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 589, Fri Mar 31 16:14:36 2000 UTC revision 853, Tue Jun 26 20:13:50 2001 UTC
# Line 13  Line 13 
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Matthias Blume
17    Date: 2001/06/26 16:15:00 EDT
18    Tag: blume-20010626-cmdoc
19    Description:
20    
21    Minor typo fixed in CM manual (syntax diagram for libraries).
22    
23    ----------------------------------------------------------------------
24    Name: Matthias Blume
25    Date: 2001/06/25 22:55:00 EDT
26    Tag: blume-20010625-x86pc
27    Description:
28    
29    Fixed a nasty bug in the X86 assembly code that caused signal
30    handlers to fail (crash) randomly.
31    
32    ----------------------------------------------------------------------
33    Name: Matthias Blume
34    Date: 2001/06/25 12:05:00 EDT
35    Tag: blume-20010625-nlffigen
36    Description:
37    
38    This update fixes a number of minor bugs in ml-nlffigen as reported by
39    Nick Carter <nbc@andrew.cmu.edu>.
40    
41      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
42      2. Default names for generated files are now derived from the name of
43         the C file *without its directory*.  In particular, this causes generated
44         files to be placed locally even if the C file is in some system directory.
45      3. Default names for generated signatures and structures are also derived
46         from the C file name without its directory.  This avoids silly things
47         like "structure GL/GL".
48         (Other silly names are still possible because ml-nlffigen does not do
49          a thorough check of whether generated names are legal ML identifiers.
50          When in doubt, use command line arguments to force particular names.)
51    
52    ----------------------------------------------------------------------
53    Name: Matthias Blume
54    Date: 2001/06/21 12:25:00 EDT
55    Tag: blume-20010621-eXene
56    Description:
57    
58    eXene now compiles and (sort of) works again.
59    
60    The library name (for version > 110.33) is $/eXene.cm.
61    
62    I also added an new example in src/eXene/examples/nbody.  See the
63    README file there for details.
64    
65    ----------------------------------------------------------------------
66    Name: Matthias Blume
67    Date: 2001/06/20 16:40:00 EDT
68    Tag: blume-20010620-cml
69    Description:
70    
71    CML now compiles and works again.
72    
73    Libraries (for version > 110.33):
74    
75      $cml/cml.cm            Main CML library.
76      $cml/basis.cm          CML's version of $/basis.cm.
77      $cml/cml-internal.cm   Internal helper library.
78      $cml/core-cml.cm       Internal helper library.
79      $cml-lib/trace-cml.cm  Tracing facility.
80      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
81    
82    The installer (config/install.sh) has been taught how to properly
83    install this stuff.
84    
85    ----------------------------------------------------------------------
86    Name: Matthias Blume
87    Date: 2001/06/19 17:55:00 EDT
88    Tag: blume-20010619-instantiate
89    Description:
90    
91    This un-breaks the fix for bug 1432.
92    (The bug was originally fixed in 110.9 but I broke it again some
93    time after that.)
94    
95    ----------------------------------------------------------------------
96    Name: Matthias Blume
97    Date: 2001/06/19 17:25:00 EDT
98    Tag: blume-20010619-signals
99    Description:
100    
101    This should (hopefully) fix the long-standing signal handling bug.
102    (The runtime system was constructing a continuation record with an
103    incorrect descriptor which would cause the GC to drop data on the floor...)
104    
105    ----------------------------------------------------------------------
106    Name: Matthias Blume
107    Date: 2001/06/15 15:05:00 EDT
108    Tag: blume-20010615-moresparc
109    Description:
110    
111    Here is a short late-hour update related to Sparc c-calls:
112    
113     -- made handling of double-word arguments a bit smarter
114    
115     -- instruction selection phase tries to collapse certain clumsily
116        constructed ML-Trees; typical example:
117    
118            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
119    
120        This currently has no further impact on SML/NJ since mlriscGen does
121        not seem to generate such patterns in the first place, and c-calls
122        (which did generate them in the beginning) has meanwhile been fixed
123        so as to avoid them as well.
124    
125    ----------------------------------------------------------------------
126    Name: Matthias Blume
127    Date: 2001/06/15 15:05:00 EDT
128    Tag: blume-20010615-sparc
129    Description:
130    
131    The purpose of this update is to provide an implementation of NLFFI
132    on Sparc machines.
133    
134    Here are the changes in detail:
135    
136       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
137       the Sparc implementation of the c-calls API.
138       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
139       for accessing the ML frame.  Thus, we have a real frame pointer and
140       can freely modify %sp without need for an omit-frame-ptr phase.
141       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
142       this case.
143       * ml-nlffigen has been taught to produce code for different architectures
144       and calling conventions.
145       * In a way similar to what was done in the x86 case, the Sparc
146       backend uses its own specific extension to mltree.  (For example,
147       it needs to be able to generate UNIMP instructions which are part
148       of the calling convention.)
149       * ml-nlffi-lib was reorganized to make it more modular (in particular,
150       to make it easier to plug in new machine- and os-dependent parts).
151    
152    There are some other fairly unrelated bug fixes and cleanups as well:
153    
154       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
155       that they properly share their libraries with existing SML/NJ libraries.
156       * I fixed a minor cosmetic bug in CM, supressing certain spurious
157       follow-up error messages.
158       * Updates to CM/CMB documentation.
159    
160    TODO items:
161    
162       * MLRISC should use a different register as its asmTemp on the Sparc.
163         (The current %o2 is a really bad choice because it is part of the
164         calling conventions, so things might interfere in unexpected ways.)
165    
166    ----------------------------------------------------------------------
167    Name: Matthias Blume
168    Date: 2001/06/07
169    Tag: blume-20010607-calls
170    Description:
171    
172    A number of internal changes related to C calls and calling conventions:
173    
174    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
175       number of bytes popped implicitly (by the callee).  In most cases
176       this field is 0 but on x86/win32 it is some non-zero value.  This
177       is information provided for the benefit of the "omit-frameptr" pass.
178    2. The CALL instruction on the x86 carries a similar "pops" field.
179       The instruction selection phase copies its value from the ML-Tree
180       CALL statement.
181    3. On all other architectures, the instruction selection phase checks
182       whether "pops=0" and complains if not.
183    4. The c-calls implementation for x86 now accepts two calling conventions:
184       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
185       up after the call and pops is set to 0.  For "stdcall", the caller
186       does nothing, leaving the cleanup to the callee; pops is set to
187       the number of bytes that were pushed onto the stack.
188    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
189       distinguish between "ccall" and "stdcall".
190    6. The UNIMP instruction has been added to the supported Sparc instruction
191       set. (This is needed for implementing the official C calling convention
192       on this architecture.)
193    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
194       work with the latest CM.
195    
196    ----------------------------------------------------------------------
197    Name: Matthias Blume
198    Date: 2001/06/05 15:10:00 EDT
199    Tag: blume-20010605-cm-index
200    Description:
201    
202    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
203    
204    1. CM can now generate "index files".  These are human-readable files
205       that list on a per-.cm-file basis each toplevel symbol defined or
206       imported.  The location of the index file for
207       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
208       To enable index-file generation, set CM.Control.generate_index to true
209       or export an environment-symbol: export CM_GENERATE_INDEX=true.
210    
211       The CM manual has been updated accordingly.
212    
213    2. I made some slight modifications to the c-calls API in MLRISC.
214    
215         a) There is now a callback to support saving/restoring of
216            dedicated but caller-save registers around the actual call
217            instruction.
218         b) One can optionally specify a comment-annotation for the
219            call instruction.
220    
221    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
222       (For example, the comment annotation shows the C prototype of
223        the function being called.)
224    
225    ----------------------------------------------------------------------
226    Name: Matthias Blume
227    Date: 2001/06/01 13:30:00 EDT
228    Tag: blume-20010601-nlffi-cleanup
229    Description:
230    
231    This is mostly a cleanup of MLFFI stuff:
232    
233       - some signature files have been put into a more exposed place
234       - the ugly 'f type parameter is gone (simplifies types tremendously!)
235       - ml-nlffigen changed accordingly
236       - tutorial updated
237    
238    Other changes:
239    
240       - author's affiliation in CM manual(s) updated
241       - some more recognized keywords added to Allen's sml.sty
242    
243    ----------------------------------------------------------------------
244    Name: Matthias Blume
245    Date: 2001/05/25 15:30:00 EDT
246    Tag: blume-20010525-iptr
247    Description:
248    
249      - put the official 110.33-README (as it appears on the ftp server) under
250        CVS
251      - fixed a small bug related to incomplete pointer types in
252        ml-nlffigen
253      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
254        (it does not need the 'f type parameter)
255    
256    ----------------------------------------------------------------------
257    Name: Matthias Blume
258    Date: 2001/05/23 14:30:00 EDT
259    Tag: Release_110_33
260    Description:
261    
262    New version number (110.33).  New bootfiles.
263    
264    ----------------------------------------------------------------------
265    Name: Matthias Blume
266    Date: 2001/05/22 18:06:00 EDT
267    Tag: blume-20010522-targets
268    Description:
269    
270    Made install.sh use file config/targets.customized if it exists, falling
271    back to config/targets if it doesn't.  This way one can have a customized
272    version of the targets file without touching the "real thing", thus
273    eliminating the constant fear of accidentally checking something bogus
274    back into the CVS repository...  (File config/targets.customized must
275    not be added to the repository!)
276    
277    ----------------------------------------------------------------------
278    Name: Matthias Blume
279    Date: 2001/05/22 16:30:00 EDT
280    Tag: blume-20010522-minitut
281    Description:
282    
283    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
284       struct returns.
285    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
286       incomplete, preliminary documentation for NLFFI.
287    
288    ----------------------------------------------------------------------
289    Name: Matthias Blume
290    Date: 2001/05/14 11:30:00 EDT
291    Tag: blume-20010514-script
292    Description:
293    
294    Some bugs in install script fixed.
295    
296    In addition to that I also made a slight change to the NLFFI API:
297    Functors generated by ml-nlffigen now take the dynamic library as a
298    straight functor argument, not as a suspended one.  (The original
299    functor code used to force the suspension right away anyway, so there
300    was nothing gained by this complication of the interface.)
301    
302    ----------------------------------------------------------------------
303    Name: Matthias Blume
304    Date: 2001/05/11 14:35:00 EDT
305    Tag: blume-20010511-ml-nlffi
306    Description:
307    
308    I finally took the plunge and added my new FFI code to the main
309    repository.  For x86-linux it is now ready for prime-time.
310    
311    There are two new subdirectories of "src":
312    
313      - ml-nlffi-lib:
314           The utility library for programs using the FFI interface.
315           Here is the implementation of $/c.cm and its associated low-level
316           partners $/c-int.cm and $/memory.cm.
