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 640, Tue May 9 02:33:09 2000 UTC revision 854, Wed Jun 27 19:11:38 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: 2001/06/27 15:10:00 EDT
18    Tag: blume-20010627-concur
19    Description:
20    
21    Fixed a small bug in CM's handling of parallel compilation.
22    (You could observe the bug by Control-C-interrupting an ordinary
23    CMB.make or CM.stabilize and then attaching some compile servers.
24    The result was that all of a sudden the previously interrupted
25    compilation would continue on its own.  This was because of
26    an over-optimization: CM did not bother to clean out certain queues
27    when no servers were attached "anyway", resulting in the contents
28    of these queues to grab control when new servers did get attached.)
29    
30    There is also another minor update to the CM manual.
31    
32    ----------------------------------------------------------------------
33    Name: Matthias Blume
34    Date: 2001/06/26 16:15:00 EDT
35    Tag: blume-20010626-cmdoc
36    Description:
37    
38    Minor typo fixed in CM manual (syntax diagram for libraries).
39    
40    ----------------------------------------------------------------------
41    Name: Matthias Blume
42    Date: 2001/06/25 22:55:00 EDT
43    Tag: blume-20010625-x86pc
44    Description:
45    
46    Fixed a nasty bug in the X86 assembly code that caused signal
47    handlers to fail (crash) randomly.
48    
49    ----------------------------------------------------------------------
50    Name: Matthias Blume
51    Date: 2001/06/25 12:05:00 EDT
52    Tag: blume-20010625-nlffigen
53    Description:
54    
55    This update fixes a number of minor bugs in ml-nlffigen as reported by
56    Nick Carter <nbc@andrew.cmu.edu>.
57    
58      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
59      2. Default names for generated files are now derived from the name of
60         the C file *without its directory*.  In particular, this causes generated
61         files to be placed locally even if the C file is in some system directory.
62      3. Default names for generated signatures and structures are also derived
63         from the C file name without its directory.  This avoids silly things
64         like "structure GL/GL".
65         (Other silly names are still possible because ml-nlffigen does not do
66          a thorough check of whether generated names are legal ML identifiers.
67          When in doubt, use command line arguments to force particular names.)
68    
69    ----------------------------------------------------------------------
70    Name: Matthias Blume
71    Date: 2001/06/21 12:25:00 EDT
72    Tag: blume-20010621-eXene
73    Description:
74    
75    eXene now compiles and (sort of) works again.
76    
77    The library name (for version > 110.33) is $/eXene.cm.
78    
79    I also added an new example in src/eXene/examples/nbody.  See the
80    README file there for details.
81    
82    ----------------------------------------------------------------------
83    Name: Matthias Blume
84    Date: 2001/06/20 16:40:00 EDT
85    Tag: blume-20010620-cml
86    Description:
87    
88    CML now compiles and works again.
89    
90    Libraries (for version > 110.33):
91    
92      $cml/cml.cm            Main CML library.
93      $cml/basis.cm          CML's version of $/basis.cm.
94      $cml/cml-internal.cm   Internal helper library.
95      $cml/core-cml.cm       Internal helper library.
96      $cml-lib/trace-cml.cm  Tracing facility.
97      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
98    
99    The installer (config/install.sh) has been taught how to properly
100    install this stuff.
101    
102    ----------------------------------------------------------------------
103    Name: Matthias Blume
104    Date: 2001/06/19 17:55:00 EDT
105    Tag: blume-20010619-instantiate
106    Description:
107    
108    This un-breaks the fix for bug 1432.
109    (The bug was originally fixed in 110.9 but I broke it again some
110    time after that.)
111    
112    ----------------------------------------------------------------------
113    Name: Matthias Blume
114    Date: 2001/06/19 17:25:00 EDT
115    Tag: blume-20010619-signals
116    Description:
117    
118    This should (hopefully) fix the long-standing signal handling bug.
119    (The runtime system was constructing a continuation record with an
120    incorrect descriptor which would cause the GC to drop data on the floor...)
121    
122    ----------------------------------------------------------------------
123    Name: Matthias Blume
124    Date: 2001/06/15 15:05:00 EDT
125    Tag: blume-20010615-moresparc
126    Description:
127    
128    Here is a short late-hour update related to Sparc c-calls:
129    
130     -- made handling of double-word arguments a bit smarter
131    
132     -- instruction selection phase tries to collapse certain clumsily
133        constructed ML-Trees; typical example:
134    
135            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
136    
137        This currently has no further impact on SML/NJ since mlriscGen does
138        not seem to generate such patterns in the first place, and c-calls
139        (which did generate them in the beginning) has meanwhile been fixed
140        so as to avoid them as well.
141    
142    ----------------------------------------------------------------------
143    Name: Matthias Blume
144    Date: 2001/06/15 15:05:00 EDT
145    Tag: blume-20010615-sparc
146    Description:
147    
148    The purpose of this update is to provide an implementation of NLFFI
149    on Sparc machines.
150    
151    Here are the changes in detail:
152    
153       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
154       the Sparc implementation of the c-calls API.
155       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
156       for accessing the ML frame.  Thus, we have a real frame pointer and
157       can freely modify %sp without need for an omit-frame-ptr phase.
158       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
159       this case.
160       * ml-nlffigen has been taught to produce code for different architectures
161       and calling conventions.
162       * In a way similar to what was done in the x86 case, the Sparc
163       backend uses its own specific extension to mltree.  (For example,
164       it needs to be able to generate UNIMP instructions which are part
165       of the calling convention.)
166       * ml-nlffi-lib was reorganized to make it more modular (in particular,
167       to make it easier to plug in new machine- and os-dependent parts).
168    
169    There are some other fairly unrelated bug fixes and cleanups as well:
170    
171       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
172       that they properly share their libraries with existing SML/NJ libraries.
173       * I fixed a minor cosmetic bug in CM, supressing certain spurious
174       follow-up error messages.
175       * Updates to CM/CMB documentation.
176    
177    TODO items:
178    
179       * MLRISC should use a different register as its asmTemp on the Sparc.
180         (The current %o2 is a really bad choice because it is part of the
181         calling conventions, so things might interfere in unexpected ways.)
182    
183    ----------------------------------------------------------------------
184    Name: Matthias Blume
185    Date: 2001/06/07
186    Tag: blume-20010607-calls
187    Description:
188    
189    A number of internal changes related to C calls and calling conventions:
190    
191    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
192       number of bytes popped implicitly (by the callee).  In most cases
193       this field is 0 but on x86/win32 it is some non-zero value.  This
194       is information provided for the benefit of the "omit-frameptr" pass.
195    2. The CALL instruction on the x86 carries a similar "pops" field.
196       The instruction selection phase copies its value from the ML-Tree
197       CALL statement.
198    3. On all other architectures, the instruction selection phase checks
199       whether "pops=0" and complains if not.
200    4. The c-calls implementation for x86 now accepts two calling conventions:
201       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
202       up after the call and pops is set to 0.  For "stdcall", the caller
203       does nothing, leaving the cleanup to the callee; pops is set to
204       the number of bytes that were pushed onto the stack.
205    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
206       distinguish between "ccall" and "stdcall".
207    6. The UNIMP instruction has been added to the supported Sparc instruction
208       set. (This is needed for implementing the official C calling convention
209       on this architecture.)
210    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
211       work with the latest CM.
212    
213    ----------------------------------------------------------------------
214    Name: Matthias Blume
215    Date: 2001/06/05 15:10:00 EDT
216    Tag: blume-20010605-cm-index
217    Description:
218    
219    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
220    
221    1. CM can now generate "index files".  These are human-readable files
222       that list on a per-.cm-file basis each toplevel symbol defined or
223       imported.  The location of the index file for
224       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
225       To enable index-file generation, set CM.Control.generate_index to true
226       or export an environment-symbol: export CM_GENERATE_INDEX=true.
227    
228       The CM manual has been updated accordingly.
229    
230    2. I made some slight modifications to the c-calls API in MLRISC.
231    
232         a) There is now a callback to support saving/restoring of
233            dedicated but caller-save registers around the actual call
234            instruction.
235         b) One can optionally specify a comment-annotation for the
236            call instruction.
237    
238    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
239       (For example, the comment annotation shows the C prototype of
240        the function being called.)
241    
242    ----------------------------------------------------------------------
243    Name: Matthias Blume
244    Date: 2001/06/01 13:30:00 EDT
245    Tag: blume-20010601-nlffi-cleanup
246    Description:
247    
248    This is mostly a cleanup of MLFFI stuff:
249    
250       - some signature files have been put into a more exposed place
251       - the ugly 'f type parameter is gone (simplifies types tremendously!)
