
426    /usr/tmp for a temporary file (pcedittmp).  The previous choice
427    of /usr/tmp caused trouble with MacOS X because of file premission
428    problems.
429    
430    ----------------------------------------------------------------------
431    Name: Matthias Blume
432    Date: 2001/04/20 11:10:00 EDT
433    Tag: blume-20010420-inMLflag
434    Description:
435    
436     - added vp_limitPtrMask to vproc-state.h
437       (for use by the raw-C-calls mechanism to implement proper interrupt
438        handling)
439     - made the ML compiler aware of various data-structure offsets so it
440       can generate code for accessing the vp_inML flag and vp_limitPtrMask
441     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
442       raw C-calls
443    
444    ----------------------------------------------------------------------
445    Name: Lal George
446    Date: 2001/04/20 09:15:28 EDT
447    Tag: george-20010420-macosX
448    Description:
449    
450     - Changes to port to Mac OS X; Darwin.
451    
452     - In the process I found that sqrt was broken on the PPC, because the
453       fsqrt instruction is not implemented.
454    
455    ----------------------------------------------------------------------
456    Name: Matthias Blume
457    Date: 2001/04/18 12:45:00 EDT
458    Tag: blume-20010418-ccalls
459    Description:
460    
461     - fixed two off-by-4 errors in the x86-specific c-calls implementation
462       (this bug prevented structure arguments containing pointers from being
463        passed correctly)
464     - changed the raw-C-call code in mlriscGen.sml in such a way that
465       structure arguments are represented as a pointer to the beginning
466       of the structure (instead of having a series of synthesized arguments,
467       one for each structure member)
468    
469     - made makeml script's verbosity level configurable via environment
470       variable (MAKEML_VERBOSITY)
471    
472     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
473       in rawmem-x86.sml; we are now using the real thing
474    
475    ----------------------------------------------------------------------
476    Name: Matthias Blume
477    Date: 2001/03/22 16:25:00 EST
478    Tag: blume-20010322-bootfiles
479    Description:
480    
481    Created a new set of bootfiles (for your automatic installation convenience).
482    
483    ----------------------------------------------------------------------
484    Name: Matthias Blume
485    Date: 2001/03/22 15:10:00 EST
486    Tag: blume-20010322-rawmem-parcm
487    Description:
488    
489    1. All "raw memory access" primitives for the new FFI are implemented now
490       (at least on the x86).
491    2. Some further cleanup of CM's parallel make mechanism.
492    
493    ----------------------------------------------------------------------
494    Name: Matthias Blume
495    Date: 2001/03/19 17:53:00 EST
496    Tag: blume-20010319-parallel
497    Description:
498    
499    Parallel make (using compile servers) now works again.
500    
501    To this end, CM.stabilize and CMB.make have been modified to work in
502    two passes when compile servers are attached:
503       1. Compile everything, do not perform stabilization; this pass
504          uses compile servers
505       2. Stabilize everything; this pass does not use compile servers
506    If there are no compile servers, the two passes are combined into one
507    (as before).  Splitting the passes increases the inherent parallelism
508    in the dependency graph because the entire graph including all
509    libraries is available at the same time.  This, in turn, improves
510    server utilization.  The downside is that the master process will
511    have to do some extra work after compilation is done (because for
512    technical reasons it must re-read all the binfiles during stabilization).
513    
514    ----------------------------------------------------------------------
515    Name: Matthias Blume
516    Date: 2001/03/16 12:22:00 EST
517    Tag: blume-20010316-bootfiles
518    Description:
519    
520    Created a new set of bootfiles (for your automatic installation convenience).
521    
522    ----------------------------------------------------------------------
523    Name: Matthias Blume
524    Date: 2001/03/16 11:00:00 EST
525    Tag: blume-20010316-MLTREE-fixup
526    Description:
527    
528    This is a minor fixup for an (untagged) earlier commit by Allen.
529    (A file was missing).
530    
531    ----------------------------------------------------------------------
532    Name: Allen Leung
533    Date: Mon Mar  5 18:54:57 EST 2001
534    Tag: leunga-20010305-cut-support
535    
536    1. New support for alternative control-flow in MLTREE.
537       Currently we support
538    
539          FLOW_TO(CALL ...., [k1,...,kn])
540    
541       This is needed for 'cuts to' in C-- and try/handle-like constructs
542       in Moby
543    
544       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
545    
546    2. Register Allocator
547    
548       Changes in interface [from Fermin, John]
549    
550    3. Alpha 8-bit SLL support [Fermin]
551    
552    4. All architectures
553    
554       A new module (ClusterExpandCopies) for expanding parallel copies.
555    
556    ----------------------------------------------------------------------
557    Name: Allen Leung
558    Date: 2001/02/27 23:07:00 EST
559    Tag: leunga-20010227-minor-stuff
560    
561    1. Alpha bug fix for CMOVNE
562    2. Handle mltree COND(..,FCMP ...,...)
563    3. Bug fix in simplifier
564    
565    ----------------------------------------------------------------------
566    Name: Matthias Blume
567    Date: 2001/01/30 17:50:00 EST
568    Tag: blume-20010130-sync
569    Description:
570    
571    This is just a minor update to sync my devel branch with the main brach.
572    The only visible change is the addition of some README files.
573    
574    ----------------------------------------------------------------------
575    Name: Matthias Blume
576    Date: 2001/01/12 23:30:00 JST
577    Tag: blume-20010112-bootfiles
578    Description:
579    
580    Made a new set of bootfiles that goes with the current state of the
581    repository.
582    
583    ----------------------------------------------------------------------
584    Name: Matthias Blume
585    Date: 2001/01/12 21:20:00 JST
586    Tag: blume-20010112-sync
587    Description:
588    
589    I am just flushing out some minor changes that had accumulated in
590    my private branch in order to sync with the main tree.  (This is
591    mainly because I had CVS trouble when trying to merge _into_ my
592    private branch.)
593    
594    Most people should be completely unaffected by this.
595    
596    ----------------------------------------------------------------------
597    Name: Allen Leung
598    Date: Thu Jan 11 21:03:00 EST 2001
599    Tag: leunga-20010111-labexp=mltree
600    Description:
601    
602    1.  Removed the type LabelExp and replace it by MLTree.
603    2.  Rewritten mltree-simplify with the pattern matcher tool.
604    3.  There were some bugs in alpha code generator which would break
605        64-bit code generation.
606    4.  Redo the tools to generate code with the
607    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
608        are now generated by perl scripts.
609    
610    ----------------------------------------------------------------------
611    Name: Matthias Blume
612    Date: 2001/01/10 21:55:00 JST
613    Tag: blume-20010110-rcc
614    Description:
615    
616    The RCC stuff now seems to work (but only on the x86).
617    This required hacking of the c-calls interface (and -implementation) in
618    MLRISC.
619    
620    Normal compiler users should be unaffected.
621    
622    ----------------------------------------------------------------------
623    Name: Matthias Blume
624    Date: 2001/01/09 01:20:00 JST
625    Tag: blume-20010109-rcc
626    Description:
627    
628    This is a fairly big patch, flushing out a large number of pending
629    changes that I made to my development copy over the last couple of days.
630    
631    Of practical relevance at this moment is a workaround for a pickling
632    bug that Allen ran into the other day.  The cause of the bug itself is
633    still unknown and it might be hard to fix it properly, but the
634    workaround has some merits of its own (namely somewhat reducing pickling
635    overhead for certain libraries).  Therefore, I think this solution should
636    be satisfactory at this time.
637    
638    The rest of the changes (i.e., the vast majority) has to do with my
639    ongoing efforts of providing direct support for C function calls from
640    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
641    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
642    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
643    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
644    handle RCC.
645    
646    The last part (the changes to mlriscGen.sml) are still known to be
647    wrong on the x86 and not implemented on all other architectures.  But
648    the infrastructure is in place. I had to change a few functor
649    signatures in the backend to be able to route the CCalls interface
650    from MLRISC there, and I had to specialize the mltree type (on the
651    x86) to include the necessary extensions. (The extensions themselves
652    were already there and redy to go in MLRISC/x86).
653    
654    Everything should be very happy as soon as someone helps me with
655    mlriscGen.sml...
