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 642, Thu May 11 07:30:29 2000 UTC revision 846, Mon Jun 25 16:06:46 2001 UTC
# Line 11  Line 11 
11  Date:  Date:
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15    ----------------------------------------------------------------------
16    Name: Matthias Blume
17    Date: 2001/06/25 12:05:00 EDT
18    Tag: blume-20010625-nlffigen
19    Description:
20    
21    This update fixes a number of minor bugs in ml-nlffigen as reported by
22    Nick Carter <nbc@andrew.cmu.edu>.
23    
24      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
25      2. Default names for generated files are now derived from the name of
26         the C file *without its directory*.  In particular, this causes generated
27         files to be placed locally even if the C file is in some system directory.
28      3. Default names for generated signatures and structures are also derived
29         from the C file name without its directory.  This avoids silly things
30         like "structure GL/GL".
31         (Other silly names are still possible because ml-nlffigen does not do
32          a thorough check of whether generated names are legal ML identifiers.)
33    
34    ----------------------------------------------------------------------
35    Name: Matthias Blume
36    Date: 2001/06/21 12:25:00 EDT
37    Tag: blume-20010621-eXene
38    Description:
39    
40    eXene now compiles and (sort of) works again.
41    
42    The library name (for version > 110.33) is $/eXene.cm.
43    
44    I also added an new example in src/eXene/examples/nbody.  See the
45    README file there for details.
46    
47    ----------------------------------------------------------------------
48    Name: Matthias Blume
49    Date: 2001/06/20 16:40:00 EDT
50    Tag: blume-20010620-cml
51    Description:
52    
53    CML now compiles and works again.
54    
55    Libraries (for version > 110.33):
56    
57      $cml/cml.cm            Main CML library.
58      $cml/basis.cm          CML's version of $/basis.cm.
59      $cml/cml-internal.cm   Internal helper library.
60      $cml/core-cml.cm       Internal helper library.
61      $cml-lib/trace-cml.cm  Tracing facility.
62      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
63    
64    The installer (config/install.sh) has been taught how to properly
65    install this stuff.
66    
67    ----------------------------------------------------------------------
68    Name: Matthias Blume
69    Date: 2001/06/19 17:55:00 EDT
70    Tag: blume-20010619-instantiate
71    Description:
72    
73    This un-breaks the fix for bug 1432.
74    (The bug was originally fixed in 110.9 but I broke it again some
75    time after that.)
76    
77    ----------------------------------------------------------------------
78    Name: Matthias Blume
79    Date: 2001/06/19 17:25:00 EDT
80    Tag: blume-20010619-signals
81    Description:
82    
83    This should (hopefully) fix the long-standing signal handling bug.
84    (The runtime system was constructing a continuation record with an
85    incorrect descriptor which would cause the GC to drop data on the floor...)
86    
87    ----------------------------------------------------------------------
88    Name: Matthias Blume
89    Date: 2001/06/15 15:05:00 EDT
90    Tag: blume-20010615-moresparc
91    Description:
92    
93    Here is a short late-hour update related to Sparc c-calls:
94    
95     -- made handling of double-word arguments a bit smarter
96    
97     -- instruction selection phase tries to collapse certain clumsily
98        constructed ML-Trees; typical example:
99    
100            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
101    
102        This currently has no further impact on SML/NJ since mlriscGen does
103        not seem to generate such patterns in the first place, and c-calls
104        (which did generate them in the beginning) has meanwhile been fixed
105        so as to avoid them as well.
106    
107    ----------------------------------------------------------------------
108    Name: Matthias Blume
109    Date: 2001/06/15 15:05:00 EDT
110    Tag: blume-20010615-sparc
111    Description:
112    
113    The purpose of this update is to provide an implementation of NLFFI
114    on Sparc machines.
115    
116    Here are the changes in detail:
117    
118       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
119       the Sparc implementation of the c-calls API.
120       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
121       for accessing the ML frame.  Thus, we have a real frame pointer and
122       can freely modify %sp without need for an omit-frame-ptr phase.
123       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
124       this case.
125       * ml-nlffigen has been taught to produce code for different architectures
126       and calling conventions.
127       * In a way similar to what was done in the x86 case, the Sparc
128       backend uses its own specific extension to mltree.  (For example,
129       it needs to be able to generate UNIMP instructions which are part
130       of the calling convention.)
131       * ml-nlffi-lib was reorganized to make it more modular (in particular,
132       to make it easier to plug in new machine- and os-dependent parts).
133    
134    There are some other fairly unrelated bug fixes and cleanups as well:
135    
136       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
137       that they properly share their libraries with existing SML/NJ libraries.
138       * I fixed a minor cosmetic bug in CM, supressing certain spurious
139       follow-up error messages.
140       * Updates to CM/CMB documentation.
141    
142    TODO items:
143    
144       * MLRISC should use a different register as its asmTemp on the Sparc.
145         (The current %o2 is a really bad choice because it is part of the
146         calling conventions, so things might interfere in unexpected ways.)
147    
148    ----------------------------------------------------------------------
149    Name: Matthias Blume
150    Date: 2001/06/07
151    Tag: blume-20010607-calls
152    Description:
153    
154    A number of internal changes related to C calls and calling conventions:
155    
156    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
157       number of bytes popped implicitly (by the callee).  In most cases
158       this field is 0 but on x86/win32 it is some non-zero value.  This
159       is information provided for the benefit of the "omit-frameptr" pass.
160    2. The CALL instruction on the x86 carries a similar "pops" field.
161       The instruction selection phase copies its value from the ML-Tree
162       CALL statement.
163    3. On all other architectures, the instruction selection phase checks
164       whether "pops=0" and complains if not.
165    4. The c-calls implementation for x86 now accepts two calling conventions:
166       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
167       up after the call and pops is set to 0.  For "stdcall", the caller
168       does nothing, leaving the cleanup to the callee; pops is set to
169       the number of bytes that were pushed onto the stack.
170    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
171       distinguish between "ccall" and "stdcall".
172    6. The UNIMP instruction has been added to the supported Sparc instruction
173       set. (This is needed for implementing the official C calling convention
174       on this architecture.)
175    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
176       work with the latest CM.
177    
178    ----------------------------------------------------------------------
179    Name: Matthias Blume
180    Date: 2001/06/05 15:10:00 EDT
181    Tag: blume-20010605-cm-index
182    Description:
183    
184    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
185    
186    1. CM can now generate "index files".  These are human-readable files
187       that list on a per-.cm-file basis each toplevel symbol defined or
188       imported.  The location of the index file for
189       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
190       To enable index-file generation, set CM.Control.generate_index to true
191       or export an environment-symbol: export CM_GENERATE_INDEX=true.
192    
193       The CM manual has been updated accordingly.
194    
195    2. I made some slight modifications to the c-calls API in MLRISC.
196    
197         a) There is now a callback to support saving/restoring of
198            dedicated but caller-save registers around the actual call
199            instruction.
200         b) One can optionally specify a comment-annotation for the
201            call instruction.
202    
203    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
204       (For example, the comment annotation shows the C prototype of
205        the function being called.)
206    
207    ----------------------------------------------------------------------
208    Name: Matthias Blume
209    Date: 2001/06/01 13:30:00 EDT
210    Tag: blume-20010601-nlffi-cleanup
211    Description:
212    
213    This is mostly a cleanup of MLFFI stuff:
214    
215       - some signature files have been put into a more exposed place
216       - the ugly 'f type parameter is gone (simplifies types tremendously!)