252       - ml-nlffigen changed accordingly
253       - tutorial updated
254    
255    Other changes:
256    
257       - author's affiliation in CM manual(s) updated
258       - some more recognized keywords added to Allen's sml.sty
259    
260    ----------------------------------------------------------------------
261    Name: Matthias Blume
262    Date: 2001/05/25 15:30:00 EDT
263    Tag: blume-20010525-iptr
264    Description:
265    
266      - put the official 110.33-README (as it appears on the ftp server) under
267        CVS
268      - fixed a small bug related to incomplete pointer types in
269        ml-nlffigen
270      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
271        (it does not need the 'f type parameter)
272    
273    ----------------------------------------------------------------------
274    Name: Matthias Blume
275    Date: 2001/05/23 14:30:00 EDT
276    Tag: Release_110_33
277    Description:
278    
279    New version number (110.33).  New bootfiles.
280    
281    ----------------------------------------------------------------------
282    Name: Matthias Blume
283    Date: 2001/05/22 18:06:00 EDT
284    Tag: blume-20010522-targets
285    Description:
286    
287    Made install.sh use file config/targets.customized if it exists, falling
288    back to config/targets if it doesn't.  This way one can have a customized
289    version of the targets file without touching the "real thing", thus
290    eliminating the constant fear of accidentally checking something bogus
291    back into the CVS repository...  (File config/targets.customized must
292    not be added to the repository!)
293    
294    ----------------------------------------------------------------------
295    Name: Matthias Blume
296    Date: 2001/05/22 16:30:00 EDT
297    Tag: blume-20010522-minitut
298    Description:
299    
300    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
301       struct returns.
302    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
303       incomplete, preliminary documentation for NLFFI.
304    
305    ----------------------------------------------------------------------
306    Name: Matthias Blume
307    Date: 2001/05/14 11:30:00 EDT
308    Tag: blume-20010514-script
309    Description:
310    
311    Some bugs in install script fixed.
312    
313    In addition to that I also made a slight change to the NLFFI API:
314    Functors generated by ml-nlffigen now take the dynamic library as a
315    straight functor argument, not as a suspended one.  (The original
316    functor code used to force the suspension right away anyway, so there
317    was nothing gained by this complication of the interface.)
318    
319    ----------------------------------------------------------------------
320    Name: Matthias Blume
321    Date: 2001/05/11 14:35:00 EDT
322    Tag: blume-20010511-ml-nlffi
323    Description:
324    
325    I finally took the plunge and added my new FFI code to the main
326    repository.  For x86-linux it is now ready for prime-time.
327    
328    There are two new subdirectories of "src":
329    
330      - ml-nlffi-lib:
331           The utility library for programs using the FFI interface.
332           Here is the implementation of $/c.cm and its associated low-level
333           partners $/c-int.cm and $/memory.cm.
334      - ml-nlffigen:
335           A stand-alone program for generating ML glue code from C source
336           code.
337    
338    Building ml-nlffigen requires $/ckit-lib.cm.
339    
340    The config/install.sh script has been updates to do the Right Thing
341    (hopefully).
342    
343    Notice that the source tree for the C-Kit will not be put under "src"
344    but directly under the installation root directory.  (This is the
345    structure that currently exists on the CVS server when you check out
346    module "sml".)  Fortunately, config/install.sh knows about this oddity.
347    
348    Bugs: No documentation yet.
349    
350    ----------------------------------------------------------------------
351    Name: Matthias Blume
352    Date: 2001/05/09 16:35:00 EDT
353    Tag: blume-20010509-cpscontract
354    Description:
355    
356    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
357    wrapper/unwrapper elimination did not decrement usage counts and some
358    dead variables got overlooked by the dead-up logic.)
359    
360    ----------------------------------------------------------------------
361    Name: Lal George
362    Date: 2001/05/08  17:26:09 EDT
363    Tag: george-20010508-omit-frameptr
364    Description:
365    
366    Changes to implement the omit-frame-pointer optimization to support
367    raw C calls. For now, there is only support on the Intel x86, but
368    other architectures will follow as more experience is gained with this.
369    
370    
371    ----------------------------------------------------------------------
372    Name: Matthias Blume
373    Date: 2001/05/07 14:40:00 EDT
374    Tag: blume-20010507-proxies
375    Description:
376    
377    I made into "proxy libraries" all libraries that qualify for such a
378    change.  (A qualifying library is a library that has another library or
379    groups as its sole member and repeats that member's export list
380    verbatim.  A proxy library avoids this repetition by omitting its export
381    list, effectively inheriting the list that its (only) member exports.
382    See the CM manual for more explanation.)
383    The main effect is that explicit export lists for these libraries
384    do not have to be kepts in sync, making maintenance a bit easier.
385    
386    I also added copyright notices to many .cm-files.
387    
388    Last but not least, I made a new set of bootfiles.
389    
390    ----------------------------------------------------------------------
391    Name: Matthias Blume
392    Date: 2001/05/04 17:00:00 EDT
393    Tag: blume-20010504-cm-lsplit
394    Description:
395    
396    0. John merged pending changes to $/smlnj-lib.cm
397    
398    1. Allen's previous change accidentally backed out of one of Lal's
399       earlier changes.  I undid this mistake (re-introducing Lal's change).
400    
401    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
402       within the compiler where applicable.  There is some code simplification
403       because of that.
404    
405    3. The "split" phase (in FLINT) is now part of the default list of phases.
406       Compiler.Control.LambdaSplitting.* can be used to globally control the
407       lambda-splitting (cross-module-inlining) engine.  In addition to that,
408       it can now also be controlled on a per-source basis: CM has been taught
409       a new tool parameter applicable to ML source files.
410    
411       - To turn lambda-splitting off completely:
412            local open Compiler.Control.LambdaSplitting in
413                val _ = set Off
414            end
415       - To make "no lambda-splitting" the global default (but allow per-source
416         overriding); this is the initial setting:
417            local open Compiler.Control.LambdaSplitting in
418                val _ = set (Default NONE)
419            end
420       - To make "lambda-splitting with aggressiveness a" the global default
421         (and allow per-source overriding):
422            local open Compiler.Control.LambdaSplitting in
423                val _ = set (Default (SOME a))
424            end
425    
426       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
427         write (in the respective .cm-file):
428            a.sml (lambdasplitting:off)
429       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
430            a.sml (lambdasplitting:on)
431       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
432         <a> is a decimal non-negative integer):
433            a.sml (lambdasplitting:<a>)
434       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
435            a.sml (lambdasplitting:infinity)
436       - To use the global default for a.sml:
437            a.sml (lambdasplitting:default)
438         or simply
439            a.sml
440    
441    ----------------------------------------------------------------------
442    Name: Allen Leung
443    Date: 2001/05/04 01:57:00 EDT
444    Tag: leunga-20010504-sync
445    Description:
446    
447      MLRISC features.
448    
449      1. Fix to CMPXCHG instructions.
450      2. Changed RA interface to allow annotations in callbacks.
451      3. Added a new method to the stream interface to allow annotations updates.
452    
453    ----------------------------------------------------------------------
454    Name: Matthias Blume
455    Date: 2001/05/01 11:45:00 EDT
456    Tag: blume-20010501-pcedittmp
457    Description:
458    
459    Changed install.sh to use the current working directory instead of
460    /usr/tmp for a temporary file (pcedittmp).  The previous choice
461    of /usr/tmp caused trouble with MacOS X because of file premission
462    problems.
463    
464    ----------------------------------------------------------------------
465    Name: Matthias Blume
466    Date: 2001/04/20 11:10:00 EDT
467    Tag: blume-20010420-inMLflag
468    Description:
469    
470     - added vp_limitPtrMask to vproc-state.h
471       (for use by the raw-C-calls mechanism to implement proper interrupt
472        handling)
473     - made the ML compiler aware of various data-structure offsets so it
474       can generate code for accessing the vp_inML flag and vp_limitPtrMask
475     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
476       raw C-calls
477    
478    ----------------------------------------------------------------------
479    Name: Lal George
480    Date: 2001/04/20 09:15:28 EDT
481    Tag: george-20010420-macosX
482    Description:
483    
484     - Changes to port to Mac OS X; Darwin.
485    
486     - In the process I found that sqrt was broken on the PPC, because the
487       fsqrt instruction is not implemented.
488    
489    ----------------------------------------------------------------------
490    Name: Matthias Blume
491    Date: 2001/04/18 12:45:00 EDT
492    Tag: blume-20010418-ccalls
493    Description:
494    
495     - fixed two off-by-4 errors in the x86-specific c-calls implementation
496       (this bug prevented structure arguments containing pointers from being
497        passed correctly)
498     - changed the raw-C-call code in mlriscGen.sml in such a way that
499       structure arguments are represented as a pointer to the beginning
500       of the structure (instead of having a series of synthesized arguments,
501       one for each structure member)
502    
503     - made makeml script's verbosity level configurable via environment
504       variable (MAKEML_VERBOSITY)
505    
506     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
507       in rawmem-x86.sml; we are now using the real thing
508    
509    ----------------------------------------------------------------------
510    Name: Matthias Blume
511    Date: 2001/03/22 16:25:00 EST
512    Tag: blume-20010322-bootfiles
513    Description:
514    
515    Created a new set of bootfiles (for your automatic installation convenience).
516    
517    ----------------------------------------------------------------------
518    Name: Matthias Blume
519    Date: 2001/03/22 15:10:00 EST
520    Tag: blume-20010322-rawmem-parcm
521    Description:
522    
523    1. All "raw memory access" primitives for the new FFI are implemented now
524       (at least on the x86).
525    2. Some further cleanup of CM's parallel make mechanism.