656    
657    In any case, nothing of this should matter to anyone as long as the
658    new primop is not being used (which is going to be the case unless you
659    find it where I hid it :). The rest of the compiler is completely
660    unaffected.
661    
662    ----------------------------------------------------------------------
663    Name: Matthias Blume
664    Date: 2001/01/05 00:30:00 JST
665    Tag: blume-20010105-primops
666    Description:
667    
668    Added some experimental support for work that I am doing right now.
669    These changes mostly concern added primops, but there is also a new
670    experimental C library in the runtime system (but currently not enabled
671    anywhere except on Linux/X86).
672    
673    In the course of adding primops (and playing with them), I discovered that
674    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
675    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
676    To recover, I made type information in INL_PRIM mandatory and changed
677    prim.sml as well as built-in.sml accordingly.  The InLine structure now
678    has complete, correct type information (i.e., no bottom types).
679    
680    Since all these changes mean that we need new binfiles, I also bumped the
681    version number to 110.32.1.
682    
683    ----------------------------------------------------------------------
684    Name: Matthias Blume
685    Date: 2000/12/30 22:10:00 JST
686    Tag: blume-20001230-various
687    Description:
688    
689    Added proxy libraries for MLRISC and let MLRISC libraries refer
690    to each other using path anchors.  (See CM manual for explanation.)
691    
692    Updated CM documentation.
693    
694    Fixed some bugs in CM.
695    
696    Implemented "proxy" libraries (= syntactic sugar for CM).
697    
698    Added "-quiet" option to makeml and changed runtime system accordingly.
699    
700    Added cleanup handler for exportML to reset timers and compiler stats.
701    
702    ----------------------------------------------------------------------
703    Name: Lal George
704    Date: 2000/12/22 22:22:58 EST 2000
705    Tag: Release_110_32
706    Description:
707    
708            Infinite precision used throughout MLRISC.
709            see MLRISC/mltree/machine-int.sig
710    
711    ----------------------------------------------------------------------
712    Name: Matthias Blume
713    Date: 2000/12/22 23:16:00 JST
714    Tag: blume-20001222-warn
715    Description:
716    
717    Corrected wording and formatting of some CM warning message which I
718    broke in my previous patch.
719    
720    ----------------------------------------------------------------------
721    Name: Matthias Blume
722    Date: 2000/12/22 21:20:00 JST
723    Tag: blume-20001222-anchorenv
724    Description:
725    
726    Fixed CM's handling of anchor environments in connection with CMB.make.
727    
728    ----------------------------------------------------------------------
729    Name: Matthias Blume
730    Date: 2000/12/22 13:15:00 JST
731    Tag: blume-20001222-cleanup
732    Description:
733    
734    Removed src/cm/ffi which does not (and did not) belong here.
735    
736    ----------------------------------------------------------------------
737    Name: Matthias Blume
738    Date: 2000/12/21 23:55:00 JST
739    Tag: blume-20001221-exn
740    Description:
741    
742    Probably most important: CM no longer silently swallows all exceptions
743    in the compiler.
744    Plus: some other minor CM changes.  For example, CM now reports some
745    sizes for generated binfiles (code, data, envpickle, lambdapickle).
746    
747    ----------------------------------------------------------------------
748    Name: Matthias Blume
749    Date: 2000/12/15 00:01:05 JST
750    Tag: blume-20001215-dirtool
751    Description:
752    
753    - "dir" tool added.
754    - improvements and cleanup to Tools structure
755    - documentation updates
756    
757    ----------------------------------------------------------------------
758    Name: Allen Leung
759    Date: Thu Dec 14 03:45:24 EST 2000
760    Description:
761    Tag:  leunga-20001214-int-inf
762    Description:
763    
764       In IntInf, added these standard functions, which are missing from our
765    implementation:
766    
767        andb : int * int -> int
768        xorb : int * int -> int
769        orb  : int * int -> int
770        notb : int -> int
771         <<   : int * word -> int
772        ~>>  : int * word -> int
773    
774       Not tested, I hope they are correct.
775    
776    ----------------------------------------------------------------------
777    Name: Allen Leung
778    Date: Fri Dec  8 19:23:26 EST 2000
779    Description:
780    Tag:  leunga-20001208-nowhere
781    Description:
782    
783      Slight improvements to the 'nowhere' tool to handle OR-patterns,
784    to generate better error messages etc.  Plus a brief manual.
785    
786    ----------------------------------------------------------------------
787    Name: Lal George
788    Date: 2000/12/08 09:54:02 EST 2000
789    Tag: Release_110_31
790    Description:
791    
792    - Version 110.31
793    ----------------------------------------------------------------------
794    Name: Allen Leung
795    Date: Thu Dec  7 22:01:04 EST 2000
796    Tag:  leunga-20001207-cell-monster-hack
797    Description:
798    
799    Major MLRISC internal changes.  Affect all clients.
800    Summary:
801    
802    1.  Type CELLS.cell = int is now replaced by a datatype.
803        As a result, the old regmap is now gone.  Almost all interfaces
804        in MLRISC change as a consequence.
805    
806    2.  A new brand version of machine description tool (v3.0) that generates
807        modules expecting the new interface.  The old version is removed.
808    
809    3.  The RA interface has been further abstracted into two new functors.
810        RISC_RA and X86RA.  These functors have much simpler interfaces.
811        [See also directory MLRISC/demo.]
812    
813    4.  Some other new source->source code generation tools are available:
814    
815        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
816        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
817           I use this tool to generate the peephole optimizers---with the new
818           cell type changes, peephole rules are becoming difficult to write
819           without conditional pattern matching.
820    
821    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
822        cover the entire MLRISC source tree so many things broke.
823    
824    6.  CM files have been moved to the subdirectory MLRISC/cm.
825        They are moved because there are a lot of them and they clutter up the
826        root dir.
827    
828    7.  More detailed documentation to come...
829    
830        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
831        a makeml -rebuild first.  This is because of other other
832        changes that Matthias has made (see below).
833    
834    
835    ----------------------------------------------------------------------
836    Name: Matthias Blume
837    Date: 2000/11/30 23:12:00 JST
838    Tag: blume-20001130-filereorg
839    Description:
840    
841    Some manual updates and some file reorganizations in CM.
842    
843    ----------------------------------------------------------------------
844    Name: Matthias Blume
845    Date: 2000/11/24 17:45:00 JST
846    Tag: blume-20001124-link
847    Description:
848    
849    Drastically improved link traversal code for the case that the dynamic
850    value was already loaded at bootstrap time.  As a result, CM and CMB
851    now both load blazingly fast -- even on a very slow machine.  Also,
852    memory consumption has been further reduced by this.
853    
854    Warning: The format of the PIDMAP file has changed.  THerefore, to
855    bootstrap you have to do this:
856    
857    1. Run CMB.make
858    2. Make a symbolic link for the boot directory:
859         ln -s sml.boot.ARCH-OS xxx
860    3. "Rebuild" the boot directory:
861         ./makeml -boot xxx -rebuild sml ; rm xxx
862    4. Boot normally:
863          ./makeml
864    
865    ----------------------------------------------------------------------
866    Name: Matthias Blume
867    Date: 2000/11/21 21:20:00 JST
868    Tag: blume-20001121-tools
869    Description:
870    
871    Continued hacking on autoloading problem -- with success this time.
872    Also changed tool-plugin mechanism.  See new CM manual.
873    
874    ----------------------------------------------------------------------
875    Name: Matthias Blume
876    Date: 2000/11/19 14:30:00 JST
877    Tag:  blume-20001119-autoload
878    Description:
879    
880    Some hacking to make autoloading faster.  Success for CMB, no success
881    so far for CM.  There is a reduced structure CM' that autoloads faster.
882    (This is a temporary, non-documented hack to be eliminated again when
883    the general problem is solved.)
884    
885    ----------------------------------------------------------------------
886    Name: Matthias Blume
887    Date: 2000/11/17 14:10:00 JST
888    Tag: blume-20001117-pickle-lib
889    Description:
890    
891    1. Eliminated comp-lib.cm
892    2. Made pickle-lib.cm
893    3. Eliminated all uses of intset.sml (from comp-lib.cm)
894    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
895       equivalent constructs from smlnj-lib.cm (INtHashTable).