217       - ml-nlffigen changed accordingly
218       - tutorial updated
219    
220    Other changes:
221    
222       - author's affiliation in CM manual(s) updated
223       - some more recognized keywords added to Allen's sml.sty
224    
225    ----------------------------------------------------------------------
226    Name: Matthias Blume
227    Date: 2001/05/25 15:30:00 EDT
228    Tag: blume-20010525-iptr
229    Description:
230    
231      - put the official 110.33-README (as it appears on the ftp server) under
232        CVS
233      - fixed a small bug related to incomplete pointer types in
234        ml-nlffigen
235      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
236        (it does not need the 'f type parameter)
237    
238    ----------------------------------------------------------------------
239    Name: Matthias Blume
240    Date: 2001/05/23 14:30:00 EDT
241    Tag: Release_110_33
242    Description:
243    
244    New version number (110.33).  New bootfiles.
245    
246    ----------------------------------------------------------------------
247    Name: Matthias Blume
248    Date: 2001/05/22 18:06:00 EDT
249    Tag: blume-20010522-targets
250    Description:
251    
252    Made install.sh use file config/targets.customized if it exists, falling
253    back to config/targets if it doesn't.  This way one can have a customized
254    version of the targets file without touching the "real thing", thus
255    eliminating the constant fear of accidentally checking something bogus
256    back into the CVS repository...  (File config/targets.customized must
257    not be added to the repository!)
258    
259    ----------------------------------------------------------------------
260    Name: Matthias Blume
261    Date: 2001/05/22 16:30:00 EDT
262    Tag: blume-20010522-minitut
263    Description:
264    
265    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
266       struct returns.
267    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
268       incomplete, preliminary documentation for NLFFI.
269    
270    ----------------------------------------------------------------------
271    Name: Matthias Blume
272    Date: 2001/05/14 11:30:00 EDT
273    Tag: blume-20010514-script
274    Description:
275    
276    Some bugs in install script fixed.
277    
278    In addition to that I also made a slight change to the NLFFI API:
279    Functors generated by ml-nlffigen now take the dynamic library as a
280    straight functor argument, not as a suspended one.  (The original
281    functor code used to force the suspension right away anyway, so there
282    was nothing gained by this complication of the interface.)
283    
284    ----------------------------------------------------------------------
285    Name: Matthias Blume
286    Date: 2001/05/11 14:35:00 EDT
287    Tag: blume-20010511-ml-nlffi
288    Description:
289    
290    I finally took the plunge and added my new FFI code to the main
291    repository.  For x86-linux it is now ready for prime-time.
292    
293    There are two new subdirectories of "src":
294    
295      - ml-nlffi-lib:
296           The utility library for programs using the FFI interface.
297           Here is the implementation of $/c.cm and its associated low-level
298           partners $/c-int.cm and $/memory.cm.
299      - ml-nlffigen:
300           A stand-alone program for generating ML glue code from C source
301           code.
302    
303    Building ml-nlffigen requires $/ckit-lib.cm.
304    
305    The config/install.sh script has been updates to do the Right Thing
306    (hopefully).
307    
308    Notice that the source tree for the C-Kit will not be put under "src"
309    but directly under the installation root directory.  (This is the
310    structure that currently exists on the CVS server when you check out
311    module "sml".)  Fortunately, config/install.sh knows about this oddity.
312    
313    Bugs: No documentation yet.
314    
315    ----------------------------------------------------------------------
316    Name: Matthias Blume
317    Date: 2001/05/09 16:35:00 EDT
318    Tag: blume-20010509-cpscontract
319    Description:
320    
321    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
322    wrapper/unwrapper elimination did not decrement usage counts and some
323    dead variables got overlooked by the dead-up logic.)
324    
325    ----------------------------------------------------------------------
326    Name: Lal George
327    Date: 2001/05/08  17:26:09 EDT
328    Tag: george-20010508-omit-frameptr
329    Description:
330    
331    Changes to implement the omit-frame-pointer optimization to support
332    raw C calls. For now, there is only support on the Intel x86, but
333    other architectures will follow as more experience is gained with this.
334    
335    
336    ----------------------------------------------------------------------
337    Name: Matthias Blume
338    Date: 2001/05/07 14:40:00 EDT
339    Tag: blume-20010507-proxies
340    Description:
341    
342    I made into "proxy libraries" all libraries that qualify for such a
343    change.  (A qualifying library is a library that has another library or
344    groups as its sole member and repeats that member's export list
345    verbatim.  A proxy library avoids this repetition by omitting its export
346    list, effectively inheriting the list that its (only) member exports.
347    See the CM manual for more explanation.)
348    The main effect is that explicit export lists for these libraries
349    do not have to be kepts in sync, making maintenance a bit easier.
350    
351    I also added copyright notices to many .cm-files.
352    
353    Last but not least, I made a new set of bootfiles.
354    
355    ----------------------------------------------------------------------
356    Name: Matthias Blume
357    Date: 2001/05/04 17:00:00 EDT
358    Tag: blume-20010504-cm-lsplit
359    Description:
360    
361    0. John merged pending changes to $/smlnj-lib.cm
362    
363    1. Allen's previous change accidentally backed out of one of Lal's
364       earlier changes.  I undid this mistake (re-introducing Lal's change).
365    
366    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
367       within the compiler where applicable.  There is some code simplification
368       because of that.
369    
370    3. The "split" phase (in FLINT) is now part of the default list of phases.
371       Compiler.Control.LambdaSplitting.* can be used to globally control the
372       lambda-splitting (cross-module-inlining) engine.  In addition to that,
373       it can now also be controlled on a per-source basis: CM has been taught
374       a new tool parameter applicable to ML source files.
375    
376       - To turn lambda-splitting off completely:
377            local open Compiler.Control.LambdaSplitting in
378                val _ = set Off
379            end
380       - To make "no lambda-splitting" the global default (but allow per-source
381         overriding); this is the initial setting:
382            local open Compiler.Control.LambdaSplitting in
383                val _ = set (Default NONE)
384            end
385       - To make "lambda-splitting with aggressiveness a" the global default
386         (and allow per-source overriding):
387            local open Compiler.Control.LambdaSplitting in
388                val _ = set (Default (SOME a))
389            end
390    
391       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
392         write (in the respective .cm-file):
393            a.sml (lambdasplitting:off)
394       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
395            a.sml (lambdasplitting:on)
396       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
397         <a> is a decimal non-negative integer):
398            a.sml (lambdasplitting:<a>)
399       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
400            a.sml (lambdasplitting:infinity)
401       - To use the global default for a.sml:
402            a.sml (lambdasplitting:default)
403         or simply
404            a.sml
405    
406    ----------------------------------------------------------------------
407    Name: Allen Leung
408    Date: 2001/05/04 01:57:00 EDT
409    Tag: leunga-20010504-sync
410    Description:
411    
412      MLRISC features.
413    
414      1. Fix to CMPXCHG instructions.
415      2. Changed RA interface to allow annotations in callbacks.
416      3. Added a new method to the stream interface to allow annotations updates.
417    
418    ----------------------------------------------------------------------
419    Name: Matthias Blume
420    Date: 2001/05/01 11:45:00 EDT
421    Tag: blume-20010501-pcedittmp
422    Description:
423    
424    Changed install.sh to use the current working directory instead of
425    /usr/tmp for a temporary file (pcedittmp).  The previous choice
426    of /usr/tmp caused trouble with MacOS X because of file premission
427    problems.
428    
429    ----------------------------------------------------------------------
430    Name: Matthias Blume
431    Date: 2001/04/20 11:10:00 EDT
432    Tag: blume-20010420-inMLflag
433    Description:
434    
435     - added vp_limitPtrMask to vproc-state.h
436       (for use by the raw-C-calls mechanism to implement proper interrupt
437        handling)
438     - made the ML compiler aware of various data-structure offsets so it
439       can generate code for accessing the vp_inML flag and vp_limitPtrMask
440     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
441       raw C-calls
442    
443    ----------------------------------------------------------------------
444    Name: Lal George
445    Date: 2001/04/20 09:15:28 EDT
446    Tag: george-20010420-macosX
447    Description:
448    
449     - Changes to port to Mac OS X; Darwin.