526    
527    ----------------------------------------------------------------------
528    Name: Matthias Blume
529    Date: 2001/03/19 17:53:00 EST
530    Tag: blume-20010319-parallel
531    Description:
532    
533    Parallel make (using compile servers) now works again.
534    
535    To this end, CM.stabilize and CMB.make have been modified to work in
536    two passes when compile servers are attached:
537       1. Compile everything, do not perform stabilization; this pass
538          uses compile servers
539       2. Stabilize everything; this pass does not use compile servers
540    If there are no compile servers, the two passes are combined into one
541    (as before).  Splitting the passes increases the inherent parallelism
542    in the dependency graph because the entire graph including all
543    libraries is available at the same time.  This, in turn, improves
544    server utilization.  The downside is that the master process will
545    have to do some extra work after compilation is done (because for
546    technical reasons it must re-read all the binfiles during stabilization).
547    
548    ----------------------------------------------------------------------
549    Name: Matthias Blume
550    Date: 2001/03/16 12:22:00 EST
551    Tag: blume-20010316-bootfiles
552    Description:
553    
554    Created a new set of bootfiles (for your automatic installation convenience).
555    
556    ----------------------------------------------------------------------
557    Name: Matthias Blume
558    Date: 2001/03/16 11:00:00 EST
559    Tag: blume-20010316-MLTREE-fixup
560    Description:
561    
562    This is a minor fixup for an (untagged) earlier commit by Allen.
563    (A file was missing).
564    
565    ----------------------------------------------------------------------
566    Name: Allen Leung
567    Date: Mon Mar  5 18:54:57 EST 2001
568    Tag: leunga-20010305-cut-support
569    
570    1. New support for alternative control-flow in MLTREE.
571       Currently we support
572    
573          FLOW_TO(CALL ...., [k1,...,kn])
574    
575       This is needed for 'cuts to' in C-- and try/handle-like constructs
576       in Moby
577    
578       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
579    
580    2. Register Allocator
581    
582       Changes in interface [from Fermin, John]
583    
584    3. Alpha 8-bit SLL support [Fermin]
585    
586    4. All architectures
587    
588       A new module (ClusterExpandCopies) for expanding parallel copies.
589    
590    ----------------------------------------------------------------------
591    Name: Allen Leung
592    Date: 2001/02/27 23:07:00 EST
593    Tag: leunga-20010227-minor-stuff
594    
595    1. Alpha bug fix for CMOVNE
596    2. Handle mltree COND(..,FCMP ...,...)
597    3. Bug fix in simplifier
598    
599    ----------------------------------------------------------------------
600    Name: Matthias Blume
601    Date: 2001/01/30 17:50:00 EST
602    Tag: blume-20010130-sync
603    Description:
604    
605    This is just a minor update to sync my devel branch with the main brach.
606    The only visible change is the addition of some README files.
607    
608    ----------------------------------------------------------------------
609    Name: Matthias Blume
610    Date: 2001/01/12 23:30:00 JST
611    Tag: blume-20010112-bootfiles
612    Description:
613    
614    Made a new set of bootfiles that goes with the current state of the
615    repository.
616    
617    ----------------------------------------------------------------------
618    Name: Matthias Blume
619    Date: 2001/01/12 21:20:00 JST
620    Tag: blume-20010112-sync
621    Description:
622    
623    I am just flushing out some minor changes that had accumulated in
624    my private branch in order to sync with the main tree.  (This is
625    mainly because I had CVS trouble when trying to merge _into_ my
626    private branch.)
627    
628    Most people should be completely unaffected by this.
629    
630    ----------------------------------------------------------------------
631    Name: Allen Leung
632    Date: Thu Jan 11 21:03:00 EST 2001
633    Tag: leunga-20010111-labexp=mltree
634    Description:
635    
636    1.  Removed the type LabelExp and replace it by MLTree.
637    2.  Rewritten mltree-simplify with the pattern matcher tool.
638    3.  There were some bugs in alpha code generator which would break
639        64-bit code generation.
640    4.  Redo the tools to generate code with the
641    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
642        are now generated by perl scripts.
643    
644    ----------------------------------------------------------------------
645    Name: Matthias Blume
646    Date: 2001/01/10 21:55:00 JST
647    Tag: blume-20010110-rcc
648    Description:
649    
650    The RCC stuff now seems to work (but only on the x86).
651    This required hacking of the c-calls interface (and -implementation) in
652    MLRISC.
653    
654    Normal compiler users should be unaffected.
655    
656    ----------------------------------------------------------------------
657    Name: Matthias Blume
658    Date: 2001/01/09 01:20:00 JST
659    Tag: blume-20010109-rcc
660    Description:
661    
662    This is a fairly big patch, flushing out a large number of pending
663    changes that I made to my development copy over the last couple of days.
664    
665    Of practical relevance at this moment is a workaround for a pickling
666    bug that Allen ran into the other day.  The cause of the bug itself is
667    still unknown and it might be hard to fix it properly, but the
668    workaround has some merits of its own (namely somewhat reducing pickling
669    overhead for certain libraries).  Therefore, I think this solution should
670    be satisfactory at this time.
671    
672    The rest of the changes (i.e., the vast majority) has to do with my
673    ongoing efforts of providing direct support for C function calls from
674    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
675    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
676    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
677    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
678    handle RCC.
679    
680    The last part (the changes to mlriscGen.sml) are still known to be
681    wrong on the x86 and not implemented on all other architectures.  But
682    the infrastructure is in place. I had to change a few functor
683    signatures in the backend to be able to route the CCalls interface
684    from MLRISC there, and I had to specialize the mltree type (on the
685    x86) to include the necessary extensions. (The extensions themselves
686    were already there and redy to go in MLRISC/x86).
687    
688    Everything should be very happy as soon as someone helps me with
689    mlriscGen.sml...
690    
691    In any case, nothing of this should matter to anyone as long as the
692    new primop is not being used (which is going to be the case unless you
693    find it where I hid it :). The rest of the compiler is completely
694    unaffected.
695    
696    ----------------------------------------------------------------------
697    Name: Matthias Blume
698    Date: 2001/01/05 00:30:00 JST
699    Tag: blume-20010105-primops
700    Description:
701    
702    Added some experimental support for work that I am doing right now.
703    These changes mostly concern added primops, but there is also a new
704    experimental C library in the runtime system (but currently not enabled
705    anywhere except on Linux/X86).
706    
707    In the course of adding primops (and playing with them), I discovered that
708    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
709    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
710    To recover, I made type information in INL_PRIM mandatory and changed
711    prim.sml as well as built-in.sml accordingly.  The InLine structure now
712    has complete, correct type information (i.e., no bottom types).
713    
714    Since all these changes mean that we need new binfiles, I also bumped the
715    version number to 110.32.1.
716    
717    ----------------------------------------------------------------------
718    Name: Matthias Blume
719    Date: 2000/12/30 22:10:00 JST
720    Tag: blume-20001230-various
721    Description:
722    
723    Added proxy libraries for MLRISC and let MLRISC libraries refer
724    to each other using path anchors.  (See CM manual for explanation.)
725    
726    Updated CM documentation.
727    
728    Fixed some bugs in CM.
729    
730    Implemented "proxy" libraries (= syntactic sugar for CM).
731    
732    Added "-quiet" option to makeml and changed runtime system accordingly.
733    
734    Added cleanup handler for exportML to reset timers and compiler stats.
735    
736    ----------------------------------------------------------------------
737    Name: Lal George
738    Date: 2000/12/22 22:22:58 EST 2000
739    Tag: Release_110_32
740    Description:
741    
742            Infinite precision used throughout MLRISC.
743            see MLRISC/mltree/machine-int.sig
744    
745    ----------------------------------------------------------------------
746    Name: Matthias Blume
747    Date: 2000/12/22 23:16:00 JST
748    Tag: blume-20001222-warn
749    Description:
750    
751    Corrected wording and formatting of some CM warning message which I
752    broke in my previous patch.
753    
754    ----------------------------------------------------------------------
755    Name: Matthias Blume
756    Date: 2000/12/22 21:20:00 JST
757    Tag: blume-20001222-anchorenv
758    Description:
759    
760    Fixed CM's handling of anchor environments in connection with CMB.make.
761    
762    ----------------------------------------------------------------------
763    Name: Matthias Blume
764    Date: 2000/12/22 13:15:00 JST
765    Tag: blume-20001222-cleanup
766    Description:
767    
768    Removed src/cm/ffi which does not (and did not) belong here.
769    
770    ----------------------------------------------------------------------
771    Name: Matthias Blume
772    Date: 2000/12/21 23:55:00 JST
773    Tag: blume-20001221-exn
774    Description:
775    
776    Probably most important: CM no longer silently swallows all exceptions
777    in the compiler.
778    Plus: some other minor CM changes.  For example, CM now reports some
779    sizes for generated binfiles (code, data, envpickle, lambdapickle).
780    
781    ----------------------------------------------------------------------
782    Name: Matthias Blume
783    Date: 2000/12/15 00:01:05 JST
784    Tag: blume-20001215-dirtool
785    Description:
786    
787    - "dir" tool added.