896    5. Point 4. also goes for those uses of intmap.* in MLRISC.
897       Duplicated intmap modules thrown out.
898    6. Hunted down all duplicated SCC code and replaced it with
899       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
900    7. Rewrote Feedback module.
901    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
902       should be thrown out and equivalent modules from smlnj-lib.cm
903       should be used (IntRedBlackSet, IntListSet, ...).
904    
905    Confirmed that compiler compiles to fixpoint.
906    
907    ----------------------------------------------------------------------
908    Name: Allen Leung
909    Date: 2000/11/10 18:00:00
910    Tag: leunga-20001110-new-x86-fp
911    
912    A new x86 floating point code generator has been added.
913    By default this is turned off.  To turn this on, do:
914    
915        CM.autoload "$smlnj/compiler.cm";
916        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
917    
918    Changes:
919    
920    1.  Changed FTAN to FPTAN so that the assembly output is correct.
921    2.  Changed the extension callback for FTANGENT to generate:
922    
923              fptan
924              fstp  %st(0)
925        instead of
926              fptan
927              fstpl ftempmem
928    
929    3.  Numerous assembly fixes for x86.
930    
931    5.  Cleaned up the machine code output module x86/x86MC.sml and added
932        support for a whole bunch of instructions and addressing modes:
933    
934          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
935          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
936          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
937          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
938          fxch %st(n)
939          fld %st(n)
940          fst %st(n)
941          fst mem
942          fstp %st(n)
943          fucom %st(n)
944          fucomp %st(n)
945    
946        All these are now generated when the fast fp mode is turned on.
947    
948    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
949    
950    ----------------------------------------------------------------------
951    Name: Matthias Blume
952    Date: 2000/11/09 11:20:00 JST
953    Tag: blume-20001109-scc
954    Description:
955    
956    Eliminated some code duplication:
957    
958    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
959       (Otherwise the functor is useless.)
960    2. Used GraphSCCFn where SCCUtilFun was used previously.
961    3. Got rid of SCCUtilFun (in comp-lib.cm).
962    
963    ----------------------------------------------------------------------
964    Name: Lal George
965    Date: 2000/11/06 09:02:21 EST 2000
966    Tag: Release_110_30
967    Description:
968    
969    - Version 110.30
970    ----------------------------------------------------------------------
971    Name: Matthias Blume
972    Date: 2000/11/04 14:45:00
973    Tag: blume-20001104-mlbuild
974    Description:
975    
976    - Made ml-build faster on startup.
977    - Documentation fixes.
978    
979    ----------------------------------------------------------------------
980    Name: Matthias Blume
981    Date: 2000/11/02 17:00:00 JST
982    Tag: blume-20001102-condcomp
983    Description:
984    
985    - Small tweaks to pickler -- new BOOTFILES!
986    - Version bumped to 110.29.2.
987    - Added conditional compilation facility to init.cmi (see comment there).
988    ----------------------------------------------------------------------
989    Name: Allen Leung
990    Date: 2000/10/23 19:31:00
991    Tag: leunga-20001023-demo-ra
992    
993    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
994    2. Test programs for the graph library updated
995    3. Some new MLRISC demo programs added
996    
997    ----------------------------------------------------------------------
998    Name: Matthias Blume
999    Date: 2000/08/31 22:15:00 JST
1000    Tag: blume-20001017-errmsg
1001    Description:
1002    
1003    More error message grief: Where there used to be no messages, there
1004    now were some that had bogus error regions.  Fixed.
1005    
1006    ----------------------------------------------------------------------
1007    Name: Matthias Blume
1008    Date: 2000/08/31 17:30:00 JST
1009    Tag: blume-20001017-v110p29p1
1010    Description:
1011    
1012    I made a version 110.29.1 with new bootfiles.
1013    
1014    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
1015              CM documentation changes and a small bugfix in CM's error reporting.
1016    
1017    ----------------------------------------------------------------------
1018    Name: Lal George
1019    Date: 2000/09/27 14:42:35 EDT
1020    Tag: george-20000927-nodestatus
1021    Description:
1022    
1023    Changed the type of the nodestatus, so that:
1024    
1025            SPILLED(~1)             is now SPILLED
1026            SPILLED(m) where m>=0   is now MEMREG(m)
1027            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
1028    
1029    ----------------------------------------------------------------------
1030    Name: Matthias Blume
1031    Date: 2000/09/07 14:45:00 JST
1032    Tag: blume-20000907-cmerrmsg
1033    Description:
1034    
1035    Small tweak to CM to avoid getting ML syntax error messages twice.
1036    
1037    ----------------------------------------------------------------------
1038    Name: Matthias Blume
1039    Date: 2000/08/31 18:00:00 JST
1040    Tag: blume-20000831-cvsbootfiles
1041    Description:
1042    
1043    New URL for boot files (because the 110.29 files on the BL server do
1044    now work correctly with my updated install scripts for yacc and lex).
1045    
1046    ----------------------------------------------------------------------
1047    Name: Matthias Blume
1048    Date: 2000/08/08 12:33:00 JST
1049    Tag: blume-20000808-manual
1050    Description:
1051    
1052    Tiny update to CM manual.
1053    
1054    ----------------------------------------------------------------------
1055    Name: Allen Leung
1056    Date: 2000/08/7 19:31:00
1057    Tag: leunga-20000807-a-whole-bunch-of-stuff
1058    
1059      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
1060    and Mobdy.
1061    
1062    1.  x86
1063    
1064       a.  Fixes to peephole module by John and Dan.
1065       b.  Assembly fix to SETcc by Allen.
1066       c.  Fix to c-call by John.
1067       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
1068       e.  Instruction selection optimization to SETcc as suggested by John.
1069    
1070           For example,
1071    
1072            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
1073    
1074           should generate:
1075    
1076            MOVL a, x
1077            SUBL b, x
1078            SHRL 31, x
1079    
1080    2.  IR stuff
1081    
1082         A bunch of new DJ-graph related algorithms added.  These
1083         speed up SSA construction.
1084    
1085    3.  SSA + Scheduling
1086    
1087         Added code for SSA and scheduling to the repository
1088    
1089    ----------------------------------------------------------------------
1090    Name: Lal George
1091    Date: 2000/07/27 11:53:14 EDT
1092    
1093    Tag: lal-20000727-linux-ppc
1094    Description:
1095    
1096     Made changes to support Linux PPC.
1097     p.s. I have confirmation that the 110.29 boot files work fine.
1098    
1099    ----------------------------------------------------------------------
1100    Name: Matthias Blume
1101    Date: 2000/07/27 17:40:00 JST
1102    Tag: blume-20000727-scripts
1103    Description:
1104    
1105    !!!! WARNING !!!!
1106    You must recompile the runtime system!
1107    !!!! WARNING !!!!
1108    
1109    This is basically another round of script-enhancements:
1110    
1111    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
1112       and undefine CM preprocessor symbols.
1113    
1114    2. ml-build avoids generating a new heap image if it finds that the
1115       existing one is still ok.  (The condition is that no ML file had to
1116       be recompiled and all ML files are found to be older that the heap
1117       file.)
1118    
1119       To make this work smoothly, I also hacked the runtime system as
1120       well as SMLofNJ.SysInfo to get access to the heap image suffix
1121       (.sparc-solaris, ...) that is currently being used.
1122    
1123       Moreover, the signature of CM.mk_standalone has changed.  See the
1124       CM manual.
1125    
1126    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
1127       CM manual for details.)
1128    
1129    4. More CM manual updates:
1130        - all of the above has been documented.
1131        - there is now a section describing the (CM-related) command line
1132          arguments that are accepted by the "sml" command
1133    
1134    ----------------------------------------------------------------------
1135    Name: Matthias Blume
1136    Date: 2000/07/25 16:20:00 JST
1137    Tag: blume-20000725-makedepend
1138    Description:
1139    
1140    Added a script called ml-makedepend.  This can be used in makefiles
1141    for Unix' make in a way very similar to the "makedepend" command for
1142    C.
1143    
1144    The script internally uses function CM.sources.
1145    
1146    Synopsis:
1147    
1148        ml-makedepend [-f makefile] cmfile targetname
1149    
1150    The default for the makefile is "makefile" (or "Makefile" should
1151    "makefile" not exist).