317      - ml-nlffigen:
318           A stand-alone program for generating ML glue code from C source
319           code.
320    
321    Building ml-nlffigen requires $/ckit-lib.cm.
322    
323    The config/install.sh script has been updates to do the Right Thing
324    (hopefully).
325    
326    Notice that the source tree for the C-Kit will not be put under "src"
327    but directly under the installation root directory.  (This is the
328    structure that currently exists on the CVS server when you check out
329    module "sml".)  Fortunately, config/install.sh knows about this oddity.
330    
331    Bugs: No documentation yet.
332    
333    ----------------------------------------------------------------------
334    Name: Matthias Blume
335    Date: 2001/05/09 16:35:00 EDT
336    Tag: blume-20010509-cpscontract
337    Description:
338    
339    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
340    wrapper/unwrapper elimination did not decrement usage counts and some
341    dead variables got overlooked by the dead-up logic.)
342    
343    ----------------------------------------------------------------------
344    Name: Lal George
345    Date: 2001/05/08  17:26:09 EDT
346    Tag: george-20010508-omit-frameptr
347    Description:
348    
349    Changes to implement the omit-frame-pointer optimization to support
350    raw C calls. For now, there is only support on the Intel x86, but
351    other architectures will follow as more experience is gained with this.
352    
353    
354    ----------------------------------------------------------------------
355    Name: Matthias Blume
356    Date: 2001/05/07 14:40:00 EDT
357    Tag: blume-20010507-proxies
358    Description:
359    
360    I made into "proxy libraries" all libraries that qualify for such a
361    change.  (A qualifying library is a library that has another library or
362    groups as its sole member and repeats that member's export list
363    verbatim.  A proxy library avoids this repetition by omitting its export
364    list, effectively inheriting the list that its (only) member exports.
365    See the CM manual for more explanation.)
366    The main effect is that explicit export lists for these libraries
367    do not have to be kepts in sync, making maintenance a bit easier.
368    
369    I also added copyright notices to many .cm-files.
370    
371    Last but not least, I made a new set of bootfiles.
372    
373    ----------------------------------------------------------------------
374    Name: Matthias Blume
375    Date: 2001/05/04 17:00:00 EDT
376    Tag: blume-20010504-cm-lsplit
377    Description:
378    
379    0. John merged pending changes to $/smlnj-lib.cm
380    
381    1. Allen's previous change accidentally backed out of one of Lal's
382       earlier changes.  I undid this mistake (re-introducing Lal's change).
383    
384    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
385       within the compiler where applicable.  There is some code simplification
386       because of that.
387    
388    3. The "split" phase (in FLINT) is now part of the default list of phases.
389       Compiler.Control.LambdaSplitting.* can be used to globally control the
390       lambda-splitting (cross-module-inlining) engine.  In addition to that,
391       it can now also be controlled on a per-source basis: CM has been taught
392       a new tool parameter applicable to ML source files.
393    
394       - To turn lambda-splitting off completely:
395            local open Compiler.Control.LambdaSplitting in
396                val _ = set Off
397            end
398       - To make "no lambda-splitting" the global default (but allow per-source
399         overriding); this is the initial setting:
400            local open Compiler.Control.LambdaSplitting in
401                val _ = set (Default NONE)
402            end
403       - To make "lambda-splitting with aggressiveness a" the global default
404         (and allow per-source overriding):
405            local open Compiler.Control.LambdaSplitting in
406                val _ = set (Default (SOME a))
407            end
408    
409       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
410         write (in the respective .cm-file):
411            a.sml (lambdasplitting:off)
412       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
413            a.sml (lambdasplitting:on)
414       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
415         <a> is a decimal non-negative integer):
416            a.sml (lambdasplitting:<a>)
417       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
418            a.sml (lambdasplitting:infinity)
419       - To use the global default for a.sml:
420            a.sml (lambdasplitting:default)
421         or simply
422            a.sml
423    
424    ----------------------------------------------------------------------
425    Name: Allen Leung
426    Date: 2001/05/04 01:57:00 EDT
427    Tag: leunga-20010504-sync
428    Description:
429    
430      MLRISC features.
431    
432      1. Fix to CMPXCHG instructions.
433      2. Changed RA interface to allow annotations in callbacks.
434      3. Added a new method to the stream interface to allow annotations updates.
435    
436    ----------------------------------------------------------------------
437    Name: Matthias Blume
438    Date: 2001/05/01 11:45:00 EDT
439    Tag: blume-20010501-pcedittmp
440    Description:
441    
442    Changed install.sh to use the current working directory instead of
443    /usr/tmp for a temporary file (pcedittmp).  The previous choice
444    of /usr/tmp caused trouble with MacOS X because of file premission
445    problems.
446    
447    ----------------------------------------------------------------------
448    Name: Matthias Blume
449    Date: 2001/04/20 11:10:00 EDT
450    Tag: blume-20010420-inMLflag
451    Description:
452    
453     - added vp_limitPtrMask to vproc-state.h
454       (for use by the raw-C-calls mechanism to implement proper interrupt
455        handling)
456     - made the ML compiler aware of various data-structure offsets so it
457       can generate code for accessing the vp_inML flag and vp_limitPtrMask
458     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
459       raw C-calls
460    
461    ----------------------------------------------------------------------
462    Name: Lal George
463    Date: 2001/04/20 09:15:28 EDT
464    Tag: george-20010420-macosX
465    Description:
466    
467     - Changes to port to Mac OS X; Darwin.
468    
469     - In the process I found that sqrt was broken on the PPC, because the
470       fsqrt instruction is not implemented.
471    
472    ----------------------------------------------------------------------
473    Name: Matthias Blume
474    Date: 2001/04/18 12:45:00 EDT
475    Tag: blume-20010418-ccalls
476    Description:
477    
478     - fixed two off-by-4 errors in the x86-specific c-calls implementation
479       (this bug prevented structure arguments containing pointers from being
480        passed correctly)
481     - changed the raw-C-call code in mlriscGen.sml in such a way that
482       structure arguments are represented as a pointer to the beginning
483       of the structure (instead of having a series of synthesized arguments,
484       one for each structure member)
485    
486     - made makeml script's verbosity level configurable via environment
487       variable (MAKEML_VERBOSITY)
488    
489     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
490       in rawmem-x86.sml; we are now using the real thing
491    
492    ----------------------------------------------------------------------
493    Name: Matthias Blume
494    Date: 2001/03/22 16:25:00 EST
495    Tag: blume-20010322-bootfiles
496    Description:
497    
498    Created a new set of bootfiles (for your automatic installation convenience).
499    
500    ----------------------------------------------------------------------
501    Name: Matthias Blume
502    Date: 2001/03/22 15:10:00 EST
503    Tag: blume-20010322-rawmem-parcm
504    Description:
505    
506    1. All "raw memory access" primitives for the new FFI are implemented now
507       (at least on the x86).
508    2. Some further cleanup of CM's parallel make mechanism.
509    
510    ----------------------------------------------------------------------
511    Name: Matthias Blume
512    Date: 2001/03/19 17:53:00 EST
513    Tag: blume-20010319-parallel
514    Description:
515    
516    Parallel make (using compile servers) now works again.
517    
518    To this end, CM.stabilize and CMB.make have been modified to work in
519    two passes when compile servers are attached:
520       1. Compile everything, do not perform stabilization; this pass
521          uses compile servers
522       2. Stabilize everything; this pass does not use compile servers
523    If there are no compile servers, the two passes are combined into one
524    (as before).  Splitting the passes increases the inherent parallelism
525    in the dependency graph because the entire graph including all
526    libraries is available at the same time.  This, in turn, improves
527    server utilization.  The downside is that the master process will
528    have to do some extra work after compilation is done (because for
529    technical reasons it must re-read all the binfiles during stabilization).
530    
531    ----------------------------------------------------------------------
532    Name: Matthias Blume
533    Date: 2001/03/16 12:22:00 EST
534    Tag: blume-20010316-bootfiles
535    Description:
536    
537    Created a new set of bootfiles (for your automatic installation convenience).
538    
539    ----------------------------------------------------------------------
540    Name: Matthias Blume
541    Date: 2001/03/16 11:00:00 EST
542    Tag: blume-20010316-MLTREE-fixup
543    Description:
544    
545    This is a minor fixup for an (untagged) earlier commit by Allen.
546    (A file was missing).
547    
548    ----------------------------------------------------------------------
549    Name: Allen Leung
550    Date: Mon Mar  5 18:54:57 EST 2001
551    Tag: leunga-20010305-cut-support
552    
553    1. New support for alternative control-flow in MLTREE.
554       Currently we support
555    
556          FLOW_TO(CALL ...., [k1,...,kn])
557    
558       This is needed for 'cuts to' in C-- and try/handle-like constructs
559       in Moby
560    
561       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
562    
563    2. Register Allocator
564    
565       Changes in interface [from Fermin, John]
566    
567    3. Alpha 8-bit SLL support [Fermin]
568    
569    4. All architectures
570    
571       A new module (ClusterExpandCopies) for expanding parallel copies.
572    
573    ----------------------------------------------------------------------
574    Name: Allen Leung
575    Date: 2001/02/27 23:07:00 EST
576    Tag: leunga-20010227-minor-stuff
577    
578    1. Alpha bug fix for CMOVNE
579    2. Handle mltree COND(..,FCMP ...,...)
580    3. Bug fix in simplifier
581    
582    ----------------------------------------------------------------------
583    Name: Matthias Blume
584    Date: 2001/01/30 17:50:00 EST
585    Tag: blume-20010130-sync
586    Description:
587    
588    This is just a minor update to sync my devel branch with the main brach.
589    The only visible change is the addition of some README files.
590    
591    ----------------------------------------------------------------------
592    Name: Matthias Blume
593    Date: 2001/01/12 23:30:00 JST
594    Tag: blume-20010112-bootfiles
595    Description:
596    
597    Made a new set of bootfiles that goes with the current state of the
598    repository.
599    
600    ----------------------------------------------------------------------
601    Name: Matthias Blume
602    Date: 2001/01/12 21:20:00 JST
603    Tag: blume-20010112-sync
604    Description:
605    
606    I am just flushing out some minor changes that had accumulated in
607    my private branch in order to sync with the main tree.  (This is
608    mainly because I had CVS trouble when trying to merge _into_ my
609    private branch.)
610    
611    Most people should be completely unaffected by this.
612    
613    ----------------------------------------------------------------------
614    Name: Allen Leung
615    Date: Thu Jan 11 21:03:00 EST 2001
616    Tag: leunga-20010111-labexp=mltree
617    Description:
618    
619    1.  Removed the type LabelExp and replace it by MLTree.
620    2.  Rewritten mltree-simplify with the pattern matcher tool.