450    
451     - In the process I found that sqrt was broken on the PPC, because the
452       fsqrt instruction is not implemented.
453    
454    ----------------------------------------------------------------------
455    Name: Matthias Blume
456    Date: 2001/04/18 12:45:00 EDT
457    Tag: blume-20010418-ccalls
458    Description:
459    
460     - fixed two off-by-4 errors in the x86-specific c-calls implementation
461       (this bug prevented structure arguments containing pointers from being
462        passed correctly)
463     - changed the raw-C-call code in mlriscGen.sml in such a way that
464       structure arguments are represented as a pointer to the beginning
465       of the structure (instead of having a series of synthesized arguments,
466       one for each structure member)
467    
468     - made makeml script's verbosity level configurable via environment
469       variable (MAKEML_VERBOSITY)
470    
471     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
472       in rawmem-x86.sml; we are now using the real thing
473    
474    ----------------------------------------------------------------------
475    Name: Matthias Blume
476    Date: 2001/03/22 16:25:00 EST
477    Tag: blume-20010322-bootfiles
478    Description:
479    
480    Created a new set of bootfiles (for your automatic installation convenience).
481    
482    ----------------------------------------------------------------------
483    Name: Matthias Blume
484    Date: 2001/03/22 15:10:00 EST
485    Tag: blume-20010322-rawmem-parcm
486    Description:
487    
488    1. All "raw memory access" primitives for the new FFI are implemented now
489       (at least on the x86).
490    2. Some further cleanup of CM's parallel make mechanism.
491    
492    ----------------------------------------------------------------------
493    Name: Matthias Blume
494    Date: 2001/03/19 17:53:00 EST
495    Tag: blume-20010319-parallel
496    Description:
497    
498    Parallel make (using compile servers) now works again.
499    
500    To this end, CM.stabilize and CMB.make have been modified to work in
501    two passes when compile servers are attached:
502       1. Compile everything, do not perform stabilization; this pass
503          uses compile servers
504       2. Stabilize everything; this pass does not use compile servers
505    If there are no compile servers, the two passes are combined into one
506    (as before).  Splitting the passes increases the inherent parallelism
507    in the dependency graph because the entire graph including all
508    libraries is available at the same time.  This, in turn, improves
509    server utilization.  The downside is that the master process will
510    have to do some extra work after compilation is done (because for
511    technical reasons it must re-read all the binfiles during stabilization).
512    
513    ----------------------------------------------------------------------
514    Name: Matthias Blume
515    Date: 2001/03/16 12:22:00 EST
516    Tag: blume-20010316-bootfiles
517    Description:
518    
519    Created a new set of bootfiles (for your automatic installation convenience).
520    
521    ----------------------------------------------------------------------
522    Name: Matthias Blume
523    Date: 2001/03/16 11:00:00 EST
524    Tag: blume-20010316-MLTREE-fixup
525    Description:
526    
527    This is a minor fixup for an (untagged) earlier commit by Allen.
528    (A file was missing).
529    
530    ----------------------------------------------------------------------
531    Name: Allen Leung
532    Date: Mon Mar  5 18:54:57 EST 2001
533    Tag: leunga-20010305-cut-support
534    
535    1. New support for alternative control-flow in MLTREE.
536       Currently we support
537    
538          FLOW_TO(CALL ...., [k1,...,kn])
539    
540       This is needed for 'cuts to' in C-- and try/handle-like constructs
541       in Moby
542    
543       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
544    
545    2. Register Allocator
546    
547       Changes in interface [from Fermin, John]
548    
549    3. Alpha 8-bit SLL support [Fermin]
550    
551    4. All architectures
552    
553       A new module (ClusterExpandCopies) for expanding parallel copies.
554    
555    ----------------------------------------------------------------------
556    Name: Allen Leung
557    Date: 2001/02/27 23:07:00 EST
558    Tag: leunga-20010227-minor-stuff
559    
560    1. Alpha bug fix for CMOVNE
561    2. Handle mltree COND(..,FCMP ...,...)
562    3. Bug fix in simplifier
563    
564    ----------------------------------------------------------------------
565    Name: Matthias Blume
566    Date: 2001/01/30 17:50:00 EST
567    Tag: blume-20010130-sync
568    Description:
569    
570    This is just a minor update to sync my devel branch with the main brach.
571    The only visible change is the addition of some README files.
572    
573    ----------------------------------------------------------------------
574    Name: Matthias Blume
575    Date: 2001/01/12 23:30:00 JST
576    Tag: blume-20010112-bootfiles
577    Description:
578    
579    Made a new set of bootfiles that goes with the current state of the
580    repository.
581    
582    ----------------------------------------------------------------------
583    Name: Matthias Blume
584    Date: 2001/01/12 21:20:00 JST
585    Tag: blume-20010112-sync
586    Description:
587    
588    I am just flushing out some minor changes that had accumulated in
589    my private branch in order to sync with the main tree.  (This is
590    mainly because I had CVS trouble when trying to merge _into_ my
591    private branch.)
592    
593    Most people should be completely unaffected by this.
594    
595    ----------------------------------------------------------------------
596    Name: Allen Leung
597    Date: Thu Jan 11 21:03:00 EST 2001
598    Tag: leunga-20010111-labexp=mltree
599    Description:
600    
601    1.  Removed the type LabelExp and replace it by MLTree.
602    2.  Rewritten mltree-simplify with the pattern matcher tool.
603    3.  There were some bugs in alpha code generator which would break
604        64-bit code generation.
605    4.  Redo the tools to generate code with the
606    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
607        are now generated by perl scripts.
608    
609    ----------------------------------------------------------------------
610    Name: Matthias Blume
611    Date: 2001/01/10 21:55:00 JST
612    Tag: blume-20010110-rcc
613    Description:
614    
615    The RCC stuff now seems to work (but only on the x86).
616    This required hacking of the c-calls interface (and -implementation) in
617    MLRISC.
618    
619    Normal compiler users should be unaffected.
620    
621    ----------------------------------------------------------------------
622    Name: Matthias Blume
623    Date: 2001/01/09 01:20:00 JST
624    Tag: blume-20010109-rcc
625    Description:
626    
627    This is a fairly big patch, flushing out a large number of pending
628    changes that I made to my development copy over the last couple of days.
629    
630    Of practical relevance at this moment is a workaround for a pickling
631    bug that Allen ran into the other day.  The cause of the bug itself is
632    still unknown and it might be hard to fix it properly, but the
633    workaround has some merits of its own (namely somewhat reducing pickling
634    overhead for certain libraries).  Therefore, I think this solution should
635    be satisfactory at this time.
636    
637    The rest of the changes (i.e., the vast majority) has to do with my
638    ongoing efforts of providing direct support for C function calls from
639    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
640    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
641    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
642    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
643    handle RCC.
644    
645    The last part (the changes to mlriscGen.sml) are still known to be
646    wrong on the x86 and not implemented on all other architectures.  But
647    the infrastructure is in place. I had to change a few functor
648    signatures in the backend to be able to route the CCalls interface
649    from MLRISC there, and I had to specialize the mltree type (on the
650    x86) to include the necessary extensions. (The extensions themselves
651    were already there and redy to go in MLRISC/x86).
652    
653    Everything should be very happy as soon as someone helps me with
654    mlriscGen.sml...
655    
656    In any case, nothing of this should matter to anyone as long as the
657    new primop is not being used (which is going to be the case unless you
658    find it where I hid it :). The rest of the compiler is completely
659    unaffected.
660    
661    ----------------------------------------------------------------------
662    Name: Matthias Blume
663    Date: 2001/01/05 00:30:00 JST
664    Tag: blume-20010105-primops
665    Description:
666    
667    Added some experimental support for work that I am doing right now.
668    These changes mostly concern added primops, but there is also a new
669    experimental C library in the runtime system (but currently not enabled
670    anywhere except on Linux/X86).