1152    
1153    ml-makedepend adds a cmfile/targetname-specific section to this
1154    makefile (after removing the previous version of this section).  The
1155    section contains a single dependency specification with targetname on
1156    the LHS (targetname is an arbitrary name), and a list of files derived
1157    from the cmfile on the RHS.  Some of the files on the RHS are
1158    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
1159    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
1160    path names.  The makefile writer is responsible for making sure that
1161    these variables have correct at the time "make" is invoked.
1162    
1163    ----------------------------------------------------------------------
1164    Name: Matthias Blume
1165    Date: 2000/07/22 23:30:00 JST
1166    Tag: blume-20000722-urlupdate
1167    Description:
1168    
1169    Changed BOOT and config/srcarchiveurl to point to BL server:
1170    
1171        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
1172    
1173    ----------------------------------------------------------------------
1174    Name: Matthias Blume
1175    Date: 2000/07/18 18:00:00 JST
1176    Tag: blume-20000718-Version_110_29
1177    Description:
1178    
1179    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
1180    
1181    2. Updated config/version to 110.29
1182    
1183    3. Updated config/srcarchiveurl
1184    
1185    3. New boot files!
1186       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
1187    
1188    ----------------------------------------------------------------------
1189    Name: Matthias Blume
1190    Date: 2000/07/11 13:58:00 JST
1191    Tag: blume-20000711-doctypo
1192    Description:
1193    
1194    Fixed a few typos in CM manual.
1195    
1196    ----------------------------------------------------------------------
1197    Name: Allen Leung
1198    Date: 2000/06/15 00:38:00
1199    Tag: leunga-20000704-sparc-x86
1200    
1201    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
1202    2. fix to x86 RET bug [found by Dan Grossman]
1203    3. sparc assembly bug fix for ticc instructions [found by Fermin]
1204    
1205       Affects c-- and moby only
1206    
1207    ----------------------------------------------------------------------
1208    Name: Matthias Blume
1209    Date: 2000/07/04 15:26:00
1210    Tag: blume-20000704-trigger
1211    Description:
1212    
1213    1. Improvements to CM manual.
1214    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
1215       of getting a back-trace.  The function, when called, raises an
1216       internal exception which explicitly carries the full back-trace history,
1217       so it is unaffected by any intervening handle-raise pairs ("trivial"
1218       or not).  The interactive loop will print that history once it arrives
1219       at top level.
1220       Short of having all exceptions implicitly carry the full history, the
1221       recommended way of using this facility is:
1222         - compile your program with instrumentation "on"
1223         - run it, when it raises an exception, look at the history
1224         - if the history is "cut off" because of some handler, go and modify
1225           your program so that it explicitly calls BTrace.trigger
1226         - recompile (still instrumented), and rerun; look at the full history
1227    
1228    ----------------------------------------------------------------------
1229    Name: Matthias Blume
1230    Date: 2000/07/03 15:36:00 JST
1231    Tag: blume-20000702-manual
1232    Description:
1233    
1234    Small corrections and updates to CM manual.
1235    
1236    ----------------------------------------------------------------------
1237    Name: Matthias Blume
1238    Date: 2000/06/29 16:04:00 JST
1239    Tag: blume-20000629-yacctool
1240    Description:
1241    
1242    Changes:
1243    
1244    1. Class "mlyacc" now takes separate arguments to pass options to
1245       generated .sml- and .sig-files independently.
1246    2. Corresponding CM manual updates.
1247    3. BTrace module now also reports call sites.  (However, for loop clusters
1248       it only shows from where the cluster was entered.)  There are associated
1249       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1250    
1251    ----------------------------------------------------------------------
1252    Name: Matthias Blume
1253    Date: 2000/06/27 16:51:00 JST
1254    Tag: blume-20000627-noweb
1255    Description:
1256    
1257    Changes:
1258    
1259     1. Implemented "subdir" and "witness" options for noweb tool.
1260        This caused some slight internal changes in CM's tool implementation.
1261     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
1262        some remaining issues from earlier path anchor changes.
1263     3. Updated CM manual accordingly.
1264    
1265     4. Changed implementation of back-tracing so that I now consider it
1266        ready for prime-time.
1267    
1268        In particular, you don't have to explicitly trigger the back-trace
1269        anymore.  Instead, if you are running BTrace-instrumented code and
1270        there is an uncaught exception (regardless of whether or not it was
1271        raised in instrumented code), the top-level evalloop will print
1272        the back-trace.
1273    
1274        Features:
1275    
1276          - Instrumented and uninstrumented code work together seemlessly.
1277            (Of course, uninstrumented code is never mentioned in actual
1278             back-traces.)
1279    
1280          - Asymptotic time- and space-complexity of instrumented code is
1281            equal to that of uninstrumented code.  (This means that
1282            tail-recursion is preserved by the instrumentation phase.)
1283    
1284          - Modules whose code has been instrumented in different sessions
1285            work together without problem.
1286    
1287          - There is no penalty whatsoever on uninstrumented code.
1288    
1289          - There is no penalty on "raise" expressions, even in
1290            instrumented code.
1291    
1292        A potential bug (or perhaps it is a feature, too):
1293    
1294          A back-trace reaches no further than the outermost instrumented
1295          non-trivial "raise".  Here, a "trivial" raise is one that is the
1296          sole RHS of a "handle" rule.  Thus, back-traces reach trough
1297    
1298               <exp> handle e => raise e
1299    
1300          and even
1301    
1302               <exp> handle Foo => raise Bar
1303    
1304          and, of course, through
1305    
1306               <exp> handle Foo => ...
1307    
1308         if the exception was not Foo.
1309    
1310         Back-traces always reach right through any un-instrumented code
1311         including any of its "handle" expressions, trivial or not.
1312    
1313       To try this out, do the following:
1314    
1315         - Erase all existing binfiles for your program.
1316           (You may keep binfiles for those modules where you think you
1317            definitely don't need back-tracing.)
1318         - Turn on back-trace instrumentation:
1319              SMLofNJ.Internals.BTrace.mode (SOME true);
1320         - Recompile your program.  (I.e., run "CM.make" or "use".)
1321         - You may now turn instrumentation off again (if you want):
1322              SMLofNJ.Internals.BTrace.mode (SOME false);
1323         - Run your program as usual.  If it raises an exception that
1324           reaches the interactive toplevel, then a back-trace will
1325           automatically be printed.  After that, the toplevel loop
1326           will print the exception history as usual.
1327    
1328    ----------------------------------------------------------------------
1329    Name: Matthias Blume
1330    Date: 2000/06/26 09:56:46 JST
1331    Tag: blume-20000626-setup
1332    Description:
1333    
1334    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1335          ML code before and after compiling a file (e.g., to set compiler
1336          flags)
1337    
1338    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1339              - associated changes to btrace.sml (BTrace instrumentation pass)
1340              - cleaner implementation of btimp.sml (BTrace tracing and report
1341                module)
1342    
1343    CM manual: * new path encoding documented
1344               * description of setup-parameter to "sml" added
1345    
1346    The biggest user-visible change to back-tracing is that it is no
1347    longer necessary to compile all traced modules within the same
1348    session.  (This was a real limitation.)
1349    
1350    ----------------------------------------------------------------------
1351    Name: Matthias Blume
1352    Date: 2000/06/24 12:40:00 JST
1353    Tag: blume-20000624-startup
1354    Description:
1355    
1356    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1357    bit too often -- to put it mildly. :)
1358    
1359    ----------------------------------------------------------------------
1360    Name: Matthias Blume
1361    Date: 2000/06/23 18:20:00 JST
1362    Tag: blume-20000623-btrace
1363    Description:
1364    
1365    This updates adds a backtrace facility to aid programmers in debugging
1366    their programs.  This involves the following changes:
1367    
1368    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1369       keeping track of the current call stack.  When programs are compiled
1370       in a special mode, the compiler will insert calls to these hooks
1371       into the user program.
1372       "Hook" means that it is possible for different implementations of
1373       back-tracing to register themselves (at different times).