621    3.  There were some bugs in alpha code generator which would break
622        64-bit code generation.
623    4.  Redo the tools to generate code with the
624    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
625        are now generated by perl scripts.
626    
627    ----------------------------------------------------------------------
628    Name: Matthias Blume
629    Date: 2001/01/10 21:55:00 JST
630    Tag: blume-20010110-rcc
631    Description:
632    
633    The RCC stuff now seems to work (but only on the x86).
634    This required hacking of the c-calls interface (and -implementation) in
635    MLRISC.
636    
637    Normal compiler users should be unaffected.
638    
639    ----------------------------------------------------------------------
640    Name: Matthias Blume
641    Date: 2001/01/09 01:20:00 JST
642    Tag: blume-20010109-rcc
643    Description:
644    
645    This is a fairly big patch, flushing out a large number of pending
646    changes that I made to my development copy over the last couple of days.
647    
648    Of practical relevance at this moment is a workaround for a pickling
649    bug that Allen ran into the other day.  The cause of the bug itself is
650    still unknown and it might be hard to fix it properly, but the
651    workaround has some merits of its own (namely somewhat reducing pickling
652    overhead for certain libraries).  Therefore, I think this solution should
653    be satisfactory at this time.
654    
655    The rest of the changes (i.e., the vast majority) has to do with my
656    ongoing efforts of providing direct support for C function calls from
657    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
658    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
659    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
660    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
661    handle RCC.
662    
663    The last part (the changes to mlriscGen.sml) are still known to be
664    wrong on the x86 and not implemented on all other architectures.  But
665    the infrastructure is in place. I had to change a few functor
666    signatures in the backend to be able to route the CCalls interface
667    from MLRISC there, and I had to specialize the mltree type (on the
668    x86) to include the necessary extensions. (The extensions themselves
669    were already there and redy to go in MLRISC/x86).
670    
671    Everything should be very happy as soon as someone helps me with
672    mlriscGen.sml...
673    
674    In any case, nothing of this should matter to anyone as long as the
675    new primop is not being used (which is going to be the case unless you
676    find it where I hid it :). The rest of the compiler is completely
677    unaffected.
678    
679    ----------------------------------------------------------------------
680    Name: Matthias Blume
681    Date: 2001/01/05 00:30:00 JST
682    Tag: blume-20010105-primops
683    Description:
684    
685    Added some experimental support for work that I am doing right now.
686    These changes mostly concern added primops, but there is also a new
687    experimental C library in the runtime system (but currently not enabled
688    anywhere except on Linux/X86).
689    
690    In the course of adding primops (and playing with them), I discovered that
691    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
692    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
693    To recover, I made type information in INL_PRIM mandatory and changed
694    prim.sml as well as built-in.sml accordingly.  The InLine structure now
695    has complete, correct type information (i.e., no bottom types).
696    
697    Since all these changes mean that we need new binfiles, I also bumped the
698    version number to 110.32.1.
699    
700    ----------------------------------------------------------------------
701    Name: Matthias Blume
702    Date: 2000/12/30 22:10:00 JST
703    Tag: blume-20001230-various
704    Description:
705    
706    Added proxy libraries for MLRISC and let MLRISC libraries refer
707    to each other using path anchors.  (See CM manual for explanation.)
708    
709    Updated CM documentation.
710    
711    Fixed some bugs in CM.
712    
713    Implemented "proxy" libraries (= syntactic sugar for CM).
714    
715    Added "-quiet" option to makeml and changed runtime system accordingly.
716    
717    Added cleanup handler for exportML to reset timers and compiler stats.
718    
719    ----------------------------------------------------------------------
720    Name: Lal George
721    Date: 2000/12/22 22:22:58 EST 2000
722    Tag: Release_110_32
723    Description:
724    
725            Infinite precision used throughout MLRISC.
726            see MLRISC/mltree/machine-int.sig
727    
728    ----------------------------------------------------------------------
729    Name: Matthias Blume
730    Date: 2000/12/22 23:16:00 JST
731    Tag: blume-20001222-warn
732    Description:
733    
734    Corrected wording and formatting of some CM warning message which I
735    broke in my previous patch.
736    
737    ----------------------------------------------------------------------
738    Name: Matthias Blume
739    Date: 2000/12/22 21:20:00 JST
740    Tag: blume-20001222-anchorenv
741    Description:
742    
743    Fixed CM's handling of anchor environments in connection with CMB.make.
744    
745    ----------------------------------------------------------------------
746    Name: Matthias Blume
747    Date: 2000/12/22 13:15:00 JST
748    Tag: blume-20001222-cleanup
749    Description:
750    
751    Removed src/cm/ffi which does not (and did not) belong here.
752    
753    ----------------------------------------------------------------------
754    Name: Matthias Blume
755    Date: 2000/12/21 23:55:00 JST
756    Tag: blume-20001221-exn
757    Description:
758    
759    Probably most important: CM no longer silently swallows all exceptions
760    in the compiler.
761    Plus: some other minor CM changes.  For example, CM now reports some
762    sizes for generated binfiles (code, data, envpickle, lambdapickle).
763    
764    ----------------------------------------------------------------------
765    Name: Matthias Blume
766    Date: 2000/12/15 00:01:05 JST
767    Tag: blume-20001215-dirtool
768    Description:
769    
770    - "dir" tool added.
771    - improvements and cleanup to Tools structure
772    - documentation updates
773    
774    ----------------------------------------------------------------------
775    Name: Allen Leung
776    Date: Thu Dec 14 03:45:24 EST 2000
777    Description:
778    Tag:  leunga-20001214-int-inf
779    Description:
780    
781       In IntInf, added these standard functions, which are missing from our
782    implementation:
783    
784        andb : int * int -> int
785        xorb : int * int -> int
786        orb  : int * int -> int
787        notb : int -> int
788         <<   : int * word -> int
789        ~>>  : int * word -> int
790    
791       Not tested, I hope they are correct.
792    
793    ----------------------------------------------------------------------
794    Name: Allen Leung
795    Date: Fri Dec  8 19:23:26 EST 2000
796    Description:
797    Tag:  leunga-20001208-nowhere
798    Description:
799    
800      Slight improvements to the 'nowhere' tool to handle OR-patterns,
801    to generate better error messages etc.  Plus a brief manual.
802    
803    ----------------------------------------------------------------------
804    Name: Lal George
805    Date: 2000/12/08 09:54:02 EST 2000
806    Tag: Release_110_31
807    Description:
808    
809    - Version 110.31
810    ----------------------------------------------------------------------
811    Name: Allen Leung
812    Date: Thu Dec  7 22:01:04 EST 2000
813    Tag:  leunga-20001207-cell-monster-hack
814    Description:
815    
816    Major MLRISC internal changes.  Affect all clients.
817    Summary:
818    
819    1.  Type CELLS.cell = int is now replaced by a datatype.
820        As a result, the old regmap is now gone.  Almost all interfaces
821        in MLRISC change as a consequence.
822    
823    2.  A new brand version of machine description tool (v3.0) that generates
824        modules expecting the new interface.  The old version is removed.
825    
826    3.  The RA interface has been further abstracted into two new functors.
827        RISC_RA and X86RA.  These functors have much simpler interfaces.
828        [See also directory MLRISC/demo.]
829    
830    4.  Some other new source->source code generation tools are available:
831    
832        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
833        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
834           I use this tool to generate the peephole optimizers---with the new
835           cell type changes, peephole rules are becoming difficult to write
836           without conditional pattern matching.
837    
838    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
839        cover the entire MLRISC source tree so many things broke.
840    
841    6.  CM files have been moved to the subdirectory MLRISC/cm.
842        They are moved because there are a lot of them and they clutter up the
843        root dir.
844    
845    7.  More detailed documentation to come...
846    
847        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
848        a makeml -rebuild first.  This is because of other other
849        changes that Matthias has made (see below).
850    
851    
852    ----------------------------------------------------------------------
853    Name: Matthias Blume
854    Date: 2000/11/30 23:12:00 JST
855    Tag: blume-20001130-filereorg
856    Description:
857    
858    Some manual updates and some file reorganizations in CM.
859    
860    ----------------------------------------------------------------------
861    Name: Matthias Blume
862    Date: 2000/11/24 17:45:00 JST
863    Tag: blume-20001124-link
864    Description:
865    
866    Drastically improved link traversal code for the case that the dynamic
867    value was already loaded at bootstrap time.  As a result, CM and CMB
868    now both load blazingly fast -- even on a very slow machine.  Also,
869    memory consumption has been further reduced by this.
870    
871    Warning: The format of the PIDMAP file has changed.  THerefore, to
872    bootstrap you have to do this:
873    
874    1. Run CMB.make
875    2. Make a symbolic link for the boot directory:
876         ln -s sml.boot.ARCH-OS xxx
877    3. "Rebuild" the boot directory:
878         ./makeml -boot xxx -rebuild sml ; rm xxx
879    4. Boot normally:
880          ./makeml
881    
882    ----------------------------------------------------------------------
883    Name: Matthias Blume
884    Date: 2000/11/21 21:20:00 JST
885    Tag: blume-20001121-tools
886    Description:
887    
888    Continued hacking on autoloading problem -- with success this time.
889    Also changed tool-plugin mechanism.  See new CM manual.
890    
891    ----------------------------------------------------------------------
892    Name: Matthias Blume
893    Date: 2000/11/19 14:30:00 JST
894    Tag:  blume-20001119-autoload
895    Description:
896    
897    Some hacking to make autoloading faster.  Success for CMB, no success
898    so far for CM.  There is a reduced structure CM' that autoloads faster.
899    (This is a temporary, non-documented hack to be eliminated again when
900    the general problem is solved.)
901    
902    ----------------------------------------------------------------------
903    Name: Matthias Blume
904    Date: 2000/11/17 14:10:00 JST
905    Tag: blume-20001117-pickle-lib
906    Description:
907    
908    1. Eliminated comp-lib.cm
909    2. Made pickle-lib.cm
910    3. Eliminated all uses of intset.sml (from comp-lib.cm)
911    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
912       equivalent constructs from smlnj-lib.cm (INtHashTable).
913    5. Point 4. also goes for those uses of intmap.* in MLRISC.
914       Duplicated intmap modules thrown out.
915    6. Hunted down all duplicated SCC code and replaced it with
916       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
917    7. Rewrote Feedback module.
918    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
919       should be thrown out and equivalent modules from smlnj-lib.cm
920       should be used (IntRedBlackSet, IntListSet, ...).
921    
922    Confirmed that compiler compiles to fixpoint.
923    
924    ----------------------------------------------------------------------
925    Name: Allen Leung
926    Date: 2000/11/10 18:00:00
927    Tag: leunga-20001110-new-x86-fp
928    
929    A new x86 floating point code generator has been added.
930    By default this is turned off.  To turn this on, do:
931    
932        CM.autoload "$smlnj/compiler.cm";
933        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
934    
935    Changes:
936    
937    1.  Changed FTAN to FPTAN so that the assembly output is correct.