671    
672    In the course of adding primops (and playing with them), I discovered that
673    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
674    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
675    To recover, I made type information in INL_PRIM mandatory and changed
676    prim.sml as well as built-in.sml accordingly.  The InLine structure now
677    has complete, correct type information (i.e., no bottom types).
678    
679    Since all these changes mean that we need new binfiles, I also bumped the
680    version number to 110.32.1.
681    
682    ----------------------------------------------------------------------
683    Name: Matthias Blume
684    Date: 2000/12/30 22:10:00 JST
685    Tag: blume-20001230-various
686    Description:
687    
688    Added proxy libraries for MLRISC and let MLRISC libraries refer
689    to each other using path anchors.  (See CM manual for explanation.)
690    
691    Updated CM documentation.
692    
693    Fixed some bugs in CM.
694    
695    Implemented "proxy" libraries (= syntactic sugar for CM).
696    
697    Added "-quiet" option to makeml and changed runtime system accordingly.
698    
699    Added cleanup handler for exportML to reset timers and compiler stats.
700    
701    ----------------------------------------------------------------------
702    Name: Lal George
703    Date: 2000/12/22 22:22:58 EST 2000
704    Tag: Release_110_32
705    Description:
706    
707            Infinite precision used throughout MLRISC.
708            see MLRISC/mltree/machine-int.sig
709    
710    ----------------------------------------------------------------------
711    Name: Matthias Blume
712    Date: 2000/12/22 23:16:00 JST
713    Tag: blume-20001222-warn
714    Description:
715    
716    Corrected wording and formatting of some CM warning message which I
717    broke in my previous patch.
718    
719    ----------------------------------------------------------------------
720    Name: Matthias Blume
721    Date: 2000/12/22 21:20:00 JST
722    Tag: blume-20001222-anchorenv
723    Description:
724    
725    Fixed CM's handling of anchor environments in connection with CMB.make.
726    
727    ----------------------------------------------------------------------
728    Name: Matthias Blume
729    Date: 2000/12/22 13:15:00 JST
730    Tag: blume-20001222-cleanup
731    Description:
732    
733    Removed src/cm/ffi which does not (and did not) belong here.
734    
735    ----------------------------------------------------------------------
736    Name: Matthias Blume
737    Date: 2000/12/21 23:55:00 JST
738    Tag: blume-20001221-exn
739    Description:
740    
741    Probably most important: CM no longer silently swallows all exceptions
742    in the compiler.
743    Plus: some other minor CM changes.  For example, CM now reports some
744    sizes for generated binfiles (code, data, envpickle, lambdapickle).
745    
746    ----------------------------------------------------------------------
747    Name: Matthias Blume
748    Date: 2000/12/15 00:01:05 JST
749    Tag: blume-20001215-dirtool
750    Description:
751    
752    - "dir" tool added.
753    - improvements and cleanup to Tools structure
754    - documentation updates
755    
756    ----------------------------------------------------------------------
757    Name: Allen Leung
758    Date: Thu Dec 14 03:45:24 EST 2000
759    Description:
760    Tag:  leunga-20001214-int-inf
761    Description:
762    
763       In IntInf, added these standard functions, which are missing from our
764    implementation:
765    
766        andb : int * int -> int
767        xorb : int * int -> int
768        orb  : int * int -> int
769        notb : int -> int
770         <<   : int * word -> int
771        ~>>  : int * word -> int
772    
773       Not tested, I hope they are correct.
774    
775    ----------------------------------------------------------------------
776    Name: Allen Leung
777    Date: Fri Dec  8 19:23:26 EST 2000
778    Description:
779    Tag:  leunga-20001208-nowhere
780    Description:
781    
782      Slight improvements to the 'nowhere' tool to handle OR-patterns,
783    to generate better error messages etc.  Plus a brief manual.
784    
785    ----------------------------------------------------------------------
786    Name: Lal George
787    Date: 2000/12/08 09:54:02 EST 2000
788    Tag: Release_110_31
789    Description:
790    
791    - Version 110.31
792    ----------------------------------------------------------------------
793    Name: Allen Leung
794    Date: Thu Dec  7 22:01:04 EST 2000
795    Tag:  leunga-20001207-cell-monster-hack
796    Description:
797    
798    Major MLRISC internal changes.  Affect all clients.
799    Summary:
800    
801    1.  Type CELLS.cell = int is now replaced by a datatype.
802        As a result, the old regmap is now gone.  Almost all interfaces
803        in MLRISC change as a consequence.
804    
805    2.  A new brand version of machine description tool (v3.0) that generates
806        modules expecting the new interface.  The old version is removed.
807    
808    3.  The RA interface has been further abstracted into two new functors.
809        RISC_RA and X86RA.  These functors have much simpler interfaces.
810        [See also directory MLRISC/demo.]
811    
812    4.  Some other new source->source code generation tools are available:
813    
814        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
815        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
816           I use this tool to generate the peephole optimizers---with the new
817           cell type changes, peephole rules are becoming difficult to write
818           without conditional pattern matching.
819    
820    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
821        cover the entire MLRISC source tree so many things broke.
822    
823    6.  CM files have been moved to the subdirectory MLRISC/cm.
824        They are moved because there are a lot of them and they clutter up the
825        root dir.
826    
827    7.  More detailed documentation to come...
828    
829        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
830        a makeml -rebuild first.  This is because of other other
831        changes that Matthias has made (see below).
832    
833    
834    ----------------------------------------------------------------------
835    Name: Matthias Blume
836    Date: 2000/11/30 23:12:00 JST
837    Tag: blume-20001130-filereorg
838    Description:
839    
840    Some manual updates and some file reorganizations in CM.
841    
842    ----------------------------------------------------------------------
843    Name: Matthias Blume
844    Date: 2000/11/24 17:45:00 JST
845    Tag: blume-20001124-link
846    Description:
847    
848    Drastically improved link traversal code for the case that the dynamic
849    value was already loaded at bootstrap time.  As a result, CM and CMB
850    now both load blazingly fast -- even on a very slow machine.  Also,
851    memory consumption has been further reduced by this.
852    
853    Warning: The format of the PIDMAP file has changed.  THerefore, to
854    bootstrap you have to do this:
855    
856    1. Run CMB.make
857    2. Make a symbolic link for the boot directory:
858         ln -s sml.boot.ARCH-OS xxx
859    3. "Rebuild" the boot directory:
860         ./makeml -boot xxx -rebuild sml ; rm xxx
861    4. Boot normally:
862          ./makeml
863    
864    ----------------------------------------------------------------------
865    Name: Matthias Blume
866    Date: 2000/11/21 21:20:00 JST
867    Tag: blume-20001121-tools
868    Description:
869    
870    Continued hacking on autoloading problem -- with success this time.
871    Also changed tool-plugin mechanism.  See new CM manual.
872    
873    ----------------------------------------------------------------------
874    Name: Matthias Blume
875    Date: 2000/11/19 14:30:00 JST
876    Tag:  blume-20001119-autoload
877    Description:
878    
879    Some hacking to make autoloading faster.  Success for CMB, no success
880    so far for CM.  There is a reduced structure CM' that autoloads faster.
881    (This is a temporary, non-documented hack to be eliminated again when
882    the general problem is solved.)
883    
884    ----------------------------------------------------------------------
885    Name: Matthias Blume
886    Date: 2000/11/17 14:10:00 JST
887    Tag: blume-20001117-pickle-lib
888    Description:
889    
890    1. Eliminated comp-lib.cm
891    2. Made pickle-lib.cm
892    3. Eliminated all uses of intset.sml (from comp-lib.cm)
893    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
894       equivalent constructs from smlnj-lib.cm (INtHashTable).