788    - improvements and cleanup to Tools structure
789    - documentation updates
790    
791    ----------------------------------------------------------------------
792    Name: Allen Leung
793    Date: Thu Dec 14 03:45:24 EST 2000
794    Description:
795    Tag:  leunga-20001214-int-inf
796    Description:
797    
798       In IntInf, added these standard functions, which are missing from our
799    implementation:
800    
801        andb : int * int -> int
802        xorb : int * int -> int
803        orb  : int * int -> int
804        notb : int -> int
805         <<   : int * word -> int
806        ~>>  : int * word -> int
807    
808       Not tested, I hope they are correct.
809    
810    ----------------------------------------------------------------------
811    Name: Allen Leung
812    Date: Fri Dec  8 19:23:26 EST 2000
813    Description:
814    Tag:  leunga-20001208-nowhere
815    Description:
816    
817      Slight improvements to the 'nowhere' tool to handle OR-patterns,
818    to generate better error messages etc.  Plus a brief manual.
819    
820    ----------------------------------------------------------------------
821    Name: Lal George
822    Date: 2000/12/08 09:54:02 EST 2000
823    Tag: Release_110_31
824    Description:
825    
826    - Version 110.31
827    ----------------------------------------------------------------------
828    Name: Allen Leung
829    Date: Thu Dec  7 22:01:04 EST 2000
830    Tag:  leunga-20001207-cell-monster-hack
831    Description:
832    
833    Major MLRISC internal changes.  Affect all clients.
834    Summary:
835    
836    1.  Type CELLS.cell = int is now replaced by a datatype.
837        As a result, the old regmap is now gone.  Almost all interfaces
838        in MLRISC change as a consequence.
839    
840    2.  A new brand version of machine description tool (v3.0) that generates
841        modules expecting the new interface.  The old version is removed.
842    
843    3.  The RA interface has been further abstracted into two new functors.
844        RISC_RA and X86RA.  These functors have much simpler interfaces.
845        [See also directory MLRISC/demo.]
846    
847    4.  Some other new source->source code generation tools are available:
848    
849        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
850        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
851           I use this tool to generate the peephole optimizers---with the new
852           cell type changes, peephole rules are becoming difficult to write
853           without conditional pattern matching.
854    
855    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
856        cover the entire MLRISC source tree so many things broke.
857    
858    6.  CM files have been moved to the subdirectory MLRISC/cm.
859        They are moved because there are a lot of them and they clutter up the
860        root dir.
861    
862    7.  More detailed documentation to come...
863    
864        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
865        a makeml -rebuild first.  This is because of other other
866        changes that Matthias has made (see below).
867    
868    
869    ----------------------------------------------------------------------
870    Name: Matthias Blume
871    Date: 2000/11/30 23:12:00 JST
872    Tag: blume-20001130-filereorg
873    Description:
874    
875    Some manual updates and some file reorganizations in CM.
876    
877    ----------------------------------------------------------------------
878    Name: Matthias Blume
879    Date: 2000/11/24 17:45:00 JST
880    Tag: blume-20001124-link
881    Description:
882    
883    Drastically improved link traversal code for the case that the dynamic
884    value was already loaded at bootstrap time.  As a result, CM and CMB
885    now both load blazingly fast -- even on a very slow machine.  Also,
886    memory consumption has been further reduced by this.
887    
888    Warning: The format of the PIDMAP file has changed.  THerefore, to
889    bootstrap you have to do this:
890    
891    1. Run CMB.make
892    2. Make a symbolic link for the boot directory:
893         ln -s sml.boot.ARCH-OS xxx
894    3. "Rebuild" the boot directory:
895         ./makeml -boot xxx -rebuild sml ; rm xxx
896    4. Boot normally:
897          ./makeml
898    
899    ----------------------------------------------------------------------
900    Name: Matthias Blume
901    Date: 2000/11/21 21:20:00 JST
902    Tag: blume-20001121-tools
903    Description:
904    
905    Continued hacking on autoloading problem -- with success this time.
906    Also changed tool-plugin mechanism.  See new CM manual.
907    
908    ----------------------------------------------------------------------
909    Name: Matthias Blume
910    Date: 2000/11/19 14:30:00 JST
911    Tag:  blume-20001119-autoload
912    Description:
913    
914    Some hacking to make autoloading faster.  Success for CMB, no success
915    so far for CM.  There is a reduced structure CM' that autoloads faster.
916    (This is a temporary, non-documented hack to be eliminated again when
917    the general problem is solved.)
918    
919    ----------------------------------------------------------------------
920    Name: Matthias Blume
921    Date: 2000/11/17 14:10:00 JST
922    Tag: blume-20001117-pickle-lib
923    Description:
924    
925    1. Eliminated comp-lib.cm
926    2. Made pickle-lib.cm
927    3. Eliminated all uses of intset.sml (from comp-lib.cm)
928    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
929       equivalent constructs from smlnj-lib.cm (INtHashTable).
930    5. Point 4. also goes for those uses of intmap.* in MLRISC.
931       Duplicated intmap modules thrown out.
932    6. Hunted down all duplicated SCC code and replaced it with
933       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
934    7. Rewrote Feedback module.
935    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
936       should be thrown out and equivalent modules from smlnj-lib.cm
937       should be used (IntRedBlackSet, IntListSet, ...).
938    
939    Confirmed that compiler compiles to fixpoint.
940    
941    ----------------------------------------------------------------------
942    Name: Allen Leung
943    Date: 2000/11/10 18:00:00
944    Tag: leunga-20001110-new-x86-fp
945    
946    A new x86 floating point code generator has been added.
947    By default this is turned off.  To turn this on, do:
948    
949        CM.autoload "$smlnj/compiler.cm";
950        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
951    
952    Changes:
953    
954    1.  Changed FTAN to FPTAN so that the assembly output is correct.
955    2.  Changed the extension callback for FTANGENT to generate:
956    
957              fptan
958              fstp  %st(0)
959        instead of
960              fptan
961              fstpl ftempmem
962    
963    3.  Numerous assembly fixes for x86.
964    
965    5.  Cleaned up the machine code output module x86/x86MC.sml and added
966        support for a whole bunch of instructions and addressing modes:
967    
968          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
969          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
970          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
971          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
972          fxch %st(n)
973          fld %st(n)
974          fst %st(n)
975          fst mem
976          fstp %st(n)
977          fucom %st(n)
978          fucomp %st(n)
979    
980        All these are now generated when the fast fp mode is turned on.
981    
982    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
983    
984    ----------------------------------------------------------------------
985    Name: Matthias Blume
986    Date: 2000/11/09 11:20:00 JST
987    Tag: blume-20001109-scc
988    Description:
989    
990    Eliminated some code duplication:
991    
992    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
993       (Otherwise the functor is useless.)
994    2. Used GraphSCCFn where SCCUtilFun was used previously.
995    3. Got rid of SCCUtilFun (in comp-lib.cm).
996    
997    ----------------------------------------------------------------------
998    Name: Lal George
999    Date: 2000/11/06 09:02:21 EST 2000
1000    Tag: Release_110_30
1001    Description:
1002    
1003    - Version 110.30
1004    ----------------------------------------------------------------------
1005    Name: Matthias Blume
1006    Date: 2000/11/04 14:45:00
1007    Tag: blume-20001104-mlbuild
1008    Description:
1009    
1010    - Made ml-build faster on startup.
1011    - Documentation fixes.
1012    
1013    ----------------------------------------------------------------------
1014    Name: Matthias Blume
1015    Date: 2000/11/02 17:00:00 JST
1016    Tag: blume-20001102-condcomp
1017    Description:
1018    
1019    - Small tweaks to pickler -- new BOOTFILES!
1020    - Version bumped to 110.29.2.
1021    - Added conditional compilation facility to init.cmi (see comment there).
1022    ----------------------------------------------------------------------
1023    Name: Allen Leung
1024    Date: 2000/10/23 19:31:00
1025    Tag: leunga-20001023-demo-ra
1026    
1027    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
1028    2. Test programs for the graph library updated
1029    3. Some new MLRISC demo programs added
1030    
1031    ----------------------------------------------------------------------
1032    Name: Matthias Blume
1033    Date: 2000/08/31 22:15:00 JST
1034    Tag: blume-20001017-errmsg
1035    Description:
1036    
1037    More error message grief: Where there used to be no messages, there
1038    now were some that had bogus error regions.  Fixed.
1039    
1040    ----------------------------------------------------------------------
1041    Name: Matthias Blume
1042    Date: 2000/08/31 17:30:00 JST
1043    Tag: blume-20001017-v110p29p1
1044    Description:
1045    
1046    I made a version 110.29.1 with new bootfiles.
1047    
1048    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
1049              CM documentation changes and a small bugfix in CM's error reporting.
1050    
1051    ----------------------------------------------------------------------
1052    Name: Lal George
1053    Date: 2000/09/27 14:42:35 EDT
1054    Tag: george-20000927-nodestatus
1055    Description:
1056    
1057    Changed the type of the nodestatus, so that:
1058    
1059            SPILLED(~1)             is now SPILLED
1060            SPILLED(m) where m>=0   is now MEMREG(m)
1061            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
1062    
1063    ----------------------------------------------------------------------
1064    Name: Matthias Blume
1065    Date: 2000/09/07 14:45:00 JST
1066    Tag: blume-20000907-cmerrmsg
1067    Description:
1068    
1069    Small tweak to CM to avoid getting ML syntax error messages twice.
1070    
1071    ----------------------------------------------------------------------
1072    Name: Matthias Blume
1073    Date: 2000/08/31 18:00:00 JST
1074    Tag: blume-20000831-cvsbootfiles
1075    Description:
1076    
1077    New URL for boot files (because the 110.29 files on the BL server do
1078    now work correctly with my updated install scripts for yacc and lex).