938    2.  Changed the extension callback for FTANGENT to generate:
939    
940              fptan
941              fstp  %st(0)
942        instead of
943              fptan
944              fstpl ftempmem
945    
946    3.  Numerous assembly fixes for x86.
947    
948    5.  Cleaned up the machine code output module x86/x86MC.sml and added
949        support for a whole bunch of instructions and addressing modes:
950    
951          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
952          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
953          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
954          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
955          fxch %st(n)
956          fld %st(n)
957          fst %st(n)
958          fst mem
959          fstp %st(n)
960          fucom %st(n)
961          fucomp %st(n)
962    
963        All these are now generated when the fast fp mode is turned on.
964    
965    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
966    
967    ----------------------------------------------------------------------
968    Name: Matthias Blume
969    Date: 2000/11/09 11:20:00 JST
970    Tag: blume-20001109-scc
971    Description:
972    
973    Eliminated some code duplication:
974    
975    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
976       (Otherwise the functor is useless.)
977    2. Used GraphSCCFn where SCCUtilFun was used previously.
978    3. Got rid of SCCUtilFun (in comp-lib.cm).
979    
980    ----------------------------------------------------------------------
981    Name: Lal George
982    Date: 2000/11/06 09:02:21 EST 2000
983    Tag: Release_110_30
984    Description:
985    
986    - Version 110.30
987    ----------------------------------------------------------------------
988    Name: Matthias Blume
989    Date: 2000/11/04 14:45:00
990    Tag: blume-20001104-mlbuild
991    Description:
992    
993    - Made ml-build faster on startup.
994    - Documentation fixes.
995    
996    ----------------------------------------------------------------------
997    Name: Matthias Blume
998    Date: 2000/11/02 17:00:00 JST
999    Tag: blume-20001102-condcomp
1000    Description:
1001    
1002    - Small tweaks to pickler -- new BOOTFILES!
1003    - Version bumped to 110.29.2.
1004    - Added conditional compilation facility to init.cmi (see comment there).
1005    ----------------------------------------------------------------------
1006    Name: Allen Leung
1007    Date: 2000/10/23 19:31:00
1008    Tag: leunga-20001023-demo-ra
1009    
1010    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
1011    2. Test programs for the graph library updated
1012    3. Some new MLRISC demo programs added
1013    
1014    ----------------------------------------------------------------------
1015    Name: Matthias Blume
1016    Date: 2000/08/31 22:15:00 JST
1017    Tag: blume-20001017-errmsg
1018    Description:
1019    
1020    More error message grief: Where there used to be no messages, there
1021    now were some that had bogus error regions.  Fixed.
1022    
1023    ----------------------------------------------------------------------
1024    Name: Matthias Blume
1025    Date: 2000/08/31 17:30:00 JST
1026    Tag: blume-20001017-v110p29p1
1027    Description:
1028    
1029    I made a version 110.29.1 with new bootfiles.
1030    
1031    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
1032              CM documentation changes and a small bugfix in CM's error reporting.
1033    
1034    ----------------------------------------------------------------------
1035    Name: Lal George
1036    Date: 2000/09/27 14:42:35 EDT
1037    Tag: george-20000927-nodestatus
1038    Description:
1039    
1040    Changed the type of the nodestatus, so that:
1041    
1042            SPILLED(~1)             is now SPILLED
1043            SPILLED(m) where m>=0   is now MEMREG(m)
1044            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
1045    
1046    ----------------------------------------------------------------------
1047    Name: Matthias Blume
1048    Date: 2000/09/07 14:45:00 JST
1049    Tag: blume-20000907-cmerrmsg
1050    Description:
1051    
1052    Small tweak to CM to avoid getting ML syntax error messages twice.
1053    
1054    ----------------------------------------------------------------------
1055    Name: Matthias Blume
1056    Date: 2000/08/31 18:00:00 JST
1057    Tag: blume-20000831-cvsbootfiles
1058    Description:
1059    
1060    New URL for boot files (because the 110.29 files on the BL server do
1061    now work correctly with my updated install scripts for yacc and lex).
1062    
1063    ----------------------------------------------------------------------
1064    Name: Matthias Blume
1065    Date: 2000/08/08 12:33:00 JST
1066    Tag: blume-20000808-manual
1067    Description:
1068    
1069    Tiny update to CM manual.
1070    
1071    ----------------------------------------------------------------------
1072    Name: Allen Leung
1073    Date: 2000/08/7 19:31:00
1074    Tag: leunga-20000807-a-whole-bunch-of-stuff
1075    
1076      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
1077    and Mobdy.
1078    
1079    1.  x86
1080    
1081       a.  Fixes to peephole module by John and Dan.
1082       b.  Assembly fix to SETcc by Allen.
1083       c.  Fix to c-call by John.
1084       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
1085       e.  Instruction selection optimization to SETcc as suggested by John.
1086    
1087           For example,
1088    
1089            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
1090    
1091           should generate:
1092    
1093            MOVL a, x
1094            SUBL b, x
1095            SHRL 31, x
1096    
1097    2.  IR stuff
1098    
1099         A bunch of new DJ-graph related algorithms added.  These
1100         speed up SSA construction.
1101    
1102    3.  SSA + Scheduling
1103    
1104         Added code for SSA and scheduling to the repository
1105    
1106    ----------------------------------------------------------------------
1107    Name: Lal George
1108    Date: 2000/07/27 11:53:14 EDT
1109    
1110    Tag: lal-20000727-linux-ppc
1111    Description:
1112    
1113     Made changes to support Linux PPC.
1114     p.s. I have confirmation that the 110.29 boot files work fine.
1115    
1116    ----------------------------------------------------------------------
1117    Name: Matthias Blume
1118    Date: 2000/07/27 17:40:00 JST
1119    Tag: blume-20000727-scripts
1120    Description:
1121    
1122    !!!! WARNING !!!!
1123    You must recompile the runtime system!
1124    !!!! WARNING !!!!
1125    
1126    This is basically another round of script-enhancements:
1127    
1128    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
1129       and undefine CM preprocessor symbols.
1130    
1131    2. ml-build avoids generating a new heap image if it finds that the
1132       existing one is still ok.  (The condition is that no ML file had to
1133       be recompiled and all ML files are found to be older that the heap
1134       file.)
1135    
1136       To make this work smoothly, I also hacked the runtime system as
1137       well as SMLofNJ.SysInfo to get access to the heap image suffix
1138       (.sparc-solaris, ...) that is currently being used.
1139    
1140       Moreover, the signature of CM.mk_standalone has changed.  See the
1141       CM manual.
1142    
1143    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
1144       CM manual for details.)
1145    
1146    4. More CM manual updates:
1147        - all of the above has been documented.
1148        - there is now a section describing the (CM-related) command line
1149          arguments that are accepted by the "sml" command
1150    
1151    ----------------------------------------------------------------------
1152    Name: Matthias Blume
1153    Date: 2000/07/25 16:20:00 JST
1154    Tag: blume-20000725-makedepend
1155    Description:
1156    
1157    Added a script called ml-makedepend.  This can be used in makefiles
1158    for Unix' make in a way very similar to the "makedepend" command for
1159    C.
1160    
1161    The script internally uses function CM.sources.
1162    
1163    Synopsis:
1164    
1165        ml-makedepend [-f makefile] cmfile targetname
1166    
1167    The default for the makefile is "makefile" (or "Makefile" should
1168    "makefile" not exist).
1169    
1170    ml-makedepend adds a cmfile/targetname-specific section to this
1171    makefile (after removing the previous version of this section).  The
1172    section contains a single dependency specification with targetname on
1173    the LHS (targetname is an arbitrary name), and a list of files derived
1174    from the cmfile on the RHS.  Some of the files on the RHS are
1175    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
1176    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
1177    path names.  The makefile writer is responsible for making sure that
1178    these variables have correct at the time "make" is invoked.
1179    
1180    ----------------------------------------------------------------------
1181    Name: Matthias Blume
1182    Date: 2000/07/22 23:30:00 JST
1183    Tag: blume-20000722-urlupdate
1184    Description:
1185    
1186    Changed BOOT and config/srcarchiveurl to point to BL server:
1187    
1188        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
1189    
1190    ----------------------------------------------------------------------
1191    Name: Matthias Blume
1192    Date: 2000/07/18 18:00:00 JST
1193    Tag: blume-20000718-Version_110_29
1194    Description:
1195    
1196    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
1197    
1198    2. Updated config/version to 110.29
1199    
1200    3. Updated config/srcarchiveurl
1201    
1202    3. New boot files!
1203       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
1204    
1205    ----------------------------------------------------------------------
1206    Name: Matthias Blume
1207    Date: 2000/07/11 13:58:00 JST
1208    Tag: blume-20000711-doctypo
1209    Description:
1210    
1211    Fixed a few typos in CM manual.
1212    
1213    ----------------------------------------------------------------------
1214    Name: Allen Leung
1215    Date: 2000/06/15 00:38:00
1216    Tag: leunga-20000704-sparc-x86
1217    
1218    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
1219    2. fix to x86 RET bug [found by Dan Grossman]
1220    3. sparc assembly bug fix for ticc instructions [found by Fermin]
1221    
1222       Affects c-- and moby only
1223    
1224    ----------------------------------------------------------------------
1225    Name: Matthias Blume
1226    Date: 2000/07/04 15:26:00
1227    Tag: blume-20000704-trigger
1228    Description:
1229    
1230    1. Improvements to CM manual.
1231    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
1232       of getting a back-trace.  The function, when called, raises an
1233       internal exception which explicitly carries the full back-trace history,
1234       so it is unaffected by any intervening handle-raise pairs ("trivial"
1235       or not).  The interactive loop will print that history once it arrives
1236       at top level.
1237       Short of having all exceptions implicitly carry the full history, the
1238       recommended way of using this facility is:
1239         - compile your program with instrumentation "on"
1240         - run it, when it raises an exception, look at the history
1241         - if the history is "cut off" because of some handler, go and modify
1242           your program so that it explicitly calls BTrace.trigger
1243         - recompile (still instrumented), and rerun; look at the full history
1244    
1245    ----------------------------------------------------------------------
1246    Name: Matthias Blume
1247    Date: 2000/07/03 15:36:00 JST
1248    Tag: blume-20000702-manual
1249    Description:
1250    
1251    Small corrections and updates to CM manual.
1252    
1253    ----------------------------------------------------------------------
1254    Name: Matthias Blume
1255    Date: 2000/06/29 16:04:00 JST
1256    Tag: blume-20000629-yacctool
1257    Description:
1258    
1259    Changes:
1260    
1261    1. Class "mlyacc" now takes separate arguments to pass options to
1262       generated .sml- and .sig-files independently.