1374    
1375    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1376       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1377       It can be turned on using this call:
1378         SMLofNJ.Internals.BTrace.mode (SOME true);
1379       Turning it off again:
1380         SMLofNJ.Internals.BTrace.mode (SOME false);
1381       Querying the current status:
1382         SMLofNJ.Internals.BTrace.mode NONE;
1383       Annotated programs are about twice as big as normal ones, and they
1384       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1385       where all hooks do nothing).  The slowdown with a plugin that is
1386       actually useful (such as the one supplied by default) is even greater,
1387       but in the case of the default plugin it is still only an constant
1388       factor (amortized).
1389    
1390    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1391       with a sub-structure BTrace for controlling back-tracing.  In particular,
1392       the above-mentioned function "mode" controls whether the annotation
1393       phase is invoked by the compiler.  Another important function is
1394       "trigger": when called it aborts the current execution and causes
1395       the top-level loop to print a full back-trace.
1396    
1397    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1398       for back-tracing.  It keeps track of the dynamic call stack and in
1399       addition to that it keeps a partial history at each "level" of that
1400       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1401       d and b (at separate times, dynamically), then the report will show:
1402    
1403       GOTO   d
1404             /c
1405       GOTO  \b
1406       CALL   a
1407    
1408       This shows that there was an initial non-tail call of a, then a
1409       tail-call to b or c, looping behavior in a cluster of functions that
1410       consist of b and c, and then a goto from that cluster (i.e., either from
1411       b or from c) to d.
1412    
1413       Note that (depending on the user program) the amount of information
1414       that the back-trace module has to keep track of at each level is bounded
1415       by a constant.  Thus, the whole implementation has the same asymptotical
1416       complexity as the original program (both in space and in time).
1417    
1418    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1419       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1420       which is raised by the "trigger" function mentioned above.
1421    
1422    Notes on usage:
1423    
1424    - Annotated code works well together with unannotated code:
1425    Unannotated calls simply do not show up at all in the backtrace.
1426    
1427    - It is not a good idea to let modules that were annotated during
1428    different sessions run at the same time.  This is because the compiler
1429    chooses small integers to identify individual functions, and there
1430    will be clashes if different modules were compiled in separate sessions.
1431    (Nothing will crash, and you will even be told about the clashes, but
1432    back-trace information will in general not be useful.)
1433    
1434    - Back-tracing can be confused by callcc and capture.
1435    
1436    - The only way of getting a back-trace right now is to explicitly
1437    invoke the "trigger" function from your user program.  Eventually, we
1438    should make every exception carry back-trace information (if
1439    available).  But since this creates more overhead at "raise"-time
1440    (similar to the current exnHistory overhead), I have not yet
1441    implemented this.  (The implementation will be rather easy.)  With
1442    exceptions carrying back-trace information, this facility will be even
1443    more useful because users don't need to modify their programs...
1444    
1445    - While it is possible to compile the compiler with back-trace
1446    annotations turned on (I did it to get some confidence in
1447    correctness), you must make absolutely sure that core.sml and
1448    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1449    be compiled with annotation because there is no core access yet, but
1450    if you compile btimp.sml with annotation, then the system will go into
1451    an infinite recursion and crash.)
1452    Since CM currently does not know about BTrace, the only way to turn
1453    annotations on and off for different modules of the compiler is to
1454    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1455    this is awkward and clumsy.
1456    
1457    Sample sessions:
1458    
1459    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1460    - SMLofNJ.Internals.BTrace.mode (SOME true);
1461    [autoloading]
1462    [autoloading done]
1463    val it = false : bool
1464    - structure X = struct
1465    -     fun main n = let
1466    -         fun a (x, 0) = d x
1467    -           | a (x, n) = b (x, n - 1)
1468    -         and b (x, n) = c (x, n)
1469    -         and c (x, n) = a (x, n)
1470    -         and d x = e (x, 3)
1471    -         and e (x, 0) = f x
1472    -           | e (x, n) = e (x, n - 1)
1473    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1474    -           | f n = n * g (n - 1)
1475    -         and g n = a (n, 3)
1476    -     in
1477    -         f n
1478    -     end
1479    - end;
1480    structure X : sig val main : int -> int end
1481    - X.main 3;
1482    *** BACK-TRACE ***
1483    GOTO   stdIn:4.2-13.20: X.main[2].f
1484    GOTO-( stdIn:4.2-13.20: X.main[2].e
1485    GOTO   stdIn:4.2-13.20: X.main[2].d
1486         / stdIn:4.2-13.20: X.main[2].a
1487         | stdIn:4.2-13.20: X.main[2].b
1488    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1489    CALL   stdIn:4.2-13.20: X.main[2].g
1490    GOTO   stdIn:4.2-13.20: X.main[2].f
1491    GOTO-( stdIn:4.2-13.20: X.main[2].e
1492    GOTO   stdIn:4.2-13.20: X.main[2].d
1493         / stdIn:4.2-13.20: X.main[2].a
1494         | stdIn:4.2-13.20: X.main[2].b
1495    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1496    CALL   stdIn:4.2-13.20: X.main[2].g
1497    GOTO   stdIn:4.2-13.20: X.main[2].f
1498    GOTO-( stdIn:4.2-13.20: X.main[2].e
1499    GOTO   stdIn:4.2-13.20: X.main[2].d
1500         / stdIn:4.2-13.20: X.main[2].a
1501         | stdIn:4.2-13.20: X.main[2].b
1502    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1503    CALL   stdIn:4.2-13.20: X.main[2].g
1504    GOTO   stdIn:4.2-13.20: X.main[2].f
1505    CALL   stdIn:2.15-17.4: X.main[2]
1506    -
1507    
1508    (Note that because of a FLINt bug the above code currently does not
1509    compile without BTrace turned on.)
1510    
1511    Here is another example, using my modified Tiger compiler:
1512    
1513    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1514    - SMLofNJ.Internals.BTrace.mode (SOME true);
1515    [autoloading]
1516    [autoloading done]
1517    val it = false : bool
1518    - CM.make "sources.cm";
1519    [autoloading]
1520    ...
1521    [autoloading done]
1522    [scanning sources.cm]
1523    [parsing (sources.cm):parse.sml]
1524    [creating directory CM/SKEL ...]
1525    [parsing (sources.cm):tiger.lex.sml]
1526    ...
1527    [wrote CM/sparc-unix/semant.sml]
1528    [compiling (sources.cm):main.sml]
1529    [wrote CM/sparc-unix/main.sml]
1530    [New bindings added.]
1531    val it = true : bool
1532    - Main.compile ("../testcases/merge.tig", "foo.out");
1533    *** BACK-TRACE ***
1534    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
1535    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1536    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1537    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1538    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1539    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1540    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
1541         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
1542    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
1543    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1544    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
1545    CALL   main.sml:1.18-118.4: Main.compile[2]
1546    -
1547    
1548    ----------------------------------------------------------------------
1549    Name: Matthias Blumen
1550    Date: 2000/06/21 18:00:00 JST
1551    Tag: blume-20000621-manual
1552    Description:
1553    
1554    CM manual update: Path environments documented.
1555    
1556    ----------------------------------------------------------------------
1557    Name: Matthias Blume
1558    Date: 2000/06/19 13:40:00
1559    Tag: blume-20000619-manual
1560    Description:
1561    
1562    CM manual and system/README update.  This only covers the fact that
1563    there are no more implicit anchors.  (Path environments and the "bind"
1564    option to "cm" have yet to be documented.)
1565    
1566    ----------------------------------------------------------------------
1567    Name: Matthias Blume
1568    Date: 2000/06/19 11:05:00 JST
1569    Tag: blume-20000619-chdir-bugfix
1570    Description:
1571    
1572    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1573    
1574    ----------------------------------------------------------------------
1575    Name: Matthias Blume
1576    Date: 2000/06/18 22:00:10 JST
1577    Tag: blume-20000618-implicit-anchors-really-gone
1578    Description:
1579    
1580    I updates the previous HISTORY entry where I forgot to mention that
1581    implicit anchors are no longer with us.
1582    
1583    The current update also gets rid of the (now useless) controller
1584    CM.Control.implicit_anchors.
1585    
1586    ----------------------------------------------------------------------
1587    Name: Matthias Blume
1588    Date: 2000/06/16 17:30:00 JST
1589    Tag: blume-20000616-anchorenv
1590    Description:
1591    
1592    This patch implements the long anticipated (just kidding :) "anchor
1593    environment" mechanism.  In the course of doing this, I also
1594    re-implemented CM's internal "SrcPath" module from scratch.  The new
1595    one should be more robust in certain boundary cases.  In any case, it
1596    is a lot cleaner than its predecessor (IMHO).