1079    
1080    ----------------------------------------------------------------------
1081    Name: Matthias Blume
1082    Date: 2000/08/08 12:33:00 JST
1083    Tag: blume-20000808-manual
1084    Description:
1085    
1086    Tiny update to CM manual.
1087    
1088    ----------------------------------------------------------------------
1089    Name: Allen Leung
1090    Date: 2000/08/7 19:31:00
1091    Tag: leunga-20000807-a-whole-bunch-of-stuff
1092    
1093      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
1094    and Mobdy.
1095    
1096    1.  x86
1097    
1098       a.  Fixes to peephole module by John and Dan.
1099       b.  Assembly fix to SETcc by Allen.
1100       c.  Fix to c-call by John.
1101       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
1102       e.  Instruction selection optimization to SETcc as suggested by John.
1103    
1104           For example,
1105    
1106            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
1107    
1108           should generate:
1109    
1110            MOVL a, x
1111            SUBL b, x
1112            SHRL 31, x
1113    
1114    2.  IR stuff
1115    
1116         A bunch of new DJ-graph related algorithms added.  These
1117         speed up SSA construction.
1118    
1119    3.  SSA + Scheduling
1120    
1121         Added code for SSA and scheduling to the repository
1122    
1123    ----------------------------------------------------------------------
1124    Name: Lal George
1125    Date: 2000/07/27 11:53:14 EDT
1126    
1127    Tag: lal-20000727-linux-ppc
1128    Description:
1129    
1130     Made changes to support Linux PPC.
1131     p.s. I have confirmation that the 110.29 boot files work fine.
1132    
1133    ----------------------------------------------------------------------
1134    Name: Matthias Blume
1135    Date: 2000/07/27 17:40:00 JST
1136    Tag: blume-20000727-scripts
1137    Description:
1138    
1139    !!!! WARNING !!!!
1140    You must recompile the runtime system!
1141    !!!! WARNING !!!!
1142    
1143    This is basically another round of script-enhancements:
1144    
1145    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
1146       and undefine CM preprocessor symbols.
1147    
1148    2. ml-build avoids generating a new heap image if it finds that the
1149       existing one is still ok.  (The condition is that no ML file had to
1150       be recompiled and all ML files are found to be older that the heap
1151       file.)
1152    
1153       To make this work smoothly, I also hacked the runtime system as
1154       well as SMLofNJ.SysInfo to get access to the heap image suffix
1155       (.sparc-solaris, ...) that is currently being used.
1156    
1157       Moreover, the signature of CM.mk_standalone has changed.  See the
1158       CM manual.
1159    
1160    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
1161       CM manual for details.)
1162    
1163    4. More CM manual updates:
1164        - all of the above has been documented.
1165        - there is now a section describing the (CM-related) command line
1166          arguments that are accepted by the "sml" command
1167    
1168    ----------------------------------------------------------------------
1169    Name: Matthias Blume
1170    Date: 2000/07/25 16:20:00 JST
1171    Tag: blume-20000725-makedepend
1172    Description:
1173    
1174    Added a script called ml-makedepend.  This can be used in makefiles
1175    for Unix' make in a way very similar to the "makedepend" command for
1176    C.
1177    
1178    The script internally uses function CM.sources.
1179    
1180    Synopsis:
1181    
1182        ml-makedepend [-f makefile] cmfile targetname
1183    
1184    The default for the makefile is "makefile" (or "Makefile" should
1185    "makefile" not exist).
1186    
1187    ml-makedepend adds a cmfile/targetname-specific section to this
1188    makefile (after removing the previous version of this section).  The
1189    section contains a single dependency specification with targetname on
1190    the LHS (targetname is an arbitrary name), and a list of files derived
1191    from the cmfile on the RHS.  Some of the files on the RHS are
1192    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
1193    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
1194    path names.  The makefile writer is responsible for making sure that
1195    these variables have correct at the time "make" is invoked.
1196    
1197    ----------------------------------------------------------------------
1198    Name: Matthias Blume
1199    Date: 2000/07/22 23:30:00 JST
1200    Tag: blume-20000722-urlupdate
1201    Description:
1202    
1203    Changed BOOT and config/srcarchiveurl to point to BL server:
1204    
1205        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
1206    
1207    ----------------------------------------------------------------------
1208    Name: Matthias Blume
1209    Date: 2000/07/18 18:00:00 JST
1210    Tag: blume-20000718-Version_110_29
1211    Description:
1212    
1213    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
1214    
1215    2. Updated config/version to 110.29
1216    
1217    3. Updated config/srcarchiveurl
1218    
1219    3. New boot files!
1220       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
1221    
1222    ----------------------------------------------------------------------
1223    Name: Matthias Blume
1224    Date: 2000/07/11 13:58:00 JST
1225    Tag: blume-20000711-doctypo
1226    Description:
1227    
1228    Fixed a few typos in CM manual.
1229    
1230    ----------------------------------------------------------------------
1231    Name: Allen Leung
1232    Date: 2000/06/15 00:38:00
1233    Tag: leunga-20000704-sparc-x86
1234    
1235    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
1236    2. fix to x86 RET bug [found by Dan Grossman]
1237    3. sparc assembly bug fix for ticc instructions [found by Fermin]
1238    
1239       Affects c-- and moby only
1240    
1241    ----------------------------------------------------------------------
1242    Name: Matthias Blume
1243    Date: 2000/07/04 15:26:00
1244    Tag: blume-20000704-trigger
1245    Description:
1246    
1247    1. Improvements to CM manual.
1248    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
1249       of getting a back-trace.  The function, when called, raises an
1250       internal exception which explicitly carries the full back-trace history,
1251       so it is unaffected by any intervening handle-raise pairs ("trivial"
1252       or not).  The interactive loop will print that history once it arrives
1253       at top level.
1254       Short of having all exceptions implicitly carry the full history, the
1255       recommended way of using this facility is:
1256         - compile your program with instrumentation "on"
1257         - run it, when it raises an exception, look at the history
1258         - if the history is "cut off" because of some handler, go and modify
1259           your program so that it explicitly calls BTrace.trigger
1260         - recompile (still instrumented), and rerun; look at the full history
1261    
1262    ----------------------------------------------------------------------
1263    Name: Matthias Blume
1264    Date: 2000/07/03 15:36:00 JST
1265    Tag: blume-20000702-manual
1266    Description:
1267    
1268    Small corrections and updates to CM manual.
1269    
1270    ----------------------------------------------------------------------
1271    Name: Matthias Blume
1272    Date: 2000/06/29 16:04:00 JST
1273    Tag: blume-20000629-yacctool
1274    Description:
1275    
1276    Changes:
1277    
1278    1. Class "mlyacc" now takes separate arguments to pass options to
1279       generated .sml- and .sig-files independently.
1280    2. Corresponding CM manual updates.
1281    3. BTrace module now also reports call sites.  (However, for loop clusters
1282       it only shows from where the cluster was entered.)  There are associated
1283       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1284    
1285    ----------------------------------------------------------------------
1286    Name: Matthias Blume
1287    Date: 2000/06/27 16:51:00 JST
1288    Tag: blume-20000627-noweb
1289    Description:
1290    
1291    Changes:
1292    
1293     1. Implemented "subdir" and "witness" options for noweb tool.
1294        This caused some slight internal changes in CM's tool implementation.
1295     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
1296        some remaining issues from earlier path anchor changes.
1297     3. Updated CM manual accordingly.
1298    
1299     4. Changed implementation of back-tracing so that I now consider it
1300        ready for prime-time.
1301    
1302        In particular, you don't have to explicitly trigger the back-trace
1303        anymore.  Instead, if you are running BTrace-instrumented code and
1304        there is an uncaught exception (regardless of whether or not it was
1305        raised in instrumented code), the top-level evalloop will print
1306        the back-trace.
1307    
1308        Features:
1309    
1310          - Instrumented and uninstrumented code work together seemlessly.
1311            (Of course, uninstrumented code is never mentioned in actual
1312             back-traces.)
1313    
1314          - Asymptotic time- and space-complexity of instrumented code is
1315            equal to that of uninstrumented code.  (This means that
1316            tail-recursion is preserved by the instrumentation phase.)
1317    
1318          - Modules whose code has been instrumented in different sessions
1319            work together without problem.
1320    
1321          - There is no penalty whatsoever on uninstrumented code.
1322    
1323          - There is no penalty on "raise" expressions, even in
1324            instrumented code.