1263    2. Corresponding CM manual updates.
1264    3. BTrace module now also reports call sites.  (However, for loop clusters
1265       it only shows from where the cluster was entered.)  There are associated
1266       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1267    
1268    ----------------------------------------------------------------------
1269    Name: Matthias Blume
1270    Date: 2000/06/27 16:51:00 JST
1271    Tag: blume-20000627-noweb
1272    Description:
1273    
1274    Changes:
1275    
1276     1. Implemented "subdir" and "witness" options for noweb tool.
1277        This caused some slight internal changes in CM's tool implementation.
1278     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
1279        some remaining issues from earlier path anchor changes.
1280     3. Updated CM manual accordingly.
1281    
1282     4. Changed implementation of back-tracing so that I now consider it
1283        ready for prime-time.
1284    
1285        In particular, you don't have to explicitly trigger the back-trace
1286        anymore.  Instead, if you are running BTrace-instrumented code and
1287        there is an uncaught exception (regardless of whether or not it was
1288        raised in instrumented code), the top-level evalloop will print
1289        the back-trace.
1290    
1291        Features:
1292    
1293          - Instrumented and uninstrumented code work together seemlessly.
1294            (Of course, uninstrumented code is never mentioned in actual
1295             back-traces.)
1296    
1297          - Asymptotic time- and space-complexity of instrumented code is
1298            equal to that of uninstrumented code.  (This means that
1299            tail-recursion is preserved by the instrumentation phase.)
1300    
1301          - Modules whose code has been instrumented in different sessions
1302            work together without problem.
1303    
1304          - There is no penalty whatsoever on uninstrumented code.
1305    
1306          - There is no penalty on "raise" expressions, even in
1307            instrumented code.
1308    
1309        A potential bug (or perhaps it is a feature, too):
1310    
1311          A back-trace reaches no further than the outermost instrumented
1312          non-trivial "raise".  Here, a "trivial" raise is one that is the
1313          sole RHS of a "handle" rule.  Thus, back-traces reach trough
1314    
1315               <exp> handle e => raise e
1316    
1317          and even
1318    
1319               <exp> handle Foo => raise Bar
1320    
1321          and, of course, through
1322    
1323               <exp> handle Foo => ...
1324    
1325         if the exception was not Foo.
1326    
1327         Back-traces always reach right through any un-instrumented code
1328         including any of its "handle" expressions, trivial or not.
1329    
1330       To try this out, do the following:
1331    
1332         - Erase all existing binfiles for your program.
1333           (You may keep binfiles for those modules where you think you
1334            definitely don't need back-tracing.)
1335         - Turn on back-trace instrumentation:
1336              SMLofNJ.Internals.BTrace.mode (SOME true);
1337         - Recompile your program.  (I.e., run "CM.make" or "use".)
1338         - You may now turn instrumentation off again (if you want):
1339              SMLofNJ.Internals.BTrace.mode (SOME false);
1340         - Run your program as usual.  If it raises an exception that
1341           reaches the interactive toplevel, then a back-trace will
1342           automatically be printed.  After that, the toplevel loop
1343           will print the exception history as usual.
1344    
1345    ----------------------------------------------------------------------
1346    Name: Matthias Blume
1347    Date: 2000/06/26 09:56:46 JST
1348    Tag: blume-20000626-setup
1349    Description:
1350    
1351    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1352          ML code before and after compiling a file (e.g., to set compiler
1353          flags)
1354    
1355    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1356              - associated changes to btrace.sml (BTrace instrumentation pass)
1357              - cleaner implementation of btimp.sml (BTrace tracing and report
1358                module)
1359    
1360    CM manual: * new path encoding documented
1361               * description of setup-parameter to "sml" added
1362    
1363    The biggest user-visible change to back-tracing is that it is no
1364    longer necessary to compile all traced modules within the same
1365    session.  (This was a real limitation.)
1366    
1367    ----------------------------------------------------------------------
1368    Name: Matthias Blume
1369    Date: 2000/06/24 12:40:00 JST
1370    Tag: blume-20000624-startup
1371    Description:
1372    
1373    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1374    bit too often -- to put it mildly. :)
1375    
1376    ----------------------------------------------------------------------
1377    Name: Matthias Blume
1378    Date: 2000/06/23 18:20:00 JST
1379    Tag: blume-20000623-btrace
1380    Description:
1381    
1382    This updates adds a backtrace facility to aid programmers in debugging
1383    their programs.  This involves the following changes:
1384    
1385    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1386       keeping track of the current call stack.  When programs are compiled
1387       in a special mode, the compiler will insert calls to these hooks
1388       into the user program.
1389       "Hook" means that it is possible for different implementations of
1390       back-tracing to register themselves (at different times).
1391    
1392    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1393       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1394       It can be turned on using this call:
1395         SMLofNJ.Internals.BTrace.mode (SOME true);
1396       Turning it off again:
1397         SMLofNJ.Internals.BTrace.mode (SOME false);
1398       Querying the current status:
1399         SMLofNJ.Internals.BTrace.mode NONE;
1400       Annotated programs are about twice as big as normal ones, and they
1401       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1402       where all hooks do nothing).  The slowdown with a plugin that is
1403       actually useful (such as the one supplied by default) is even greater,
1404       but in the case of the default plugin it is still only an constant
1405       factor (amortized).
1406    
1407    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1408       with a sub-structure BTrace for controlling back-tracing.  In particular,
1409       the above-mentioned function "mode" controls whether the annotation
1410       phase is invoked by the compiler.  Another important function is
1411       "trigger": when called it aborts the current execution and causes
1412       the top-level loop to print a full back-trace.
1413    
1414    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1415       for back-tracing.  It keeps track of the dynamic call stack and in
1416       addition to that it keeps a partial history at each "level" of that
1417       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1418       d and b (at separate times, dynamically), then the report will show:
1419    
1420       GOTO   d
1421             /c
1422       GOTO  \b
1423       CALL   a
1424    
1425       This shows that there was an initial non-tail call of a, then a
1426       tail-call to b or c, looping behavior in a cluster of functions that
1427       consist of b and c, and then a goto from that cluster (i.e., either from
1428       b or from c) to d.
1429    
1430       Note that (depending on the user program) the amount of information
1431       that the back-trace module has to keep track of at each level is bounded
1432       by a constant.  Thus, the whole implementation has the same asymptotical
1433       complexity as the original program (both in space and in time).
1434    
1435    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1436       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1437       which is raised by the "trigger" function mentioned above.
1438    
1439    Notes on usage:
1440    
1441    - Annotated code works well together with unannotated code:
1442    Unannotated calls simply do not show up at all in the backtrace.
1443    
1444    - It is not a good idea to let modules that were annotated during
1445    different sessions run at the same time.  This is because the compiler
1446    chooses small integers to identify individual functions, and there
1447    will be clashes if different modules were compiled in separate sessions.
1448    (Nothing will crash, and you will even be told about the clashes, but
1449    back-trace information will in general not be useful.)
1450    
1451    - Back-tracing can be confused by callcc and capture.
1452    
1453    - The only way of getting a back-trace right now is to explicitly
1454    invoke the "trigger" function from your user program.  Eventually, we
1455    should make every exception carry back-trace information (if
1456    available).  But since this creates more overhead at "raise"-time
1457    (similar to the current exnHistory overhead), I have not yet
1458    implemented this.  (The implementation will be rather easy.)  With
1459    exceptions carrying back-trace information, this facility will be even
1460    more useful because users don't need to modify their programs...
1461    
1462    - While it is possible to compile the compiler with back-trace
1463    annotations turned on (I did it to get some confidence in
1464    correctness), you must make absolutely sure that core.sml and
1465    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1466    be compiled with annotation because there is no core access yet, but
1467    if you compile btimp.sml with annotation, then the system will go into
1468    an infinite recursion and crash.)
1469    Since CM currently does not know about BTrace, the only way to turn
1470    annotations on and off for different modules of the compiler is to
1471    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1472    this is awkward and clumsy.
1473    
1474    Sample sessions:
1475    
1476    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1477    - SMLofNJ.Internals.BTrace.mode (SOME true);
1478    [autoloading]
1479    [autoloading done]
1480    val it = false : bool
1481    - structure X = struct
1482    -     fun main n = let
1483    -         fun a (x, 0) = d x
1484    -           | a (x, n) = b (x, n - 1)
1485    -         and b (x, n) = c (x, n)
1486    -         and c (x, n) = a (x, n)
1487    -         and d x = e (x, 3)
1488    -         and e (x, 0) = f x
1489    -           | e (x, n) = e (x, n - 1)
1490    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1491    -           | f n = n * g (n - 1)
1492    -         and g n = a (n, 3)
1493    -     in
1494    -         f n
1495    -     end
1496    - end;
1497    structure X : sig val main : int -> int end
1498    - X.main 3;
1499    *** BACK-TRACE ***
1500    GOTO   stdIn:4.2-13.20: X.main[2].f
1501    GOTO-( stdIn:4.2-13.20: X.main[2].e
1502    GOTO   stdIn:4.2-13.20: X.main[2].d
1503         / stdIn:4.2-13.20: X.main[2].a
1504         | stdIn:4.2-13.20: X.main[2].b
1505    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1506    CALL   stdIn:4.2-13.20: X.main[2].g
1507    GOTO   stdIn:4.2-13.20: X.main[2].f
1508    GOTO-( stdIn:4.2-13.20: X.main[2].e
1509    GOTO   stdIn:4.2-13.20: X.main[2].d
1510         / stdIn:4.2-13.20: X.main[2].a
1511         | stdIn:4.2-13.20: X.main[2].b
1512    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1513    CALL   stdIn:4.2-13.20: X.main[2].g
1514    GOTO   stdIn:4.2-13.20: X.main[2].f
1515    GOTO-( stdIn:4.2-13.20: X.main[2].e
1516    GOTO   stdIn:4.2-13.20: X.main[2].d
1517         / stdIn:4.2-13.20: X.main[2].a
1518         | stdIn:4.2-13.20: X.main[2].b
1519    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1520    CALL   stdIn:4.2-13.20: X.main[2].g
1521    GOTO   stdIn:4.2-13.20: X.main[2].f
1522    CALL   stdIn:2.15-17.4: X.main[2]
1523    -
1524    
1525    (Note that because of a FLINt bug the above code currently does not
1526    compile without BTrace turned on.)
1527    
1528    Here is another example, using my modified Tiger compiler:
1529    
1530    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1531    - SMLofNJ.Internals.BTrace.mode (SOME true);
1532    [autoloading]
1533    [autoloading done]
1534    val it = false : bool
1535    - CM.make "sources.cm";
1536    [autoloading]
1537    ...