1597    
1598    This time, although there is yet another boot file format change, I
1599    kept the unpickler backward-compatible.  As a result, no new bootfiles
1600    are necessary and bootstrapping is straightforward.  (You cannot read
1601    new bootfiles into an old system, but the other way around is no
1602    problem.)
1603    
1604    Visible changes:
1605    
1606    ** 0. Implicit path anchors (without the leading $-symbol) are no
1607    longer recognized at all. This means that such path names are not
1608    illegal either.  For example, the name basis.cm simply refers to a
1609    local file called "basis.cm" (i.e, the name is an ordinary path
1610    relative to .cm-files directory).  Or, to put it differently, only
1611    names that start with $ are anchored paths.
1612    
1613    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1614    vanished.
1615    
1616    John (Reppy) had critizised this as soon as I originally proposed and
1617    implemented it, but at that time I did not really deeply believe
1618    him. :) Now I came full-circle because I need the $<singlearc> syntax
1619    in another place where it cannot be seen as an abbreviation for
1620    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1621    seems to mean (i.e., it "expands" into the corresponding anchor
1622    value).
1623    
1624    However, when paths are used as members in CM description files, it
1625    continues to be true that there must be at least another arc after the
1626    anchor.  This is now enforced separately during semantic analysis
1627    (i.e., from a lexical/syntactical point of view, the notation is ok.)
1628    
1629    ** 2. The "cm" class now accepts an option "bind".  The option's value
1630    is a sub-option list of precisely two items -- one labeled "anchor"
1631    and the other one labeled "value".  As you might expect, "anchor" is
1632    used to specify an anchor name to be bound, and "value" specifies what
1633    the anchor is being bound to.
1634    
1635    The value must be a directory name and can be given in either standard
1636    syntax (including the possibility that it is itself an anchored path)
1637    or native syntax.
1638    
1639    Examples:
1640    
1641       foo.cm (bind:(anchor:bar value:$mystuff/bar))
1642       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1643    
1644    and so on.
1645    
1646    The meaning of this is that the .cm-file will be processed with an
1647    augmented anchor environment where the given anchor(s) is/are bound to
1648    the given values(s).
1649    
1650    The rationale for having this feature is this: Suppose you are trying
1651    to use two different (already stable) libraries a.cm and b.cm (that
1652    you perhaps didn't write yourself).  Further, suppose each of these
1653    two libraries internally uses its own auxiliary library $aux/lib.cm.
1654    Normally you would now have a problem because the anchor "lib" can not
1655    be bound to more than one value globally.  Therefore, the project that
1656    uses both a.cm and b.cm must locally redirect the anchor to some other
1657    place:
1658    
1659       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1660       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1661    
1662    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1663    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
1664    
1665    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1666    you when you do so at the time of CM.stabilize).  Therefore, you can
1667    also use an anchored path as the value:
1668    
1669      a.cm (bind:(anchor:lib value:$a-lib))
1670      b.cm (bind:(anchor:lib value:$b-lib))
1671    
1672    Now you can globally configure (using the usual CM.Anchor.anchor or
1673    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1674    itself is always locally bound, setting it globally is no longer
1675    meaningful or necessary (but it does not hurt either).  In fact, "lib"
1676    can still be used as a global anchor for separate purposes.  As a
1677    matter of fact, one can locally define "lib" in terms of a global
1678    "lib":
1679    
1680      a.cm (bind:(anchor:lib value:$lib/a))
1681      b.cm (bind:(anchor:lib value:$lib/b))
1682    
1683    ** 3: The encoding of path names has changed.  This affects the way
1684    path names are shown in CM's progress report and also the internal
1685    protocol encoding used for parallel make.
1686    
1687    The encoding now uses one or more ':'-separated segments.  Each
1688    segments corresponds to a file that has been specified relative to the
1689    file given by its preceding segment.  The first segment is either
1690    relative to the CWD, absolute, or anchored.  Each segment itself is
1691    basically a Unix pathname; all segments but the first are relative.
1692    
1693    Example:
1694    
1695       $foo/bar/baz.cm:a/b/c.sml
1696    
1697    This path denotes the file bar/a/b/c.sml relative to the directory
1698    denoted by anchor "foo".  Notice that the encoding also includes
1699    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1700    paths are resolved relative to the .cm-files directory, so baz.cm must
1701    be ignored to get the "real" pathname.
1702    
1703    To make this fact more obvious, CM puts the names of such "virtual
1704    arcs" into parentheses when they appear in progress reports. (No
1705    parentheses will appear in the internal protocol encoding.)  Thus,
1706    what you really see is:
1707    
1708      $foo/bar/(baz.cm):a/b/c.sml
1709    
1710    I find this notation to be much more informative than before.
1711    
1712    Another new feature of the encoding is that special characters
1713    including parentheses, colons, (back)slashes, and white space are
1714    written as \ddd (where ddd is the decimal encoding of the character).
1715    
1716    *** The CM manual still needs to be updated.
1717    
1718    ----------------------------------------------------------------------
1719    Name: Allen Leung
1720    Date: 2000/06/15 00:38:00
1721    Tag: leunga-20000615-x86-peephole
1722    
1723    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1724    
1725    ----------------------------------------------------------------------
1726    Name: Matthias Blume
1727    Date: 2000/06/12 11:40:00
1728    Tag: blume-20000612-parmakefix
1729    Description:
1730    
1731    More cleanup after changing the file naming scheme: This time I
1732    repaired the parallel make mechanism for CMB.make which I broke earlier.
1733    
1734    ----------------------------------------------------------------------
1735    Name: Allen Leung
1736    Date: 2000/06/09 01:25:00
1737    Tag: leunga-20000609-various
1738    
1739    None of these things should affect normal SML/NJ operations
1740    
1741    1. Peephole improvements provided by Fermin (c--)
1742    2. New annotation DEFUSE for adding extra dependence (moby)
1743    3. New X86 LOCK instructions (moby)
1744    4. New machine description language for reservation tables (scheduling)
1745    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1746       trees etc.)
1747    6. I've changed the CM files so that they can work with versions
1748       110.0.6, 110.25 and 110.28
1749    
1750    ----------------------------------------------------------------------
1751    Name: Matthias Blume
1752    Date: 2000/06/09 12:40:00
1753    Tag: blume-20000609-log
1754    Description:
1755    
1756    - Removed all(?) remaining RCS Log entries from sources.
1757    
1758    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1759      anchored paths).
1760    
1761    ----------------------------------------------------------------------
1762    Name: Matthias Blume
1763    Date: 2000/06/07 17:00:00 JST
1764    Tag: blume-20000607-no-implicit-anchors
1765    Description:
1766    
1767    1. This update changes the default setting for
1768    CM.Control.implicit_anchors from true to false.  This means that
1769    implicit anchors are no longer permitted by default.  I also tried to
1770    make sure that nothing else still relies on implicit anchors.
1771    (This is the next step on the schedule towards a CM that does not even
1772    have the notion of implicit anchors anymore.)
1773    
1774    2. More CM manual updates.
1775    
1776    3. I managed to track down and fix the pickling bug I mentioned last
1777    time.  Because of the previously existing workaround, this entails no
1778    immediate practical changes.
1779    
1780    ----------------------------------------------------------------------
1781    Name: Matthias Blume
1782    Date: 2000/06/06 11:15:00 JST
1783    Tag: blume-20000606-lazierpickle
1784    Description:
1785    
1786    !!!! NEW BOOT FILES !!!!
1787    
1788    * The main purpose of this update is to make library pickles lazier in
1789    order to reduce the initial space penalty for autoloading a library.
1790    As a result, it is now possible to have $smlnj/compiler.cm
1791    pre-registered.  This should take care of the many complaints or
1792    inquiries about missing structure Compiler.  This required changes to
1793    CM's internal data structures and small tweaks to some algorithms.
1794    
1795    As a neat additional effect, it is no longer necessary (for the sake
1796    of lean heap image files) to distinguish between a "minimal" CM and a
1797    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1798    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1799    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1800    
1801    To make the life of compiler-hackers easier, "makeml" now also
1802    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1803    words, after you bootstrap a new sml for the first time, you will not
1804    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1805    around you will still have to do it, though.)