895    5. Point 4. also goes for those uses of intmap.* in MLRISC.
896       Duplicated intmap modules thrown out.
897    6. Hunted down all duplicated SCC code and replaced it with
898       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
899    7. Rewrote Feedback module.
900    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
901       should be thrown out and equivalent modules from smlnj-lib.cm
902       should be used (IntRedBlackSet, IntListSet, ...).
903    
904    Confirmed that compiler compiles to fixpoint.
905    
906    ----------------------------------------------------------------------
907    Name: Allen Leung
908    Date: 2000/11/10 18:00:00
909    Tag: leunga-20001110-new-x86-fp
910    
911    A new x86 floating point code generator has been added.
912    By default this is turned off.  To turn this on, do:
913    
914        CM.autoload "$smlnj/compiler.cm";
915        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
916    
917    Changes:
918    
919    1.  Changed FTAN to FPTAN so that the assembly output is correct.
920    2.  Changed the extension callback for FTANGENT to generate:
921    
922              fptan
923              fstp  %st(0)
924        instead of
925              fptan
926              fstpl ftempmem
927    
928    3.  Numerous assembly fixes for x86.
929    
930    5.  Cleaned up the machine code output module x86/x86MC.sml and added
931        support for a whole bunch of instructions and addressing modes:
932    
933          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
934          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
935          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
936          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
937          fxch %st(n)
938          fld %st(n)
939          fst %st(n)
940          fst mem
941          fstp %st(n)
942          fucom %st(n)
943          fucomp %st(n)
944    
945        All these are now generated when the fast fp mode is turned on.
946    
947    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
948    
949    ----------------------------------------------------------------------
950    Name: Matthias Blume
951    Date: 2000/11/09 11:20:00 JST
952    Tag: blume-20001109-scc
953    Description:
954    
955    Eliminated some code duplication:
956    
957    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
958       (Otherwise the functor is useless.)
959    2. Used GraphSCCFn where SCCUtilFun was used previously.
960    3. Got rid of SCCUtilFun (in comp-lib.cm).
961    
962    ----------------------------------------------------------------------
963    Name: Lal George
964    Date: 2000/11/06 09:02:21 EST 2000
965    Tag: Release_110_30
966    Description:
967    
968    - Version 110.30
969    ----------------------------------------------------------------------
970    Name: Matthias Blume
971    Date: 2000/11/04 14:45:00
972    Tag: blume-20001104-mlbuild
973    Description:
974    
975    - Made ml-build faster on startup.
976    - Documentation fixes.
977    
978    ----------------------------------------------------------------------
979    Name: Matthias Blume
980    Date: 2000/11/02 17:00:00 JST
981    Tag: blume-20001102-condcomp
982    Description:
983    
984    - Small tweaks to pickler -- new BOOTFILES!
985    - Version bumped to 110.29.2.
986    - Added conditional compilation facility to init.cmi (see comment there).
987    ----------------------------------------------------------------------
988    Name: Allen Leung
989    Date: 2000/10/23 19:31:00
990    Tag: leunga-20001023-demo-ra
991    
992    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
993    2. Test programs for the graph library updated
994    3. Some new MLRISC demo programs added
995    
996    ----------------------------------------------------------------------
997    Name: Matthias Blume
998    Date: 2000/08/31 22:15:00 JST
999    Tag: blume-20001017-errmsg
1000    Description:
1001    
1002    More error message grief: Where there used to be no messages, there
1003    now were some that had bogus error regions.  Fixed.
1004    
1005    ----------------------------------------------------------------------
1006    Name: Matthias Blume
1007    Date: 2000/08/31 17:30:00 JST
1008    Tag: blume-20001017-v110p29p1
1009    Description:
1010    
1011    I made a version 110.29.1 with new bootfiles.
1012    
1013    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
1014              CM documentation changes and a small bugfix in CM's error reporting.
1015    
1016    ----------------------------------------------------------------------
1017    Name: Lal George
1018    Date: 2000/09/27 14:42:35 EDT
1019    Tag: george-20000927-nodestatus
1020    Description:
1021    
1022    Changed the type of the nodestatus, so that:
1023    
1024            SPILLED(~1)             is now SPILLED
1025            SPILLED(m) where m>=0   is now MEMREG(m)
1026            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
1027    
1028    ----------------------------------------------------------------------
1029    Name: Matthias Blume
1030    Date: 2000/09/07 14:45:00 JST
1031    Tag: blume-20000907-cmerrmsg
1032    Description:
1033    
1034    Small tweak to CM to avoid getting ML syntax error messages twice.
1035    
1036    ----------------------------------------------------------------------
1037    Name: Matthias Blume
1038    Date: 2000/08/31 18:00:00 JST
1039    Tag: blume-20000831-cvsbootfiles
1040    Description:
1041    
1042    New URL for boot files (because the 110.29 files on the BL server do
1043    now work correctly with my updated install scripts for yacc and lex).
1044    
1045    ----------------------------------------------------------------------
1046    Name: Matthias Blume
1047    Date: 2000/08/08 12:33:00 JST
1048    Tag: blume-20000808-manual
1049    Description:
1050    
1051    Tiny update to CM manual.
1052    
1053    ----------------------------------------------------------------------
1054    Name: Allen Leung
1055    Date: 2000/08/7 19:31:00
1056    Tag: leunga-20000807-a-whole-bunch-of-stuff
1057    
1058      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
1059    and Mobdy.
1060    
1061    1.  x86
1062    
1063       a.  Fixes to peephole module by John and Dan.
1064       b.  Assembly fix to SETcc by Allen.
1065       c.  Fix to c-call by John.
1066       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
1067       e.  Instruction selection optimization to SETcc as suggested by John.
1068    
1069           For example,
1070    
1071            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
1072    
1073           should generate:
1074    
1075            MOVL a, x
1076            SUBL b, x
1077            SHRL 31, x
1078    
1079    2.  IR stuff
1080    
1081         A bunch of new DJ-graph related algorithms added.  These
1082         speed up SSA construction.
1083    
1084    3.  SSA + Scheduling
1085    
1086         Added code for SSA and scheduling to the repository
1087    
1088    ----------------------------------------------------------------------
1089    Name: Lal George
1090    Date: 2000/07/27 11:53:14 EDT
1091    
1092    Tag: lal-20000727-linux-ppc
1093    Description:
1094    
1095     Made changes to support Linux PPC.
1096     p.s. I have confirmation that the 110.29 boot files work fine.
1097    
1098    ----------------------------------------------------------------------
1099    Name: Matthias Blume
1100    Date: 2000/07/27 17:40:00 JST
1101    Tag: blume-20000727-scripts
1102    Description:
1103    
1104    !!!! WARNING !!!!
1105    You must recompile the runtime system!
1106    !!!! WARNING !!!!
1107    
1108    This is basically another round of script-enhancements:
1109    
1110    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
1111       and undefine CM preprocessor symbols.
1112    
1113    2. ml-build avoids generating a new heap image if it finds that the
1114       existing one is still ok.  (The condition is that no ML file had to
1115       be recompiled and all ML files are found to be older that the heap
1116       file.)
1117    
1118       To make this work smoothly, I also hacked the runtime system as
1119       well as SMLofNJ.SysInfo to get access to the heap image suffix
1120       (.sparc-solaris, ...) that is currently being used.
1121    
1122       Moreover, the signature of CM.mk_standalone has changed.  See the
1123       CM manual.
1124    
1125    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
1126       CM manual for details.)
1127    
1128    4. More CM manual updates:
1129        - all of the above has been documented.