1538    [autoloading done]
1539    [scanning sources.cm]
1540    [parsing (sources.cm):parse.sml]
1541    [creating directory CM/SKEL ...]
1542    [parsing (sources.cm):tiger.lex.sml]
1543    ...
1544    [wrote CM/sparc-unix/semant.sml]
1545    [compiling (sources.cm):main.sml]
1546    [wrote CM/sparc-unix/main.sml]
1547    [New bindings added.]
1548    val it = true : bool
1549    - Main.compile ("../testcases/merge.tig", "foo.out");
1550    *** BACK-TRACE ***
1551    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
1552    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1553    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1554    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1555    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1556    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1557    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
1558         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
1559    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
1560    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1561    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
1562    CALL   main.sml:1.18-118.4: Main.compile[2]
1563    -
1564    
1565    ----------------------------------------------------------------------
1566    Name: Matthias Blumen
1567    Date: 2000/06/21 18:00:00 JST
1568    Tag: blume-20000621-manual
1569    Description:
1570    
1571    CM manual update: Path environments documented.
1572    
1573    ----------------------------------------------------------------------
1574    Name: Matthias Blume
1575    Date: 2000/06/19 13:40:00
1576    Tag: blume-20000619-manual
1577    Description:
1578    
1579    CM manual and system/README update.  This only covers the fact that
1580    there are no more implicit anchors.  (Path environments and the "bind"
1581    option to "cm" have yet to be documented.)
1582    
1583    ----------------------------------------------------------------------
1584    Name: Matthias Blume
1585    Date: 2000/06/19 11:05:00 JST
1586    Tag: blume-20000619-chdir-bugfix
1587    Description:
1588    
1589    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1590    
1591    ----------------------------------------------------------------------
1592    Name: Matthias Blume
1593    Date: 2000/06/18 22:00:10 JST
1594    Tag: blume-20000618-implicit-anchors-really-gone
1595    Description:
1596    
1597    I updates the previous HISTORY entry where I forgot to mention that
1598    implicit anchors are no longer with us.
1599    
1600    The current update also gets rid of the (now useless) controller
1601    CM.Control.implicit_anchors.
1602    
1603    ----------------------------------------------------------------------
1604    Name: Matthias Blume
1605    Date: 2000/06/16 17:30:00 JST
1606    Tag: blume-20000616-anchorenv
1607    Description:
1608    
1609    This patch implements the long anticipated (just kidding :) "anchor
1610    environment" mechanism.  In the course of doing this, I also
1611    re-implemented CM's internal "SrcPath" module from scratch.  The new
1612    one should be more robust in certain boundary cases.  In any case, it
1613    is a lot cleaner than its predecessor (IMHO).
1614    
1615    This time, although there is yet another boot file format change, I
1616    kept the unpickler backward-compatible.  As a result, no new bootfiles
1617    are necessary and bootstrapping is straightforward.  (You cannot read
1618    new bootfiles into an old system, but the other way around is no
1619    problem.)
1620    
1621    Visible changes:
1622    
1623    ** 0. Implicit path anchors (without the leading $-symbol) are no
1624    longer recognized at all. This means that such path names are not
1625    illegal either.  For example, the name basis.cm simply refers to a
1626    local file called "basis.cm" (i.e, the name is an ordinary path
1627    relative to .cm-files directory).  Or, to put it differently, only
1628    names that start with $ are anchored paths.
1629    
1630    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1631    vanished.
1632    
1633    John (Reppy) had critizised this as soon as I originally proposed and
1634    implemented it, but at that time I did not really deeply believe
1635    him. :) Now I came full-circle because I need the $<singlearc> syntax
1636    in another place where it cannot be seen as an abbreviation for
1637    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1638    seems to mean (i.e., it "expands" into the corresponding anchor
1639    value).
1640    
1641    However, when paths are used as members in CM description files, it
1642    continues to be true that there must be at least another arc after the
1643    anchor.  This is now enforced separately during semantic analysis
1644    (i.e., from a lexical/syntactical point of view, the notation is ok.)
1645    
1646    ** 2. The "cm" class now accepts an option "bind".  The option's value
1647    is a sub-option list of precisely two items -- one labeled "anchor"
1648    and the other one labeled "value".  As you might expect, "anchor" is
1649    used to specify an anchor name to be bound, and "value" specifies what
1650    the anchor is being bound to.
1651    
1652    The value must be a directory name and can be given in either standard
1653    syntax (including the possibility that it is itself an anchored path)
1654    or native syntax.
1655    
1656    Examples:
1657    
1658       foo.cm (bind:(anchor:bar value:$mystuff/bar))
1659       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1660    
1661    and so on.
1662    
1663    The meaning of this is that the .cm-file will be processed with an
1664    augmented anchor environment where the given anchor(s) is/are bound to
1665    the given values(s).
1666    
1667    The rationale for having this feature is this: Suppose you are trying
1668    to use two different (already stable) libraries a.cm and b.cm (that
1669    you perhaps didn't write yourself).  Further, suppose each of these
1670    two libraries internally uses its own auxiliary library $aux/lib.cm.
1671    Normally you would now have a problem because the anchor "lib" can not
1672    be bound to more than one value globally.  Therefore, the project that
1673    uses both a.cm and b.cm must locally redirect the anchor to some other
1674    place:
1675    
1676       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1677       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1678    
1679    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1680    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
1681    
1682    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1683    you when you do so at the time of CM.stabilize).  Therefore, you can
1684    also use an anchored path as the value:
1685    
1686      a.cm (bind:(anchor:lib value:$a-lib))
1687      b.cm (bind:(anchor:lib value:$b-lib))
1688    
1689    Now you can globally configure (using the usual CM.Anchor.anchor or
1690    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1691    itself is always locally bound, setting it globally is no longer
1692    meaningful or necessary (but it does not hurt either).  In fact, "lib"
1693    can still be used as a global anchor for separate purposes.  As a
1694    matter of fact, one can locally define "lib" in terms of a global
1695    "lib":
1696    
1697      a.cm (bind:(anchor:lib value:$lib/a))
1698      b.cm (bind:(anchor:lib value:$lib/b))
1699    
1700    ** 3: The encoding of path names has changed.  This affects the way
1701    path names are shown in CM's progress report and also the internal
1702    protocol encoding used for parallel make.
1703    
1704    The encoding now uses one or more ':'-separated segments.  Each
1705    segments corresponds to a file that has been specified relative to the
1706    file given by its preceding segment.  The first segment is either
1707    relative to the CWD, absolute, or anchored.  Each segment itself is
1708    basically a Unix pathname; all segments but the first are relative.
1709    
1710    Example:
1711    
1712       $foo/bar/baz.cm:a/b/c.sml
1713    
1714    This path denotes the file bar/a/b/c.sml relative to the directory
1715    denoted by anchor "foo".  Notice that the encoding also includes
1716    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1717    paths are resolved relative to the .cm-files directory, so baz.cm must
1718    be ignored to get the "real" pathname.
1719    
1720    To make this fact more obvious, CM puts the names of such "virtual
1721    arcs" into parentheses when they appear in progress reports. (No
1722    parentheses will appear in the internal protocol encoding.)  Thus,
1723    what you really see is:
1724    
1725      $foo/bar/(baz.cm):a/b/c.sml
1726    
1727    I find this notation to be much more informative than before.
1728    
1729    Another new feature of the encoding is that special characters
1730    including parentheses, colons, (back)slashes, and white space are
1731    written as \ddd (where ddd is the decimal encoding of the character).
1732    
1733    *** The CM manual still needs to be updated.
1734    
1735    ----------------------------------------------------------------------
1736    Name: Allen Leung
1737    Date: 2000/06/15 00:38:00
1738    Tag: leunga-20000615-x86-peephole
1739    
1740    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1741    
1742    ----------------------------------------------------------------------
1743    Name: Matthias Blume
1744    Date: 2000/06/12 11:40:00
1745    Tag: blume-20000612-parmakefix
1746    Description:
1747    
1748    More cleanup after changing the file naming scheme: This time I
1749    repaired the parallel make mechanism for CMB.make which I broke earlier.
1750    
1751    ----------------------------------------------------------------------
1752    Name: Allen Leung
1753    Date: 2000/06/09 01:25:00
1754    Tag: leunga-20000609-various
1755    
1756    None of these things should affect normal SML/NJ operations
1757    
1758    1. Peephole improvements provided by Fermin (c--)
1759    2. New annotation DEFUSE for adding extra dependence (moby)
1760    3. New X86 LOCK instructions (moby)
1761    4. New machine description language for reservation tables (scheduling)
1762    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1763       trees etc.)
1764    6. I've changed the CM files so that they can work with versions
1765       110.0.6, 110.25 and 110.28
1766    
1767    ----------------------------------------------------------------------
1768    Name: Matthias Blume
1769    Date: 2000/06/09 12:40:00
1770    Tag: blume-20000609-log
1771    Description:
1772    
1773    - Removed all(?) remaining RCS Log entries from sources.
1774    
1775    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1776      anchored paths).
1777    
1778    ----------------------------------------------------------------------
1779    Name: Matthias Blume
1780    Date: 2000/06/07 17:00:00 JST
1781    Tag: blume-20000607-no-implicit-anchors
1782    Description:
1783    
1784    1. This update changes the default setting for
1785    CM.Control.implicit_anchors from true to false.  This means that
1786    implicit anchors are no longer permitted by default.  I also tried to
1787    make sure that nothing else still relies on implicit anchors.
1788    (This is the next step on the schedule towards a CM that does not even
1789    have the notion of implicit anchors anymore.)
1790    
1791    2. More CM manual updates.
1792    
1793    3. I managed to track down and fix the pickling bug I mentioned last
1794    time.  Because of the previously existing workaround, this entails no
1795    immediate practical changes.
1796    
1797    ----------------------------------------------------------------------
1798    Name: Matthias Blume
1799    Date: 2000/06/06 11:15:00 JST
1800    Tag: blume-20000606-lazierpickle
1801    Description:
1802    
1803    !!!! NEW BOOT FILES !!!!
1804    
1805    * The main purpose of this update is to make library pickles lazier in
1806    order to reduce the initial space penalty for autoloading a library.
1807    As a result, it is now possible to have $smlnj/compiler.cm
1808    pre-registered.  This should take care of the many complaints or
1809    inquiries about missing structure Compiler.  This required changes to
1810    CM's internal data structures and small tweaks to some algorithms.
1811    
1812    As a neat additional effect, it is no longer necessary (for the sake
1813    of lean heap image files) to distinguish between a "minimal" CM and a
1814    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1815    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1816    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1817    
1818    To make the life of compiler-hackers easier, "makeml" now also
1819    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1820    words, after you bootstrap a new sml for the first time, you will not
1821    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1822    around you will still have to do it, though.)