1325    
1326        A potential bug (or perhaps it is a feature, too):
1327    
1328          A back-trace reaches no further than the outermost instrumented
1329          non-trivial "raise".  Here, a "trivial" raise is one that is the
1330          sole RHS of a "handle" rule.  Thus, back-traces reach trough
1331    
1332               <exp> handle e => raise e
1333    
1334          and even
1335    
1336               <exp> handle Foo => raise Bar
1337    
1338          and, of course, through
1339    
1340               <exp> handle Foo => ...
1341    
1342         if the exception was not Foo.
1343    
1344         Back-traces always reach right through any un-instrumented code
1345         including any of its "handle" expressions, trivial or not.
1346    
1347       To try this out, do the following:
1348    
1349         - Erase all existing binfiles for your program.
1350           (You may keep binfiles for those modules where you think you
1351            definitely don't need back-tracing.)
1352         - Turn on back-trace instrumentation:
1353              SMLofNJ.Internals.BTrace.mode (SOME true);
1354         - Recompile your program.  (I.e., run "CM.make" or "use".)
1355         - You may now turn instrumentation off again (if you want):
1356              SMLofNJ.Internals.BTrace.mode (SOME false);
1357         - Run your program as usual.  If it raises an exception that
1358           reaches the interactive toplevel, then a back-trace will
1359           automatically be printed.  After that, the toplevel loop
1360           will print the exception history as usual.
1361    
1362    ----------------------------------------------------------------------
1363    Name: Matthias Blume
1364    Date: 2000/06/26 09:56:46 JST
1365    Tag: blume-20000626-setup
1366    Description:
1367    
1368    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1369          ML code before and after compiling a file (e.g., to set compiler
1370          flags)
1371    
1372    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1373              - associated changes to btrace.sml (BTrace instrumentation pass)
1374              - cleaner implementation of btimp.sml (BTrace tracing and report
1375                module)
1376    
1377    CM manual: * new path encoding documented
1378               * description of setup-parameter to "sml" added
1379    
1380    The biggest user-visible change to back-tracing is that it is no
1381    longer necessary to compile all traced modules within the same
1382    session.  (This was a real limitation.)
1383    
1384    ----------------------------------------------------------------------
1385    Name: Matthias Blume
1386    Date: 2000/06/24 12:40:00 JST
1387    Tag: blume-20000624-startup
1388    Description:
1389    
1390    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1391    bit too often -- to put it mildly. :)
1392    
1393    ----------------------------------------------------------------------
1394    Name: Matthias Blume
1395    Date: 2000/06/23 18:20:00 JST
1396    Tag: blume-20000623-btrace
1397    Description:
1398    
1399    This updates adds a backtrace facility to aid programmers in debugging
1400    their programs.  This involves the following changes:
1401    
1402    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1403       keeping track of the current call stack.  When programs are compiled
1404       in a special mode, the compiler will insert calls to these hooks
1405       into the user program.
1406       "Hook" means that it is possible for different implementations of
1407       back-tracing to register themselves (at different times).
1408    
1409    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1410       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1411       It can be turned on using this call:
1412         SMLofNJ.Internals.BTrace.mode (SOME true);
1413       Turning it off again:
1414         SMLofNJ.Internals.BTrace.mode (SOME false);
1415       Querying the current status:
1416         SMLofNJ.Internals.BTrace.mode NONE;
1417       Annotated programs are about twice as big as normal ones, and they
1418       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1419       where all hooks do nothing).  The slowdown with a plugin that is
1420       actually useful (such as the one supplied by default) is even greater,
1421       but in the case of the default plugin it is still only an constant
1422       factor (amortized).
1423    
1424    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1425       with a sub-structure BTrace for controlling back-tracing.  In particular,
1426       the above-mentioned function "mode" controls whether the annotation
1427       phase is invoked by the compiler.  Another important function is
1428       "trigger": when called it aborts the current execution and causes
1429       the top-level loop to print a full back-trace.
1430    
1431    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1432       for back-tracing.  It keeps track of the dynamic call stack and in
1433       addition to that it keeps a partial history at each "level" of that
1434       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1435       d and b (at separate times, dynamically), then the report will show:
1436    
1437       GOTO   d
1438             /c
1439       GOTO  \b
1440       CALL   a
1441    
1442       This shows that there was an initial non-tail call of a, then a
1443       tail-call to b or c, looping behavior in a cluster of functions that
1444       consist of b and c, and then a goto from that cluster (i.e., either from
1445       b or from c) to d.
1446    
1447       Note that (depending on the user program) the amount of information
1448       that the back-trace module has to keep track of at each level is bounded
1449       by a constant.  Thus, the whole implementation has the same asymptotical
1450       complexity as the original program (both in space and in time).
1451    
1452    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1453       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1454       which is raised by the "trigger" function mentioned above.
1455    
1456    Notes on usage:
1457    
1458    - Annotated code works well together with unannotated code:
1459    Unannotated calls simply do not show up at all in the backtrace.
1460    
1461    - It is not a good idea to let modules that were annotated during
1462    different sessions run at the same time.  This is because the compiler
1463    chooses small integers to identify individual functions, and there
1464    will be clashes if different modules were compiled in separate sessions.
1465    (Nothing will crash, and you will even be told about the clashes, but
1466    back-trace information will in general not be useful.)
1467    
1468    - Back-tracing can be confused by callcc and capture.
1469    
1470    - The only way of getting a back-trace right now is to explicitly
1471    invoke the "trigger" function from your user program.  Eventually, we
1472    should make every exception carry back-trace information (if
1473    available).  But since this creates more overhead at "raise"-time
1474    (similar to the current exnHistory overhead), I have not yet
1475    implemented this.  (The implementation will be rather easy.)  With
1476    exceptions carrying back-trace information, this facility will be even
1477    more useful because users don't need to modify their programs...
1478    
1479    - While it is possible to compile the compiler with back-trace
1480    annotations turned on (I did it to get some confidence in
1481    correctness), you must make absolutely sure that core.sml and
1482    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1483    be compiled with annotation because there is no core access yet, but
1484    if you compile btimp.sml with annotation, then the system will go into
1485    an infinite recursion and crash.)
1486    Since CM currently does not know about BTrace, the only way to turn
1487    annotations on and off for different modules of the compiler is to
1488    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1489    this is awkward and clumsy.
1490    
1491    Sample sessions:
1492    
1493    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1494    - SMLofNJ.Internals.BTrace.mode (SOME true);
1495    [autoloading]
1496    [autoloading done]
1497    val it = false : bool
1498    - structure X = struct
1499    -     fun main n = let
1500    -         fun a (x, 0) = d x
1501    -           | a (x, n) = b (x, n - 1)
1502    -         and b (x, n) = c (x, n)
1503    -         and c (x, n) = a (x, n)
1504    -         and d x = e (x, 3)
1505    -         and e (x, 0) = f x
1506    -           | e (x, n) = e (x, n - 1)
1507    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1508    -           | f n = n * g (n - 1)
1509    -         and g n = a (n, 3)
1510    -     in
1511    -         f n
1512    -     end
1513    - end;
1514    structure X : sig val main : int -> int end
1515    - X.main 3;
1516    *** BACK-TRACE ***
1517    GOTO   stdIn:4.2-13.20: X.main[2].f
1518    GOTO-( stdIn:4.2-13.20: X.main[2].e
1519    GOTO   stdIn:4.2-13.20: X.main[2].d
1520         / stdIn:4.2-13.20: X.main[2].a
1521         | stdIn:4.2-13.20: X.main[2].b
1522    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1523    CALL   stdIn:4.2-13.20: X.main[2].g
1524    GOTO   stdIn:4.2-13.20: X.main[2].f
1525    GOTO-( stdIn:4.2-13.20: X.main[2].e
1526    GOTO   stdIn:4.2-13.20: X.main[2].d
1527         / stdIn:4.2-13.20: X.main[2].a
1528         | stdIn:4.2-13.20: X.main[2].b
1529    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1530    CALL   stdIn:4.2-13.20: X.main[2].g
1531    GOTO   stdIn:4.2-13.20: X.main[2].f
1532    GOTO-( stdIn:4.2-13.20: X.main[2].e
1533    GOTO   stdIn:4.2-13.20: X.main[2].d
1534         / stdIn:4.2-13.20: X.main[2].a
1535         | stdIn:4.2-13.20: X.main[2].b
1536    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1537    CALL   stdIn:4.2-13.20: X.main[2].g
1538    GOTO   stdIn:4.2-13.20: X.main[2].f
1539    CALL   stdIn:2.15-17.4: X.main[2]
1540    -
1541    
1542    (Note that because of a FLINt bug the above code currently does not
1543    compile without BTrace turned on.)
1544    
1545    Here is another example, using my modified Tiger compiler:
1546    
1547    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1548    - SMLofNJ.Internals.BTrace.mode (SOME true);
1549    [autoloading]
1550    [autoloading done]
1551    val it = false : bool
1552    - CM.make "sources.cm";
1553    [autoloading]
1554    ...
1555    [autoloading done]
1556    [scanning sources.cm]
1557    [parsing (sources.cm):parse.sml]
1558    [creating directory CM/SKEL ...]
1559    [parsing (sources.cm):tiger.lex.sml]
1560    ...