1130        - there is now a section describing the (CM-related) command line
1131          arguments that are accepted by the "sml" command
1132    
1133    ----------------------------------------------------------------------
1134    Name: Matthias Blume
1135    Date: 2000/07/25 16:20:00 JST
1136    Tag: blume-20000725-makedepend
1137    Description:
1138    
1139    Added a script called ml-makedepend.  This can be used in makefiles
1140    for Unix' make in a way very similar to the "makedepend" command for
1141    C.
1142    
1143    The script internally uses function CM.sources.
1144    
1145    Synopsis:
1146    
1147        ml-makedepend [-f makefile] cmfile targetname
1148    
1149    The default for the makefile is "makefile" (or "Makefile" should
1150    "makefile" not exist).
1151    
1152    ml-makedepend adds a cmfile/targetname-specific section to this
1153    makefile (after removing the previous version of this section).  The
1154    section contains a single dependency specification with targetname on
1155    the LHS (targetname is an arbitrary name), and a list of files derived
1156    from the cmfile on the RHS.  Some of the files on the RHS are
1157    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
1158    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
1159    path names.  The makefile writer is responsible for making sure that
1160    these variables have correct at the time "make" is invoked.
1161    
1162    ----------------------------------------------------------------------
1163    Name: Matthias Blume
1164    Date: 2000/07/22 23:30:00 JST
1165    Tag: blume-20000722-urlupdate
1166    Description:
1167    
1168    Changed BOOT and config/srcarchiveurl to point to BL server:
1169    
1170        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
1171    
1172    ----------------------------------------------------------------------
1173    Name: Matthias Blume
1174    Date: 2000/07/18 18:00:00 JST
1175    Tag: blume-20000718-Version_110_29
1176    Description:
1177    
1178    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
1179    
1180    2. Updated config/version to 110.29
1181    
1182    3. Updated config/srcarchiveurl
1183    
1184    3. New boot files!
1185       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
1186    
1187    ----------------------------------------------------------------------
1188    Name: Matthias Blume
1189    Date: 2000/07/11 13:58:00 JST
1190    Tag: blume-20000711-doctypo
1191    Description:
1192    
1193    Fixed a few typos in CM manual.
1194    
1195    ----------------------------------------------------------------------
1196    Name: Allen Leung
1197    Date: 2000/06/15 00:38:00
1198    Tag: leunga-20000704-sparc-x86
1199    
1200    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
1201    2. fix to x86 RET bug [found by Dan Grossman]
1202    3. sparc assembly bug fix for ticc instructions [found by Fermin]
1203    
1204       Affects c-- and moby only
1205    
1206    ----------------------------------------------------------------------
1207    Name: Matthias Blume
1208    Date: 2000/07/04 15:26:00
1209    Tag: blume-20000704-trigger
1210    Description:
1211    
1212    1. Improvements to CM manual.
1213    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
1214       of getting a back-trace.  The function, when called, raises an
1215       internal exception which explicitly carries the full back-trace history,
1216       so it is unaffected by any intervening handle-raise pairs ("trivial"
1217       or not).  The interactive loop will print that history once it arrives
1218       at top level.
1219       Short of having all exceptions implicitly carry the full history, the
1220       recommended way of using this facility is:
1221         - compile your program with instrumentation "on"
1222         - run it, when it raises an exception, look at the history
1223         - if the history is "cut off" because of some handler, go and modify
1224           your program so that it explicitly calls BTrace.trigger
1225         - recompile (still instrumented), and rerun; look at the full history
1226    
1227    ----------------------------------------------------------------------
1228    Name: Matthias Blume
1229    Date: 2000/07/03 15:36:00 JST
1230    Tag: blume-20000702-manual
1231    Description:
1232    
1233    Small corrections and updates to CM manual.
1234    
1235    ----------------------------------------------------------------------
1236    Name: Matthias Blume
1237    Date: 2000/06/29 16:04:00 JST
1238    Tag: blume-20000629-yacctool
1239    Description:
1240    
1241    Changes:
1242    
1243    1. Class "mlyacc" now takes separate arguments to pass options to
1244       generated .sml- and .sig-files independently.
1245    2. Corresponding CM manual updates.
1246    3. BTrace module now also reports call sites.  (However, for loop clusters
1247       it only shows from where the cluster was entered.)  There are associated
1248       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1249    
1250    ----------------------------------------------------------------------
1251    Name: Matthias Blume
1252    Date: 2000/06/27 16:51:00 JST
1253    Tag: blume-20000627-noweb
1254    Description:
1255    
1256    Changes:
1257    
1258     1. Implemented "subdir" and "witness" options for noweb tool.
1259        This caused some slight internal changes in CM's tool implementation.
1260     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
1261        some remaining issues from earlier path anchor changes.
1262     3. Updated CM manual accordingly.
1263    
1264     4. Changed implementation of back-tracing so that I now consider it
1265        ready for prime-time.
1266    
1267        In particular, you don't have to explicitly trigger the back-trace
1268        anymore.  Instead, if you are running BTrace-instrumented code and
1269        there is an uncaught exception (regardless of whether or not it was
1270        raised in instrumented code), the top-level evalloop will print
1271        the back-trace.
1272    
1273        Features:
1274    
1275          - Instrumented and uninstrumented code work together seemlessly.
1276            (Of course, uninstrumented code is never mentioned in actual
1277             back-traces.)
1278    
1279          - Asymptotic time- and space-complexity of instrumented code is
1280            equal to that of uninstrumented code.  (This means that
1281            tail-recursion is preserved by the instrumentation phase.)
1282    
1283          - Modules whose code has been instrumented in different sessions
1284            work together without problem.
1285    
1286          - There is no penalty whatsoever on uninstrumented code.
1287    
1288          - There is no penalty on "raise" expressions, even in
1289            instrumented code.
1290    
1291        A potential bug (or perhaps it is a feature, too):
1292    
1293          A back-trace reaches no further than the outermost instrumented
1294          non-trivial "raise".  Here, a "trivial" raise is one that is the
1295          sole RHS of a "handle" rule.  Thus, back-traces reach trough
1296    
1297               <exp> handle e => raise e
1298    
1299          and even
1300    
1301               <exp> handle Foo => raise Bar
1302    
1303          and, of course, through
1304    
1305               <exp> handle Foo => ...
1306    
1307         if the exception was not Foo.
1308    
1309         Back-traces always reach right through any un-instrumented code
1310         including any of its "handle" expressions, trivial or not.
1311    
1312       To try this out, do the following:
1313    
1314         - Erase all existing binfiles for your program.
1315           (You may keep binfiles for those modules where you think you
1316            definitely don't need back-tracing.)
1317         - Turn on back-trace instrumentation:
1318              SMLofNJ.Internals.BTrace.mode (SOME true);
1319         - Recompile your program.  (I.e., run "CM.make" or "use".)
1320         - You may now turn instrumentation off again (if you want):
1321              SMLofNJ.Internals.BTrace.mode (SOME false);
1322         - Run your program as usual.  If it raises an exception that
1323           reaches the interactive toplevel, then a back-trace will
1324           automatically be printed.  After that, the toplevel loop
1325           will print the exception history as usual.
1326    
1327    ----------------------------------------------------------------------
1328    Name: Matthias Blume
1329    Date: 2000/06/26 09:56:46 JST
1330    Tag: blume-20000626-setup
1331    Description:
1332    
1333    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1334          ML code before and after compiling a file (e.g., to set compiler
1335          flags)
1336    
1337    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1338              - associated changes to btrace.sml (BTrace instrumentation pass)
1339              - cleaner implementation of btimp.sml (BTrace tracing and report
1340                module)
1341    
1342    CM manual: * new path encoding documented
1343               * description of setup-parameter to "sml" added
1344    
1345    The biggest user-visible change to back-tracing is that it is no
1346    longer necessary to compile all traced modules within the same
1347    session.  (This was a real limitation.)