1806    
1807    * A second change consists of major updates to the CM manual.  There
1808    are now several appendices with summary information and also a full
1809    specification of the CM description file syntax.
1810    
1811    * In directory src/system I added the script "allcross".  This script
1812    invokes sml and cross-compiles the compiler for all supported
1813    architectures.  (Useful when providing a new set of boot files.)
1814    
1815    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1816    added a small tweak to pickle-util.sml to work around this problem,
1817    but it is not a proper fix yet.  I will investigate further.  (The
1818    effect of the bug was an inflation of library pickle size.)
1819    
1820    * Version number increased to 110.28.1 (to avoid compatibility problems).
1821    
1822    ----------------------------------------------------------------------
1823    Name: Allen Leung
1824    Date: 2000/05/25 17:28 EDT
1825    Tag: leunga-20000525-ra
1826    Description:
1827    
1828      Fixed a bug in freezing phase of the register allocator.
1829    
1830    ----------------------------------------------------------------------
1831    Name: Allen Leung
1832    Date: 2000/05/15 22:53 EDT
1833    Tag: leunga-20000515-alpha-x86-ra
1834    Description:
1835    
1836      1. Alpha
1837    
1838          Slight cleanup.  Removed the instruction SGNXL
1839    
1840      2. X86
1841    
1842          Added the following instructions to the instruction set:
1843    
1844            ROLx, RORx,
1845            BTx, BTSx, BTLx, BTRx,
1846            XCHGx, and variants with the LOCK prefix
1847    
1848      3. Register Allocation
1849    
1850          The module ra-rewrite-with-renaming has been improved.
1851    
1852      These have no effect on SML/NJ.
1853    
1854    ----------------------------------------------------------------------
1855    Name: Matthias Blume
1856    Date: 2000/05/15 16:20:00 JST
1857    Tag: blume-20000515-lightrebuild
1858    Description:
1859    
1860    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1861       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1862       defined.  In effect, the command will not build any cross-compiler
1863       backends and therefore finish more quickly.
1864    
1865       The "fixpt" script also takes a "-light" switch to be able to use
1866       this new facility while compiling for a fixpoint.
1867    
1868    2. I replaced all mentions of anchored paths in group owner specifications
1869       with simple relative paths (usually starting with "..").
1870       The rationale is that a library's internal workings should not be
1871       compromised by the lack of some anchor.  (An anchor is necessary
1872       for someone who wants to refer to the library by an anchored path,
1873       but it should not be necessary to build the same library in the first
1874       place.)
1875    
1876    3. I changed the way CM's tool mechanism determines the shell command
1877       string used for things like ml-yacc etc. so that it does not break
1878       when CM.Control.implicit_anchors is turned off.
1879    
1880    ----------------------------------------------------------------------
1881    Name: Matthias Blume
1882    Date: 2000/05/12 18:20:00 JST
1883    Tag: blume-20000512-ml-build
1884    Description:
1885    
1886    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1887    getting installed properly (by config/install.sh).
1888    
1889    ----------------------------------------------------------------------
1890    Name: Matthias Blume
1891    Date: 2000/05/12 17:30:00 JST
1892    Tag: blume-20000512-anchors
1893    Description:
1894    
1895    !!! NEW BOOT FILES !!!
1896    
1897    This change is in preparation of fading out support for "implicitly
1898    anchored path names".  I went through all sources and used the
1899    explicit (and relatively new) $-notation.  See system/README and the
1900    CM manual for more info on this.
1901    
1902    I also modified the anchoring scheme for some things such as "smlnj",
1903    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1904    anchors are more expressive: anchor name and first arc do not have to
1905    coincide.  This entails the following user-visible change:
1906    
1907    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1908    particular, when you fire up sml with a command-line argument, say,
1909    e.g.:
1910    
1911       sml '$smlnj/cmb.cm'
1912    
1913    At the ML toplevel prompt:
1914    
1915       CM.autoload "$smlnj/cmb.cm";
1916    
1917    There is also a new controller in CM.Control that can be used to turn
1918    off all remaining support for implicit anchors by saying:
1919    
1920        CM.autoload "$smlnj/
1921        #set CM.Control.implicit_anchors false;
1922    
1923    This causes CM to reject implicitly anchored paths.  This is (for the
1924    time being) less permissive than the "final" version where there will
1925    be no more such implicit anchors and relative paths will be just that:
1926    relative.
1927    
1928    The next step (version after next version?) will be to make the
1929    default for CM.Control.implicit_anchors false.  After the dust has
1930    settled, I can then produce the "final" version of this...
1931    
1932    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1933    
1934    ----------------------------------------------------------------------
1935    Name: Matthias Blume
1936    Date: 2000/05/11 16:30:00 JST
1937    Tag: blume-20000511-sources
1938    Description:
1939    
1940    The main change is that I added function CM.sources as a generalized
1941    version of the earlier CM.makedepend.  This entails the following
1942    additional changes:
1943    
1944      - CM.makedepend has been dropped.
1945    
1946      - CM manual has been updated.
1947    
1948      - TOOLS signature and API have been changed.
1949    
1950    ----------------------------------------------------------------------
1951    Name: Allen Leung
1952    Date: 2000/05/10 21:17 EDT
1953    Tag: leunga-20000510-moby-c--ssa
1954    Description:
1955    
1956      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1957    None of these affect SML/NJ.
1958    
1959    1. Register Allocation
1960    
1961        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1962           This module tries to remove local (i.e. basic block level) redundancies
1963           during spilling.
1964    
1965        b. A new framework for performing region based register allocation.
1966           Not yet entirely functional.
1967    
1968    2. X86
1969    
1970       a. DefUse for POP was missing the stack pointer [found by Lal]
1971       b. Reload for CALL was incorrect in X86Spill [found by John]
1972       c. Various fixes in X86Spill so that it can be used correctly for
1973          the new spilling module.
1974    
1975    3. SSA/IR
1976    
1977       a. New module ir/dj-dataflow.sml implements elimination based
1978          data flow analysis.
1979    
1980    4. MLRiscGen
1981    
1982       a. Fix for gc type annotation
1983    
1984    5. MDGen
1985    
1986       Various fixes for machine description -> ml code translation.  For ssa
1987       only.
1988    
1989    ----------------------------------------------------------------------
1990    Name: Allen Leung
1991    Date: 2000/05/08 22:17 EDT
1992    Tag: leunga-20000508-labexp
1993    Description:
1994    
1995      Fermin has found a few assembly problems with constant expressions
1996      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1997      which choke on dumb assemblers.  This is his fix.
1998    
1999    ----------------------------------------------------------------------
2000    Name: Dave MacQueen
2001    Date: 2000/04/09 14:00 EDT
2002    Tag: dbm-20000502-Version_110_28
2003    Description:
2004    
2005    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
2006    
2007    2. Updated config/version to 110.28
2008    
2009    3. Updated config/srcarchiveurl
2010    
2011    3. New boot files!
2012       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
2013    
2014    ----------------------------------------------------------------------
2015    Name: Matthias Blume
2016    Date: 2000/05/01 19:05:00 JST
2017    Tag: blume-20000501-noweb
2018    Description:
2019    
2020    A new noweb tool has been added.  The existing system is entirely
2021    unaffected by this, but some CM users have asked for renewed noweb
2022    support.  Everything is documented in the CM manual.
2023    
2024    New (plugin) libraries:
2025    
2026       noweb-tool.cm
2027       nw-ext.cm
2028    
2029    ----------------------------------------------------------------------
2030    Name: Dave MacQueen
2031    Date: 2000/04/30 12:40PM EDT
2032    Tag: dbm-20000430-bug_fixes
2033    Description:
2034    
2035    1. Fix for bug 1498
2036       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
2037       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
2038         added toRealArray function
2039       smlnj/src/compiler/MiscUtil/print/ppobj.sml
2040         added check for tag Obj.RealArray to array printing case in ppObj
2041    
2042    2. Fix for bug 1510
2043       smlnj/src/compiler/Semant/types/typesutil.sml
2044         fixed definition of dummyargs (used by equalTycon) so that
2045         dummy args are distinct types
2046    
2047    ----------------------------------------------------------------------
2048    Name: Matthias Blume
2049    Date: 2000/04/30 01:00:00 JST
2050    Tag: blume-20000430-versions
2051    Description:
2052    
2053    1. CM version numbering added.  This is an implementation of Lal's
2054       proposal for adding version numbers and version checking to .cm
2055       files.  Lal said that his proposal was just that -- a proposal.