1823    
1824    * A second change consists of major updates to the CM manual.  There
1825    are now several appendices with summary information and also a full
1826    specification of the CM description file syntax.
1827    
1828    * In directory src/system I added the script "allcross".  This script
1829    invokes sml and cross-compiles the compiler for all supported
1830    architectures.  (Useful when providing a new set of boot files.)
1831    
1832    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1833    added a small tweak to pickle-util.sml to work around this problem,
1834    but it is not a proper fix yet.  I will investigate further.  (The
1835    effect of the bug was an inflation of library pickle size.)
1836    
1837    * Version number increased to 110.28.1 (to avoid compatibility problems).
1838    
1839    ----------------------------------------------------------------------
1840    Name: Allen Leung
1841    Date: 2000/05/25 17:28 EDT
1842    Tag: leunga-20000525-ra
1843    Description:
1844    
1845      Fixed a bug in freezing phase of the register allocator.
1846    
1847    ----------------------------------------------------------------------
1848    Name: Allen Leung
1849    Date: 2000/05/15 22:53 EDT
1850    Tag: leunga-20000515-alpha-x86-ra
1851    Description:
1852    
1853      1. Alpha
1854    
1855          Slight cleanup.  Removed the instruction SGNXL
1856    
1857      2. X86
1858    
1859          Added the following instructions to the instruction set:
1860    
1861            ROLx, RORx,
1862            BTx, BTSx, BTLx, BTRx,
1863            XCHGx, and variants with the LOCK prefix
1864    
1865      3. Register Allocation
1866    
1867          The module ra-rewrite-with-renaming has been improved.
1868    
1869      These have no effect on SML/NJ.
1870    
1871    ----------------------------------------------------------------------
1872    Name: Matthias Blume
1873    Date: 2000/05/15 16:20:00 JST
1874    Tag: blume-20000515-lightrebuild
1875    Description:
1876    
1877    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1878       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1879       defined.  In effect, the command will not build any cross-compiler
1880       backends and therefore finish more quickly.
1881    
1882       The "fixpt" script also takes a "-light" switch to be able to use
1883       this new facility while compiling for a fixpoint.
1884    
1885    2. I replaced all mentions of anchored paths in group owner specifications
1886       with simple relative paths (usually starting with "..").
1887       The rationale is that a library's internal workings should not be
1888       compromised by the lack of some anchor.  (An anchor is necessary
1889       for someone who wants to refer to the library by an anchored path,
1890       but it should not be necessary to build the same library in the first
1891       place.)
1892    
1893    3. I changed the way CM's tool mechanism determines the shell command
1894       string used for things like ml-yacc etc. so that it does not break
1895       when CM.Control.implicit_anchors is turned off.
1896    
1897    ----------------------------------------------------------------------
1898    Name: Matthias Blume
1899    Date: 2000/05/12 18:20:00 JST
1900    Tag: blume-20000512-ml-build
1901    Description:
1902    
1903    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1904    getting installed properly (by config/install.sh).
1905    
1906    ----------------------------------------------------------------------
1907    Name: Matthias Blume
1908    Date: 2000/05/12 17:30:00 JST
1909    Tag: blume-20000512-anchors
1910    Description:
1911    
1912    !!! NEW BOOT FILES !!!
1913    
1914    This change is in preparation of fading out support for "implicitly
1915    anchored path names".  I went through all sources and used the
1916    explicit (and relatively new) $-notation.  See system/README and the
1917    CM manual for more info on this.
1918    
1919    I also modified the anchoring scheme for some things such as "smlnj",
1920    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1921    anchors are more expressive: anchor name and first arc do not have to
1922    coincide.  This entails the following user-visible change:
1923    
1924    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1925    particular, when you fire up sml with a command-line argument, say,
1926    e.g.:
1927    
1928       sml '$smlnj/cmb.cm'
1929    
1930    At the ML toplevel prompt:
1931    
1932       CM.autoload "$smlnj/cmb.cm";
1933    
1934    There is also a new controller in CM.Control that can be used to turn
1935    off all remaining support for implicit anchors by saying:
1936    
1937        CM.autoload "$smlnj/
1938        #set CM.Control.implicit_anchors false;
1939    
1940    This causes CM to reject implicitly anchored paths.  This is (for the
1941    time being) less permissive than the "final" version where there will
1942    be no more such implicit anchors and relative paths will be just that:
1943    relative.
1944    
1945    The next step (version after next version?) will be to make the
1946    default for CM.Control.implicit_anchors false.  After the dust has
1947    settled, I can then produce the "final" version of this...
1948    
1949    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1950    
1951    ----------------------------------------------------------------------
1952    Name: Matthias Blume
1953    Date: 2000/05/11 16:30:00 JST
1954    Tag: blume-20000511-sources
1955    Description:
1956    
1957    The main change is that I added function CM.sources as a generalized
1958    version of the earlier CM.makedepend.  This entails the following
1959    additional changes:
1960    
1961      - CM.makedepend has been dropped.
1962    
1963      - CM manual has been updated.
1964    
1965      - TOOLS signature and API have been changed.
1966    
1967    ----------------------------------------------------------------------
1968    Name: Allen Leung
1969    Date: 2000/05/10 21:17 EDT
1970    Tag: leunga-20000510-moby-c--ssa
1971    Description:
1972    
1973      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1974    None of these affect SML/NJ.
1975    
1976    1. Register Allocation
1977    
1978        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1979           This module tries to remove local (i.e. basic block level) redundancies
1980           during spilling.
1981    
1982        b. A new framework for performing region based register allocation.
1983           Not yet entirely functional.
1984    
1985    2. X86
1986    
1987       a. DefUse for POP was missing the stack pointer [found by Lal]
1988       b. Reload for CALL was incorrect in X86Spill [found by John]
1989       c. Various fixes in X86Spill so that it can be used correctly for
1990          the new spilling module.
1991    
1992    3. SSA/IR
1993    
1994       a. New module ir/dj-dataflow.sml implements elimination based
1995          data flow analysis.
1996    
1997    4. MLRiscGen
1998    
1999       a. Fix for gc type annotation
2000    
2001    5. MDGen
2002    
2003       Various fixes for machine description -> ml code translation.  For ssa
2004       only.
2005    
2006    ----------------------------------------------------------------------
2007    Name: Allen Leung
2008    Date: 2000/05/08 22:17 EDT
2009    Tag: leunga-20000508-labexp
2010    Description:
2011    
2012      Fermin has found a few assembly problems with constant expressions
2013      generated in LabelExp.  Mostly, the problems involve extra parentheses,
2014      which choke on dumb assemblers.  This is his fix.
2015    
2016    ----------------------------------------------------------------------
2017    Name: Dave MacQueen
2018    Date: 2000/04/09 14:00 EDT
2019    Tag: dbm-20000502-Version_110_28
2020    Description:
2021    
2022    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
2023    
2024    2. Updated config/version to 110.28
2025    
2026    3. Updated config/srcarchiveurl
2027    
2028    3. New boot files!
2029       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
2030    
2031    ----------------------------------------------------------------------
2032    Name: Matthias Blume
2033    Date: 2000/05/01 19:05:00 JST
2034    Tag: blume-20000501-noweb
2035    Description:
2036    
2037    A new noweb tool has been added.  The existing system is entirely
2038    unaffected by this, but some CM users have asked for renewed noweb
2039    support.  Everything is documented in the CM manual.
2040    
2041    New (plugin) libraries:
2042    
2043       noweb-tool.cm
2044       nw-ext.cm
2045    
2046    ----------------------------------------------------------------------
2047    Name: Dave MacQueen
2048    Date: 2000/04/30 12:40PM EDT
2049    Tag: dbm-20000430-bug_fixes
2050    Description:
2051    
2052    1. Fix for bug 1498
2053       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
2054       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
2055         added toRealArray function
2056       smlnj/src/compiler/MiscUtil/print/ppobj.sml
2057         added check for tag Obj.RealArray to array printing case in ppObj
2058    
2059    2. Fix for bug 1510
2060       smlnj/src/compiler/Semant/types/typesutil.sml
2061         fixed definition of dummyargs (used by equalTycon) so that
2062         dummy args are distinct types
2063    
2064    ----------------------------------------------------------------------
2065    Name: Matthias Blume
2066    Date: 2000/04/30 01:00:00 JST
2067    Tag: blume-20000430-versions
2068    Description:
2069    
2070    1. CM version numbering added.  This is an implementation of Lal's
2071       proposal for adding version numbers and version checking to .cm
2072       files.  Lal said that his proposal was just that -- a proposal.
2073       For the time being I went ahead and implemented it so that people
2074       can comment on it.  Everything is completely backward-compatible
2075       (except for the stable library format, i.e., new bootfiles!).
2076    
2077       As usual, see the CM manual for details.
2078    
2079    2. An alternative syntax for anchored paths has been implemented.
2080       Dave has recently voiced the same concerns that I had when I did
2081       this, so there should be some support.  My take is that eventually
2082       I will let support for the current syntax (where anchors are
2083       "implicit") fade out in favor of the new, explicit syntax.
2084       In order to be backward-compatible, both old and new syntax are
2085       currently supported.
2086    
2087       Again, see the CM manual for details.
2088    
2089    3. Parallel make is trying to be slightly smarter:  When the master
2090       process finds a "bottleneck", i.e., when there is only one
2091       compilation unit that can be compiled and everybody else is
2092       waiting on it, then it will simply compile it directly instead
2093       of clumsily telling one of the slaves to do it.
2094    
2095    4. Support for "unsharing" added.  This is necessary in order to be
2096       able to have two different versions of the same library running
2097       at the same time (e.g., for trying out a new MLRISC while still
2098       having the old MLRISC linked into the current compiler, etc.)
2099       See the CM manual.
2100    
2101    5. Simple "makedepend" functionality added for generating Makefile
2102       dependency information.  (This is rather crude at the moment.
2103       Expect some changes here in the future.)
2104    
2105    6. ".fun" added as a recognized suffix for ML files. Also documented
2106       explicitly in the manual that the fallback behavior (unknown suffix
2107       -> ML file) is not an official feature!
2108    
2109    7. Small changes to the pickler for stable libraries.
2110    
2111    8. Several internal changes to CM (for cleanup/improvement).
2112    
2113    
2114    !!!! NEW BINFILES !!!!
2115    
2116    ----------------------------------------------------------------------
2117    Name: Matthias Blume
2118    Date: 2000/04/28 17:30:00 JST
2119    Tag: blume-20000428-pathconfig
2120    Description:
2121    
2122    1. I changed config/install.sh to remove duplicate entries from the
2123       lib/pathconfig file at the end.  Moreover, the final version of
2124       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
2125       in src/system/installml.