1348    
1349    ----------------------------------------------------------------------
1350    Name: Matthias Blume
1351    Date: 2000/06/24 12:40:00 JST
1352    Tag: blume-20000624-startup
1353    Description:
1354    
1355    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1356    bit too often -- to put it mildly. :)
1357    
1358    ----------------------------------------------------------------------
1359    Name: Matthias Blume
1360    Date: 2000/06/23 18:20:00 JST
1361    Tag: blume-20000623-btrace
1362    Description:
1363    
1364    This updates adds a backtrace facility to aid programmers in debugging
1365    their programs.  This involves the following changes:
1366    
1367    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1368       keeping track of the current call stack.  When programs are compiled
1369       in a special mode, the compiler will insert calls to these hooks
1370       into the user program.
1371       "Hook" means that it is possible for different implementations of
1372       back-tracing to register themselves (at different times).
1373    
1374    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1375       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1376       It can be turned on using this call:
1377         SMLofNJ.Internals.BTrace.mode (SOME true);
1378       Turning it off again:
1379         SMLofNJ.Internals.BTrace.mode (SOME false);
1380       Querying the current status:
1381         SMLofNJ.Internals.BTrace.mode NONE;
1382       Annotated programs are about twice as big as normal ones, and they
1383       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1384       where all hooks do nothing).  The slowdown with a plugin that is
1385       actually useful (such as the one supplied by default) is even greater,
1386       but in the case of the default plugin it is still only an constant
1387       factor (amortized).
1388    
1389    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1390       with a sub-structure BTrace for controlling back-tracing.  In particular,
1391       the above-mentioned function "mode" controls whether the annotation
1392       phase is invoked by the compiler.  Another important function is
1393       "trigger": when called it aborts the current execution and causes
1394       the top-level loop to print a full back-trace.
1395    
1396    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1397       for back-tracing.  It keeps track of the dynamic call stack and in
1398       addition to that it keeps a partial history at each "level" of that
1399       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1400       d and b (at separate times, dynamically), then the report will show:
1401    
1402       GOTO   d
1403             /c
1404       GOTO  \b
1405       CALL   a
1406    
1407       This shows that there was an initial non-tail call of a, then a
1408       tail-call to b or c, looping behavior in a cluster of functions that
1409       consist of b and c, and then a goto from that cluster (i.e., either from
1410       b or from c) to d.
1411    
1412       Note that (depending on the user program) the amount of information
1413       that the back-trace module has to keep track of at each level is bounded
1414       by a constant.  Thus, the whole implementation has the same asymptotical
1415       complexity as the original program (both in space and in time).
1416    
1417    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1418       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1419       which is raised by the "trigger" function mentioned above.
1420    
1421    Notes on usage:
1422    
1423    - Annotated code works well together with unannotated code:
1424    Unannotated calls simply do not show up at all in the backtrace.
1425    
1426    - It is not a good idea to let modules that were annotated during
1427    different sessions run at the same time.  This is because the compiler
1428    chooses small integers to identify individual functions, and there
1429    will be clashes if different modules were compiled in separate sessions.
1430    (Nothing will crash, and you will even be told about the clashes, but
1431    back-trace information will in general not be useful.)
1432    
1433    - Back-tracing can be confused by callcc and capture.
1434    
1435    - The only way of getting a back-trace right now is to explicitly
1436    invoke the "trigger" function from your user program.  Eventually, we
1437    should make every exception carry back-trace information (if
1438    available).  But since this creates more overhead at "raise"-time
1439    (similar to the current exnHistory overhead), I have not yet
1440    implemented this.  (The implementation will be rather easy.)  With
1441    exceptions carrying back-trace information, this facility will be even
1442    more useful because users don't need to modify their programs...
1443    
1444    - While it is possible to compile the compiler with back-trace
1445    annotations turned on (I did it to get some confidence in
1446    correctness), you must make absolutely sure that core.sml and
1447    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1448    be compiled with annotation because there is no core access yet, but
1449    if you compile btimp.sml with annotation, then the system will go into
1450    an infinite recursion and crash.)
1451    Since CM currently does not know about BTrace, the only way to turn
1452    annotations on and off for different modules of the compiler is to
1453    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1454    this is awkward and clumsy.
1455    
1456    Sample sessions:
1457    
1458    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1459    - SMLofNJ.Internals.BTrace.mode (SOME true);
1460    [autoloading]
1461    [autoloading done]
1462    val it = false : bool
1463    - structure X = struct
1464    -     fun main n = let
1465    -         fun a (x, 0) = d x
1466    -           | a (x, n) = b (x, n - 1)
1467    -         and b (x, n) = c (x, n)
1468    -         and c (x, n) = a (x, n)
1469    -         and d x = e (x, 3)
1470    -         and e (x, 0) = f x
1471    -           | e (x, n) = e (x, n - 1)
1472    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1473    -           | f n = n * g (n - 1)
1474    -         and g n = a (n, 3)
1475    -     in
1476    -         f n
1477    -     end
1478    - end;
1479    structure X : sig val main : int -> int end
1480    - X.main 3;
1481    *** BACK-TRACE ***
1482    GOTO   stdIn:4.2-13.20: X.main[2].f
1483    GOTO-( stdIn:4.2-13.20: X.main[2].e
1484    GOTO   stdIn:4.2-13.20: X.main[2].d
1485         / stdIn:4.2-13.20: X.main[2].a
1486         | stdIn:4.2-13.20: X.main[2].b
1487    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1488    CALL   stdIn:4.2-13.20: X.main[2].g
1489    GOTO   stdIn:4.2-13.20: X.main[2].f
1490    GOTO-( stdIn:4.2-13.20: X.main[2].e
1491    GOTO   stdIn:4.2-13.20: X.main[2].d
1492         / stdIn:4.2-13.20: X.main[2].a
1493         | stdIn:4.2-13.20: X.main[2].b
1494    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1495    CALL   stdIn:4.2-13.20: X.main[2].g
1496    GOTO   stdIn:4.2-13.20: X.main[2].f
1497    GOTO-( stdIn:4.2-13.20: X.main[2].e
1498    GOTO   stdIn:4.2-13.20: X.main[2].d
1499         / stdIn:4.2-13.20: X.main[2].a
1500         | stdIn:4.2-13.20: X.main[2].b
1501    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1502    CALL   stdIn:4.2-13.20: X.main[2].g
1503    GOTO   stdIn:4.2-13.20: X.main[2].f
1504    CALL   stdIn:2.15-17.4: X.main[2]
1505    -
1506    
1507    (Note that because of a FLINt bug the above code currently does not
1508    compile without BTrace turned on.)
1509    
1510    Here is another example, using my modified Tiger compiler:
1511    
1512    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1513    - SMLofNJ.Internals.BTrace.mode (SOME true);
1514    [autoloading]
1515    [autoloading done]
1516    val it = false : bool
1517    - CM.make "sources.cm";
1518    [autoloading]
1519    ...
1520    [autoloading done]
1521    [scanning sources.cm]
1522    [parsing (sources.cm):parse.sml]
1523    [creating directory CM/SKEL ...]
1524    [parsing (sources.cm):tiger.lex.sml]
1525    ...
1526    [wrote CM/sparc-unix/semant.sml]
1527    [compiling (sources.cm):main.sml]
1528    [wrote CM/sparc-unix/main.sml]
1529    [New bindings added.]
1530    val it = true : bool
1531    - Main.compile ("../testcases/merge.tig", "foo.out");
1532    *** BACK-TRACE ***
1533    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
1534    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1535    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1536    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1537    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1538    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1539    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
1540         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
1541    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
1542    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1543    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
1544    CALL   main.sml:1.18-118.4: Main.compile[2]
1545    -
1546    
1547    ----------------------------------------------------------------------
1548    Name: Matthias Blumen
1549    Date: 2000/06/21 18:00:00 JST
1550    Tag: blume-20000621-manual
1551    Description:
1552    
1553    CM manual update: Path environments documented.
1554    
1555    ----------------------------------------------------------------------
1556    Name: Matthias Blume
1557    Date: 2000/06/19 13:40:00
1558    Tag: blume-20000619-manual
1559    Description:
1560    
1561    CM manual and system/README update.  This only covers the fact that
1562    there are no more implicit anchors.  (Path environments and the "bind"
1563    option to "cm" have yet to be documented.)