1561    [wrote CM/sparc-unix/semant.sml]
1562    [compiling (sources.cm):main.sml]
1563    [wrote CM/sparc-unix/main.sml]
1564    [New bindings added.]
1565    val it = true : bool
1566    - Main.compile ("../testcases/merge.tig", "foo.out");
1567    *** BACK-TRACE ***
1568    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
1569    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1570    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1571    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1572    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1573    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1574    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
1575         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
1576    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
1577    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1578    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
1579    CALL   main.sml:1.18-118.4: Main.compile[2]
1580    -
1581    
1582    ----------------------------------------------------------------------
1583    Name: Matthias Blumen
1584    Date: 2000/06/21 18:00:00 JST
1585    Tag: blume-20000621-manual
1586    Description:
1587    
1588    CM manual update: Path environments documented.
1589    
1590    ----------------------------------------------------------------------
1591    Name: Matthias Blume
1592    Date: 2000/06/19 13:40:00
1593    Tag: blume-20000619-manual
1594    Description:
1595    
1596    CM manual and system/README update.  This only covers the fact that
1597    there are no more implicit anchors.  (Path environments and the "bind"
1598    option to "cm" have yet to be documented.)
1599    
1600    ----------------------------------------------------------------------
1601    Name: Matthias Blume
1602    Date: 2000/06/19 11:05:00 JST
1603    Tag: blume-20000619-chdir-bugfix
1604    Description:
1605    
1606    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1607    
1608    ----------------------------------------------------------------------
1609    Name: Matthias Blume
1610    Date: 2000/06/18 22:00:10 JST
1611    Tag: blume-20000618-implicit-anchors-really-gone
1612    Description:
1613    
1614    I updates the previous HISTORY entry where I forgot to mention that
1615    implicit anchors are no longer with us.
1616    
1617    The current update also gets rid of the (now useless) controller
1618    CM.Control.implicit_anchors.
1619    
1620    ----------------------------------------------------------------------
1621    Name: Matthias Blume
1622    Date: 2000/06/16 17:30:00 JST
1623    Tag: blume-20000616-anchorenv
1624    Description:
1625    
1626    This patch implements the long anticipated (just kidding :) "anchor
1627    environment" mechanism.  In the course of doing this, I also
1628    re-implemented CM's internal "SrcPath" module from scratch.  The new
1629    one should be more robust in certain boundary cases.  In any case, it
1630    is a lot cleaner than its predecessor (IMHO).
1631    
1632    This time, although there is yet another boot file format change, I
1633    kept the unpickler backward-compatible.  As a result, no new bootfiles
1634    are necessary and bootstrapping is straightforward.  (You cannot read
1635    new bootfiles into an old system, but the other way around is no
1636    problem.)
1637    
1638    Visible changes:
1639    
1640    ** 0. Implicit path anchors (without the leading $-symbol) are no
1641    longer recognized at all. This means that such path names are not
1642    illegal either.  For example, the name basis.cm simply refers to a
1643    local file called "basis.cm" (i.e, the name is an ordinary path
1644    relative to .cm-files directory).  Or, to put it differently, only
1645    names that start with $ are anchored paths.
1646    
1647    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1648    vanished.
1649    
1650    John (Reppy) had critizised this as soon as I originally proposed and
1651    implemented it, but at that time I did not really deeply believe
1652    him. :) Now I came full-circle because I need the $<singlearc> syntax
1653    in another place where it cannot be seen as an abbreviation for
1654    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1655    seems to mean (i.e., it "expands" into the corresponding anchor
1656    value).
1657    
1658    However, when paths are used as members in CM description files, it
1659    continues to be true that there must be at least another arc after the
1660    anchor.  This is now enforced separately during semantic analysis
1661    (i.e., from a lexical/syntactical point of view, the notation is ok.)
1662    
1663    ** 2. The "cm" class now accepts an option "bind".  The option's value
1664    is a sub-option list of precisely two items -- one labeled "anchor"
1665    and the other one labeled "value".  As you might expect, "anchor" is
1666    used to specify an anchor name to be bound, and "value" specifies what
1667    the anchor is being bound to.
1668    
1669    The value must be a directory name and can be given in either standard
1670    syntax (including the possibility that it is itself an anchored path)
1671    or native syntax.
1672    
1673    Examples:
1674    
1675       foo.cm (bind:(anchor:bar value:$mystuff/bar))
1676       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1677    
1678    and so on.
1679    
1680    The meaning of this is that the .cm-file will be processed with an
1681    augmented anchor environment where the given anchor(s) is/are bound to
1682    the given values(s).
1683    
1684    The rationale for having this feature is this: Suppose you are trying
1685    to use two different (already stable) libraries a.cm and b.cm (that
1686    you perhaps didn't write yourself).  Further, suppose each of these
1687    two libraries internally uses its own auxiliary library $aux/lib.cm.
1688    Normally you would now have a problem because the anchor "lib" can not
1689    be bound to more than one value globally.  Therefore, the project that
1690    uses both a.cm and b.cm must locally redirect the anchor to some other
1691    place:
1692    
1693       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1694       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1695    
1696    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1697    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
1698    
1699    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1700    you when you do so at the time of CM.stabilize).  Therefore, you can
1701    also use an anchored path as the value:
1702    
1703      a.cm (bind:(anchor:lib value:$a-lib))
1704      b.cm (bind:(anchor:lib value:$b-lib))
1705    
1706    Now you can globally configure (using the usual CM.Anchor.anchor or
1707    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1708    itself is always locally bound, setting it globally is no longer
1709    meaningful or necessary (but it does not hurt either).  In fact, "lib"
1710    can still be used as a global anchor for separate purposes.  As a
1711    matter of fact, one can locally define "lib" in terms of a global
1712    "lib":
1713    
1714      a.cm (bind:(anchor:lib value:$lib/a))
1715      b.cm (bind:(anchor:lib value:$lib/b))
1716    
1717    ** 3: The encoding of path names has changed.  This affects the way
1718    path names are shown in CM's progress report and also the internal
1719    protocol encoding used for parallel make.
1720    
1721    The encoding now uses one or more ':'-separated segments.  Each
1722    segments corresponds to a file that has been specified relative to the
1723    file given by its preceding segment.  The first segment is either
1724    relative to the CWD, absolute, or anchored.  Each segment itself is
1725    basically a Unix pathname; all segments but the first are relative.
1726    
1727    Example:
1728    
1729       $foo/bar/baz.cm:a/b/c.sml
1730    
1731    This path denotes the file bar/a/b/c.sml relative to the directory
1732    denoted by anchor "foo".  Notice that the encoding also includes
1733    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1734    paths are resolved relative to the .cm-files directory, so baz.cm must
1735    be ignored to get the "real" pathname.
1736    
1737    To make this fact more obvious, CM puts the names of such "virtual
1738    arcs" into parentheses when they appear in progress reports. (No
1739    parentheses will appear in the internal protocol encoding.)  Thus,
1740    what you really see is:
1741    
1742      $foo/bar/(baz.cm):a/b/c.sml
1743    
1744    I find this notation to be much more informative than before.
1745    
1746    Another new feature of the encoding is that special characters
1747    including parentheses, colons, (back)slashes, and white space are
1748    written as \ddd (where ddd is the decimal encoding of the character).
1749    
1750    *** The CM manual still needs to be updated.
1751    
1752    ----------------------------------------------------------------------
1753    Name: Allen Leung
1754    Date: 2000/06/15 00:38:00
1755    Tag: leunga-20000615-x86-peephole
1756    
1757    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1758    
1759    ----------------------------------------------------------------------
1760    Name: Matthias Blume
1761    Date: 2000/06/12 11:40:00
1762    Tag: blume-20000612-parmakefix
1763    Description:
1764    
1765    More cleanup after changing the file naming scheme: This time I
1766    repaired the parallel make mechanism for CMB.make which I broke earlier.
1767    
1768    ----------------------------------------------------------------------
1769    Name: Allen Leung
1770    Date: 2000/06/09 01:25:00
1771    Tag: leunga-20000609-various
1772    
1773    None of these things should affect normal SML/NJ operations
1774    
1775    1. Peephole improvements provided by Fermin (c--)
1776    2. New annotation DEFUSE for adding extra dependence (moby)
1777    3. New X86 LOCK instructions (moby)
1778    4. New machine description language for reservation tables (scheduling)
1779    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1780       trees etc.)
1781    6. I've changed the CM files so that they can work with versions
1782       110.0.6, 110.25 and 110.28
1783    
1784    ----------------------------------------------------------------------
1785    Name: Matthias Blume
1786    Date: 2000/06/09 12:40:00
1787    Tag: blume-20000609-log
1788    Description:
1789    
1790    - Removed all(?) remaining RCS Log entries from sources.
1791    
1792    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1793      anchored paths).
1794    
1795    ----------------------------------------------------------------------
1796    Name: Matthias Blume
1797    Date: 2000/06/07 17:00:00 JST
1798    Tag: blume-20000607-no-implicit-anchors
1799    Description:
1800    
1801    1. This update changes the default setting for
1802    CM.Control.implicit_anchors from true to false.  This means that
1803    implicit anchors are no longer permitted by default.  I also tried to
1804    make sure that nothing else still relies on implicit anchors.