2126    
2127    2. The config/install.sh script now consistently uses relative
2128       pathnames in lib/pathconfig whenever the anchor is in the lib
2129       directory.  (So far this was true for the libraries that come
2130       pre-compiled and bundled as part of the bootfiles but not for
2131       libraries that are compiled by the script itself.)
2132    
2133    ----------------------------------------------------------------------
2134    Name: Matthias Blume
2135    Date: 2000/04/26 13:10:00 JST
2136    Tag: blume-20000426-fun_suffix
2137    Description:
2138    
2139    Added ".fun" as a recognized file name suffix (for ML code).
2140    
2141    ----------------------------------------------------------------------
2142    Name: Allen Leung
2143    Date: 2000/04/25 17:00:00 EST
2144    Tag: leunga-20000425-alpha-ra
2145    Description:
2146    
2147    1. Alpha
2148    
2149        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
2150    in C--.
2151    
2152    2. RA
2153    
2154       Added a flag "ra-dump-size" to print out the size of the flowgraph
2155       and the interference graph.
2156    
2157    ----------------------------------------------------------------------
2158    Name: Dave MacQueen
2159    Date: 2000/04/25/
2160    Tag: dbm-20000425-mlyacc_doc_examples
2161    Description:
2162      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
2163      Updated all three examples in src/ml-yacc/examples to run
2164      under 110.* using CM.make.
2165    
2166    ----------------------------------------------------------------------
2167    Name: Allen Leung
2168    Date: 2000/04/20 23:04:00 EST
2169    Tag: leunga-20000420-ssa-c---stuff
2170    Description:
2171    
2172      This update synchronizes my repository with Yale's.  Most of these
2173    changes, however, do not affect SML/NJ at all (the RA is an exception).
2174    
2175    1. Register Allocator
2176    
2177       a. An improvement in the interference graph construction:
2178          Given a copy
2179    
2180                s <- t
2181    
2182          no interference edge between s and t is added for this definition of s.
2183    
2184       b. I've added two new spill heuristic modules that Fermin and I developed
2185          (in the new library RA.cm). These are unused in SML/NJ but maybe
2186          useful for others (Moby?)
2187    
2188    2. X86
2189    
2190       a. Various fixes in the backend provided by Fermin [C--] and Lal.
2191    
2192    3. Alpha
2193    
2194       a. Added the BSR instruction and code generation that goes with it [C--]
2195       b. Other fixes too numerous to recount provided by Fermin [C--]
2196    
2197    4. Regmaps
2198    
2199       a. The regmaps are not initialized with the identity physical bindings
2200          at creation time.  This is unneeded.
2201    
2202    5. MLRISC Optimizations
2203    
2204       a. The DJ-Graph module can now compute the iterated dominance frontiers
2205          intersects with liveness incrementally in linear time! Woohoo!
2206          This is now used in my new SSA construction algorithm.
2207    
2208       b. THe branch reorganization module is now smarter about linear chains of
2209          basic blocks.
2210    
2211    
2212    ----------------------------------------------------------------------
2213    Name: Matthias Blume
2214    Date: 2000/04/12 13:52:00 JST
2215    Tag: blume_main_v110p27_1
2216    Description:
2217    
2218    Changed install.sh script to handle archive files without version number
2219    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
2220    name of the boot file archive.
2221    
2222    ----------------------------------------------------------------------
2223    Name: Dave MacQueen
2224    Date: 2000/04/09 14:00 EDT
2225    Tag: dbm-20000410-Version_110_27
2226    Description:
2227    
2228    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
2229    
2230    2. Updated src/config/version to 110.27
2231    
2232    3. New boot files!
2233    
2234    ----------------------------------------------------------------------
2235    Name: Allen Leung
2236    Date: 2000/04/09 19:09:00 EST
2237    Tag: leunga-20000409-misc
2238    Description:
2239    
2240    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
2241    
2242    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
2243    
2244    ----------------------------------------------------------------------
2245    Name: Stefan
2246    Date: 2000/04/07 10:00:00 EDT
2247    Tag: monnier-20000406-branch-handling
2248    Description:
2249    
2250    Improved handling of branches (mostly those generated from
2251    polymorphic equality), removed switchoff and changed the
2252    default optimization settings (more cpsopt and less flintopt).
2253    
2254    ----------------------------------------------------------------------
2255    Name: Allen Leung
2256    Date: 2000/04/06 01:30:00 EST
2257    Tag: leunga-20000406-peephole-x86-SSA-2
2258    Description:
2259    
2260       Forgot a few files.
2261    
2262    ----------------------------------------------------------------------
2263    Name: Allen Leung
2264    Date: 2000/04/06 00:36:00 EST
2265    Tag: leunga-20000406-peephole-x86-SSA
2266    Description:
2267    
2268    1.  New Peephole code
2269    
2270    2.  Minor improvement to X86 instruction selection
2271    
2272    3.  Various fixes to SSA and machine description -> code translator
2273    
2274    ----------------------------------------------------------------------
2275    Name: Matthias Blume
2276    Date: 2000/04/05 12:30:00 JST
2277    Tag: blume_main_v110p26p2_3
2278    Description:
2279    
2280    This update just merges three minor cosmetic updates to CM's sources
2281    to get ready for the 110.27 code freeze on Friday.  No functionality
2282    has changed.
2283    
2284    ----------------------------------------------------------------------
2285    Name: Allen Leung
2286    Date: 2000/04/04 19:39:00 EST
2287    Tag: leunga-20000404-x86-asm
2288    Description:
2289    
2290    1.  Fixed a problem in X86 assembly.
2291    
2292        Things like
2293    
2294           jmp %eax
2295           jmp (%eax)
2296    
2297        should be output as
2298    
2299           jmp *%eax
2300           jmp *(%eax)
2301    
2302    2.  Assembly output
2303    
2304          Added a new flag
2305    
2306              "asm-indent-copies" (default to false)
2307    
2308          When this flag is on, parallel copies will be indented an extra level.
2309    
2310    ----------------------------------------------------------------------
2311    Name: Allen Leung
2312    Date: 2000/04/04 03:18:00 EST
2313    Tag: leunga-20000404-C--Moby
2314    Description:
2315    
2316        All of these fixes are related to C--, Moby, and my own optimization
2317        stuff; so they shouldn't affect SML/NJ.
2318    
2319    1.  X86
2320    
2321        Various fixes related floating point, and extensions.
2322    
2323    2.  Alpha
2324    
2325        Some extra patterns related to loads with signed/zero extension
2326        provided by Fermin.
2327    
2328    3.  Assembly
2329    
2330        When generating assembly, resolve the value of client defined constants,
2331        instead of generating symbolic values.  This is controlled by the
2332        new flag "asm-resolve-constants", which is default to true.
2333    
2334    4.  Machine Descriptions
2335    
2336        a. The precedence parser was slightly broken when parsing infixr symbols.
2337        b. The type generalizing code had the bound variables reversed, resulting
2338           in a problem during arity raising.
2339        c. Various fixes in machine descriptions.
2340    
2341    ----------------------------------------------------------------------
2342    Name: Matthias Blume
2343    Date: 2000/04/03 16:05:00 JST
2344    Tag: blume_main_v110p26p2_2
2345    Description:
2346    
2347    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
2348    now done via the ordinary static environment that is context to each
2349    compilation unit.
2350    
2351    To this end, I arranged that instead of "structure Core" as "structure
2352    _Core" is bound in the pervasive environment.  Core access is done via
2353    _Core (which can never be accidentally rebound because _Core is not a
2354    legal surface-syntax symbol).
2355    
2356    The current solution is much cleaner because the core environment is
2357    now simply part of the pervasive environment which is part of every
2358    compilation unit's context anyway.  In particular, this eliminates all
2359    special-case handling that was necessary until now in order to deal
2360    with dynamic and symbolic parts of the core environment.
2361    
2362    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2363    compilation manager's bootstrap compiler (actually: in the "init group"
2364    handling).  See the comments in src/system/smlnj/init/init.cmi for
2365    more details.
2366    
2367    I also tried to track down all mentions of "Core" (as string argument
2368    to Symbol.strSymbol) in the compiler and replaced them with a
2369    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
2370    name appears in one place only.
2371    
2372    Binfile and bootfile format have not changed, but the switchover from
2373    the old "init.cmi" to the new one is a bit tricky, so I supplied new
2374    bootfiles anyway.
2375    
2376    ----------------------------------------------------------------------
2377    Name: Allen Leung
2378    Date: 2000/04/02 21:17:00 EST
2379    Tag: leunga-20000402-mltree
2380    Description:
2381    
2382       1. Renamed the constructor CALL in MLTREE by popular demand.
2383       2. Added a bunch of files from my repository.  These are currently
2384          used by other non-SMLNJ backends.
2385    
2386    ----------------------------------------------------------------------
2387    Name: Allen Leung
2388    Date: 2000/03/31 21:15:00 EST
2389    Tag: leunga-20000331-aliasing
2390    Description:
2391    
2392    This update contains a rewritten (and hopefully more correct) module
2393    for extracting aliasing information from CPS.
2394    
2395       To turn on this feature:
2396    
2397            Compiler.Control.CG.memDisambiguate := true
2398    
2399       To pretty print the region information with assembly
2400    
2401           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
2402    
2403       To control how many levels of aliasing information are printed, use:
2404    
2405           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
2406    
2407       The default of n is 3.
2408    
2409    ----------------------------------------------------------------------
2410  Name: David MacQueen  Name: David MacQueen
2411  Date: 2000/03/31 11:15:00 EST  Date: 2000/03/31 11:15:00 EST
2412  Tag: dbm-20000331-runtime_fix  Tag: dbm-20000331-runtime_fix
# Line 81  Line 2475 
2475     (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
2476     used).     used).
2477    
2478     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
2479     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
2480     Makefiles or Shell scripts.     Makefiles or Shell scripts.
2481    
2482     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 94  Line 2488 
2488    
2489     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
2490    
2491     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
2492     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
2493    
2494  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 228  Line 2622 
2622    
2623       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
2624       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
2625       testl is perferable to andl.       testl is preferable to andl.
2626    
2627  7. RA (x86 only)  7. RA (x86 only)
2628    
# Line 410  Line 2804 
2804    
2805  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2806  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2807  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
2808  explanation.  explanation.
2809    
2810  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2811  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2812  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2813  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 457  Line 2851 
2851  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
2852  <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,
2853  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
2854  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2855  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2856  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2857    
# Line 507  Line 2901 
2901    it from that remote directory.    it from that remote directory.
2902    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2903    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2904    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2905    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2906    
2907    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.589  
changed lines
  Added in v.853

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