2056       For the time being I went ahead and implemented it so that people
2057       can comment on it.  Everything is completely backward-compatible
2058       (except for the stable library format, i.e., new bootfiles!).
2059    
2060       As usual, see the CM manual for details.
2061    
2062    2. An alternative syntax for anchored paths has been implemented.
2063       Dave has recently voiced the same concerns that I had when I did
2064       this, so there should be some support.  My take is that eventually
2065       I will let support for the current syntax (where anchors are
2066       "implicit") fade out in favor of the new, explicit syntax.
2067       In order to be backward-compatible, both old and new syntax are
2068       currently supported.
2069    
2070       Again, see the CM manual for details.
2071    
2072    3. Parallel make is trying to be slightly smarter:  When the master
2073       process finds a "bottleneck", i.e., when there is only one
2074       compilation unit that can be compiled and everybody else is
2075       waiting on it, then it will simply compile it directly instead
2076       of clumsily telling one of the slaves to do it.
2077    
2078    4. Support for "unsharing" added.  This is necessary in order to be
2079       able to have two different versions of the same library running
2080       at the same time (e.g., for trying out a new MLRISC while still
2081       having the old MLRISC linked into the current compiler, etc.)
2082       See the CM manual.
2083    
2084    5. Simple "makedepend" functionality added for generating Makefile
2085       dependency information.  (This is rather crude at the moment.
2086       Expect some changes here in the future.)
2087    
2088    6. ".fun" added as a recognized suffix for ML files. Also documented
2089       explicitly in the manual that the fallback behavior (unknown suffix
2090       -> ML file) is not an official feature!
2091    
2092    7. Small changes to the pickler for stable libraries.
2093    
2094    8. Several internal changes to CM (for cleanup/improvement).
2095    
2096    
2097    !!!! NEW BINFILES !!!!
2098    
2099    ----------------------------------------------------------------------
2100    Name: Matthias Blume
2101    Date: 2000/04/28 17:30:00 JST
2102    Tag: blume-20000428-pathconfig
2103    Description:
2104    
2105    1. I changed config/install.sh to remove duplicate entries from the
2106       lib/pathconfig file at the end.  Moreover, the final version of
2107       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
2108       in src/system/installml.
2109    
2110    2. The config/install.sh script now consistently uses relative
2111       pathnames in lib/pathconfig whenever the anchor is in the lib
2112       directory.  (So far this was true for the libraries that come
2113       pre-compiled and bundled as part of the bootfiles but not for
2114       libraries that are compiled by the script itself.)
2115    
2116    ----------------------------------------------------------------------
2117    Name: Matthias Blume
2118    Date: 2000/04/26 13:10:00 JST
2119    Tag: blume-20000426-fun_suffix
2120    Description:
2121    
2122    Added ".fun" as a recognized file name suffix (for ML code).
2123    
2124    ----------------------------------------------------------------------
2125    Name: Allen Leung
2126    Date: 2000/04/25 17:00:00 EST
2127    Tag: leunga-20000425-alpha-ra
2128    Description:
2129    
2130    1. Alpha
2131    
2132        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
2133    in C--.
2134    
2135    2. RA
2136    
2137       Added a flag "ra-dump-size" to print out the size of the flowgraph
2138       and the interference graph.
2139    
2140    ----------------------------------------------------------------------
2141    Name: Dave MacQueen
2142    Date: 2000/04/25/
2143    Tag: dbm-20000425-mlyacc_doc_examples
2144    Description:
2145      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
2146      Updated all three examples in src/ml-yacc/examples to run
2147      under 110.* using CM.make.
2148    
2149    ----------------------------------------------------------------------
2150    Name: Allen Leung
2151    Date: 2000/04/20 23:04:00 EST
2152    Tag: leunga-20000420-ssa-c---stuff
2153    Description:
2154    
2155      This update synchronizes my repository with Yale's.  Most of these
2156    changes, however, do not affect SML/NJ at all (the RA is an exception).
2157    
2158    1. Register Allocator
2159    
2160       a. An improvement in the interference graph construction:
2161          Given a copy
2162    
2163                s <- t
2164    
2165          no interference edge between s and t is added for this definition of s.
2166    
2167       b. I've added two new spill heuristic modules that Fermin and I developed
2168          (in the new library RA.cm). These are unused in SML/NJ but maybe
2169          useful for others (Moby?)
2170    
2171    2. X86
2172    
2173       a. Various fixes in the backend provided by Fermin [C--] and Lal.
2174    
2175    3. Alpha
2176    
2177       a. Added the BSR instruction and code generation that goes with it [C--]
2178       b. Other fixes too numerous to recount provided by Fermin [C--]
2179    
2180    4. Regmaps
2181    
2182       a. The regmaps are not initialized with the identity physical bindings
2183          at creation time.  This is unneeded.
2184    
2185    5. MLRISC Optimizations
2186    
2187       a. The DJ-Graph module can now compute the iterated dominance frontiers
2188          intersects with liveness incrementally in linear time! Woohoo!
2189          This is now used in my new SSA construction algorithm.
2190    
2191       b. THe branch reorganization module is now smarter about linear chains of
2192          basic blocks.
2193    
2194    
2195    ----------------------------------------------------------------------
2196    Name: Matthias Blume
2197    Date: 2000/04/12 13:52:00 JST
2198    Tag: blume_main_v110p27_1
2199    Description:
2200    
2201    Changed install.sh script to handle archive files without version number
2202    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
2203    name of the boot file archive.
2204    
2205    ----------------------------------------------------------------------
2206    Name: Dave MacQueen
2207    Date: 2000/04/09 14:00 EDT
2208    Tag: dbm-20000410-Version_110_27
2209    Description:
2210    
2211    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
2212    
2213    2. Updated src/config/version to 110.27
2214    
2215    3. New boot files!
2216    
2217    ----------------------------------------------------------------------
2218    Name: Allen Leung
2219    Date: 2000/04/09 19:09:00 EST
2220    Tag: leunga-20000409-misc
2221    Description:
2222    
2223    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
2224    
2225    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
2226    
2227  ----------------------------------------------------------------------  ----------------------------------------------------------------------
2228  Name: Stefan  Name: Stefan
2229  Date: 2000/04/07 10:00:00 EDT  Date: 2000/04/07 10:00:00 EDT
# Line 97  Line 2310 
2310    
2311  3.  Assembly  3.  Assembly
2312    
2313      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
2314      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
2315      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
2316    
# Line 120  Line 2333 
2333    
2334  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
2335  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
2336  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
2337  legal surface-syntax symbol).  legal surface-syntax symbol).
2338    
2339  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 130  Line 2343 
2343  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
2344    
2345  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2346  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
2347  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
2348  more details.  more details.
2349    
# Line 245  Line 2458 
2458     (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
2459     used).     used).
2460    
2461     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
2462     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
2463     Makefiles or Shell scripts.     Makefiles or Shell scripts.
2464    
2465     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 258  Line 2471 
2471    
2472     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
2473    
2474     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
2475     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
2476    
2477  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 392  Line 2605 
2605    
2606       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
2607       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
2608       testl is perferable to andl.       testl is preferable to andl.
2609    
2610  7. RA (x86 only)  7. RA (x86 only)
2611    
# Line 574  Line 2787 
2787    
2788  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2789  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2790  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
2791  explanation.  explanation.
2792    
2793  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2794  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2795  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2796  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 621  Line 2834 
2834  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
2835  <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,
2836  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
2837  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2838  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2839  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2840    
# Line 671  Line 2884 
2884    it from that remote directory.    it from that remote directory.
2885    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2886    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2887    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2888    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2889    
2890    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.605  
changed lines
  Added in v.847

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