1805    (This is the next step on the schedule towards a CM that does not even
1806    have the notion of implicit anchors anymore.)
1807    
1808    2. More CM manual updates.
1809    
1810    3. I managed to track down and fix the pickling bug I mentioned last
1811    time.  Because of the previously existing workaround, this entails no
1812    immediate practical changes.
1813    
1814    ----------------------------------------------------------------------
1815    Name: Matthias Blume
1816    Date: 2000/06/06 11:15:00 JST
1817    Tag: blume-20000606-lazierpickle
1818    Description:
1819    
1820    !!!! NEW BOOT FILES !!!!
1821    
1822    * The main purpose of this update is to make library pickles lazier in
1823    order to reduce the initial space penalty for autoloading a library.
1824    As a result, it is now possible to have $smlnj/compiler.cm
1825    pre-registered.  This should take care of the many complaints or
1826    inquiries about missing structure Compiler.  This required changes to
1827    CM's internal data structures and small tweaks to some algorithms.
1828    
1829    As a neat additional effect, it is no longer necessary (for the sake
1830    of lean heap image files) to distinguish between a "minimal" CM and a
1831    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1832    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1833    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1834    
1835    To make the life of compiler-hackers easier, "makeml" now also
1836    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1837    words, after you bootstrap a new sml for the first time, you will not
1838    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1839    around you will still have to do it, though.)
1840    
1841    * A second change consists of major updates to the CM manual.  There
1842    are now several appendices with summary information and also a full
1843    specification of the CM description file syntax.
1844    
1845    * In directory src/system I added the script "allcross".  This script
1846    invokes sml and cross-compiles the compiler for all supported
1847    architectures.  (Useful when providing a new set of boot files.)
1848    
1849    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1850    added a small tweak to pickle-util.sml to work around this problem,
1851    but it is not a proper fix yet.  I will investigate further.  (The
1852    effect of the bug was an inflation of library pickle size.)
1853    
1854    * Version number increased to 110.28.1 (to avoid compatibility problems).
1855    
1856    ----------------------------------------------------------------------
1857    Name: Allen Leung
1858    Date: 2000/05/25 17:28 EDT
1859    Tag: leunga-20000525-ra
1860    Description:
1861    
1862      Fixed a bug in freezing phase of the register allocator.
1863    
1864    ----------------------------------------------------------------------
1865    Name: Allen Leung
1866    Date: 2000/05/15 22:53 EDT
1867    Tag: leunga-20000515-alpha-x86-ra
1868    Description:
1869    
1870      1. Alpha
1871    
1872          Slight cleanup.  Removed the instruction SGNXL
1873    
1874      2. X86
1875    
1876          Added the following instructions to the instruction set:
1877    
1878            ROLx, RORx,
1879            BTx, BTSx, BTLx, BTRx,
1880            XCHGx, and variants with the LOCK prefix
1881    
1882      3. Register Allocation
1883    
1884          The module ra-rewrite-with-renaming has been improved.
1885    
1886      These have no effect on SML/NJ.
1887    
1888    ----------------------------------------------------------------------
1889    Name: Matthias Blume
1890    Date: 2000/05/15 16:20:00 JST
1891    Tag: blume-20000515-lightrebuild
1892    Description:
1893    
1894    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1895       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1896       defined.  In effect, the command will not build any cross-compiler
1897       backends and therefore finish more quickly.
1898    
1899       The "fixpt" script also takes a "-light" switch to be able to use
1900       this new facility while compiling for a fixpoint.
1901    
1902    2. I replaced all mentions of anchored paths in group owner specifications
1903       with simple relative paths (usually starting with "..").
1904       The rationale is that a library's internal workings should not be
1905       compromised by the lack of some anchor.  (An anchor is necessary
1906       for someone who wants to refer to the library by an anchored path,
1907       but it should not be necessary to build the same library in the first
1908       place.)
1909    
1910    3. I changed the way CM's tool mechanism determines the shell command
1911       string used for things like ml-yacc etc. so that it does not break
1912       when CM.Control.implicit_anchors is turned off.
1913    
1914    ----------------------------------------------------------------------
1915    Name: Matthias Blume
1916    Date: 2000/05/12 18:20:00 JST
1917    Tag: blume-20000512-ml-build
1918    Description:
1919    
1920    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1921    getting installed properly (by config/install.sh).
1922    
1923    ----------------------------------------------------------------------
1924    Name: Matthias Blume
1925    Date: 2000/05/12 17:30:00 JST
1926    Tag: blume-20000512-anchors
1927    Description:
1928    
1929    !!! NEW BOOT FILES !!!
1930    
1931    This change is in preparation of fading out support for "implicitly
1932    anchored path names".  I went through all sources and used the
1933    explicit (and relatively new) $-notation.  See system/README and the
1934    CM manual for more info on this.
1935    
1936    I also modified the anchoring scheme for some things such as "smlnj",
1937    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1938    anchors are more expressive: anchor name and first arc do not have to
1939    coincide.  This entails the following user-visible change:
1940    
1941    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1942    particular, when you fire up sml with a command-line argument, say,
1943    e.g.:
1944    
1945       sml '$smlnj/cmb.cm'
1946    
1947    At the ML toplevel prompt:
1948    
1949       CM.autoload "$smlnj/cmb.cm";
1950    
1951    There is also a new controller in CM.Control that can be used to turn
1952    off all remaining support for implicit anchors by saying:
1953    
1954        CM.autoload "$smlnj/
1955        #set CM.Control.implicit_anchors false;
1956    
1957    This causes CM to reject implicitly anchored paths.  This is (for the
1958    time being) less permissive than the "final" version where there will
1959    be no more such implicit anchors and relative paths will be just that:
1960    relative.
1961    
1962    The next step (version after next version?) will be to make the
1963    default for CM.Control.implicit_anchors false.  After the dust has
1964    settled, I can then produce the "final" version of this...
1965    
1966    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1967    
1968    ----------------------------------------------------------------------
1969    Name: Matthias Blume
1970    Date: 2000/05/11 16:30:00 JST
1971    Tag: blume-20000511-sources
1972    Description:
1973    
1974    The main change is that I added function CM.sources as a generalized
1975    version of the earlier CM.makedepend.  This entails the following
1976    additional changes:
1977    
1978      - CM.makedepend has been dropped.
1979    
1980      - CM manual has been updated.
1981    
1982      - TOOLS signature and API have been changed.
1983    
1984    ----------------------------------------------------------------------
1985    Name: Allen Leung
1986    Date: 2000/05/10 21:17 EDT
1987    Tag: leunga-20000510-moby-c--ssa
1988    Description:
1989    
1990      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1991    None of these affect SML/NJ.
1992    
1993    1. Register Allocation
1994    
1995        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1996           This module tries to remove local (i.e. basic block level) redundancies
1997           during spilling.
1998    
1999        b. A new framework for performing region based register allocation.
2000           Not yet entirely functional.
2001    
2002    2. X86
2003    
2004       a. DefUse for POP was missing the stack pointer [found by Lal]
2005       b. Reload for CALL was incorrect in X86Spill [found by John]
2006       c. Various fixes in X86Spill so that it can be used correctly for
2007          the new spilling module.
2008    
2009    3. SSA/IR
2010    
2011       a. New module ir/dj-dataflow.sml implements elimination based
2012          data flow analysis.
2013    
2014    4. MLRiscGen
2015    
2016       a. Fix for gc type annotation
2017    
2018    5. MDGen
2019    
2020       Various fixes for machine description -> ml code translation.  For ssa
2021       only.
2022    
2023  ----------------------------------------------------------------------  ----------------------------------------------------------------------
2024  Name: Allen Leung  Name: Allen Leung
2025  Date: 2000/05/08 22:17 EDT  Date: 2000/05/08 22:17 EDT
# Line 335  Line 2344 
2344    
2345  3.  Assembly  3.  Assembly
2346    
2347      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
2348      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
2349      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
2350    
# Line 358  Line 2367 
2367    
2368  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
2369  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
2370  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
2371  legal surface-syntax symbol).  legal surface-syntax symbol).
2372    
2373  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 368  Line 2377 
2377  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
2378    
2379  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2380  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
2381  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
2382  more details.  more details.
2383    
# Line 483  Line 2492 
2492     (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
2493     used).     used).
2494    
2495     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
2496     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
2497     Makefiles or Shell scripts.     Makefiles or Shell scripts.
2498    
2499     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 496  Line 2505 
2505    
2506     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
2507    
2508     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
2509     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
2510    
2511  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 630  Line 2639 
2639    
2640       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
2641       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
2642       testl is perferable to andl.       testl is preferable to andl.
2643    
2644  7. RA (x86 only)  7. RA (x86 only)
2645    
# Line 812  Line 2821 
2821    
2822  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2823  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2824  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
2825  explanation.  explanation.
2826    
2827  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2828  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2829  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2830  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 859  Line 2868 
2868  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
2869  <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,
2870  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
2871  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2872  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2873  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2874    
# Line 909  Line 2918 
2918    it from that remote directory.    it from that remote directory.
2919    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2920    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2921    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2922    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2923    
2924    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.640  
changed lines
  Added in v.854

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