1564    
1565    ----------------------------------------------------------------------
1566    Name: Matthias Blume
1567    Date: 2000/06/19 11:05:00 JST
1568    Tag: blume-20000619-chdir-bugfix
1569    Description:
1570    
1571    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1572    
1573    ----------------------------------------------------------------------
1574    Name: Matthias Blume
1575    Date: 2000/06/18 22:00:10 JST
1576    Tag: blume-20000618-implicit-anchors-really-gone
1577    Description:
1578    
1579    I updates the previous HISTORY entry where I forgot to mention that
1580    implicit anchors are no longer with us.
1581    
1582    The current update also gets rid of the (now useless) controller
1583    CM.Control.implicit_anchors.
1584    
1585    ----------------------------------------------------------------------
1586    Name: Matthias Blume
1587    Date: 2000/06/16 17:30:00 JST
1588    Tag: blume-20000616-anchorenv
1589    Description:
1590    
1591    This patch implements the long anticipated (just kidding :) "anchor
1592    environment" mechanism.  In the course of doing this, I also
1593    re-implemented CM's internal "SrcPath" module from scratch.  The new
1594    one should be more robust in certain boundary cases.  In any case, it
1595    is a lot cleaner than its predecessor (IMHO).
1596    
1597    This time, although there is yet another boot file format change, I
1598    kept the unpickler backward-compatible.  As a result, no new bootfiles
1599    are necessary and bootstrapping is straightforward.  (You cannot read
1600    new bootfiles into an old system, but the other way around is no
1601    problem.)
1602    
1603    Visible changes:
1604    
1605    ** 0. Implicit path anchors (without the leading $-symbol) are no
1606    longer recognized at all. This means that such path names are not
1607    illegal either.  For example, the name basis.cm simply refers to a
1608    local file called "basis.cm" (i.e, the name is an ordinary path
1609    relative to .cm-files directory).  Or, to put it differently, only
1610    names that start with $ are anchored paths.
1611    
1612    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1613    vanished.
1614    
1615    John (Reppy) had critizised this as soon as I originally proposed and
1616    implemented it, but at that time I did not really deeply believe
1617    him. :) Now I came full-circle because I need the $<singlearc> syntax
1618    in another place where it cannot be seen as an abbreviation for
1619    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1620    seems to mean (i.e., it "expands" into the corresponding anchor
1621    value).
1622    
1623    However, when paths are used as members in CM description files, it
1624    continues to be true that there must be at least another arc after the
1625    anchor.  This is now enforced separately during semantic analysis
1626    (i.e., from a lexical/syntactical point of view, the notation is ok.)
1627    
1628    ** 2. The "cm" class now accepts an option "bind".  The option's value
1629    is a sub-option list of precisely two items -- one labeled "anchor"
1630    and the other one labeled "value".  As you might expect, "anchor" is
1631    used to specify an anchor name to be bound, and "value" specifies what
1632    the anchor is being bound to.
1633    
1634    The value must be a directory name and can be given in either standard
1635    syntax (including the possibility that it is itself an anchored path)
1636    or native syntax.
1637    
1638    Examples:
1639    
1640       foo.cm (bind:(anchor:bar value:$mystuff/bar))
1641       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1642    
1643    and so on.
1644    
1645    The meaning of this is that the .cm-file will be processed with an
1646    augmented anchor environment where the given anchor(s) is/are bound to
1647    the given values(s).
1648    
1649    The rationale for having this feature is this: Suppose you are trying
1650    to use two different (already stable) libraries a.cm and b.cm (that
1651    you perhaps didn't write yourself).  Further, suppose each of these
1652    two libraries internally uses its own auxiliary library $aux/lib.cm.
1653    Normally you would now have a problem because the anchor "lib" can not
1654    be bound to more than one value globally.  Therefore, the project that
1655    uses both a.cm and b.cm must locally redirect the anchor to some other
1656    place:
1657    
1658       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1659       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1660    
1661    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1662    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
1663    
1664    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1665    you when you do so at the time of CM.stabilize).  Therefore, you can
1666    also use an anchored path as the value:
1667    
1668      a.cm (bind:(anchor:lib value:$a-lib))
1669      b.cm (bind:(anchor:lib value:$b-lib))
1670    
1671    Now you can globally configure (using the usual CM.Anchor.anchor or
1672    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1673    itself is always locally bound, setting it globally is no longer
1674    meaningful or necessary (but it does not hurt either).  In fact, "lib"
1675    can still be used as a global anchor for separate purposes.  As a
1676    matter of fact, one can locally define "lib" in terms of a global
1677    "lib":
1678    
1679      a.cm (bind:(anchor:lib value:$lib/a))
1680      b.cm (bind:(anchor:lib value:$lib/b))
1681    
1682    ** 3: The encoding of path names has changed.  This affects the way
1683    path names are shown in CM's progress report and also the internal
1684    protocol encoding used for parallel make.
1685    
1686    The encoding now uses one or more ':'-separated segments.  Each
1687    segments corresponds to a file that has been specified relative to the
1688    file given by its preceding segment.  The first segment is either
1689    relative to the CWD, absolute, or anchored.  Each segment itself is
1690    basically a Unix pathname; all segments but the first are relative.
1691    
1692    Example:
1693    
1694       $foo/bar/baz.cm:a/b/c.sml
1695    
1696    This path denotes the file bar/a/b/c.sml relative to the directory
1697    denoted by anchor "foo".  Notice that the encoding also includes
1698    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1699    paths are resolved relative to the .cm-files directory, so baz.cm must
1700    be ignored to get the "real" pathname.
1701    
1702    To make this fact more obvious, CM puts the names of such "virtual
1703    arcs" into parentheses when they appear in progress reports. (No
1704    parentheses will appear in the internal protocol encoding.)  Thus,
1705    what you really see is:
1706    
1707      $foo/bar/(baz.cm):a/b/c.sml
1708    
1709    I find this notation to be much more informative than before.
1710    
1711    Another new feature of the encoding is that special characters
1712    including parentheses, colons, (back)slashes, and white space are
1713    written as \ddd (where ddd is the decimal encoding of the character).
1714    
1715    *** The CM manual still needs to be updated.
1716    
1717    ----------------------------------------------------------------------
1718    Name: Allen Leung
1719    Date: 2000/06/15 00:38:00
1720    Tag: leunga-20000615-x86-peephole
1721    
1722    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1723    
1724    ----------------------------------------------------------------------
1725    Name: Matthias Blume
1726    Date: 2000/06/12 11:40:00
1727    Tag: blume-20000612-parmakefix
1728    Description:
1729    
1730    More cleanup after changing the file naming scheme: This time I
1731    repaired the parallel make mechanism for CMB.make which I broke earlier.
1732    
1733    ----------------------------------------------------------------------
1734    Name: Allen Leung
1735    Date: 2000/06/09 01:25:00
1736    Tag: leunga-20000609-various
1737    
1738    None of these things should affect normal SML/NJ operations
1739    
1740    1. Peephole improvements provided by Fermin (c--)
1741    2. New annotation DEFUSE for adding extra dependence (moby)
1742    3. New X86 LOCK instructions (moby)
1743    4. New machine description language for reservation tables (scheduling)
1744    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1745       trees etc.)
1746    6. I've changed the CM files so that they can work with versions
1747       110.0.6, 110.25 and 110.28
1748    
1749    ----------------------------------------------------------------------
1750    Name: Matthias Blume
1751    Date: 2000/06/09 12:40:00
1752    Tag: blume-20000609-log
1753    Description:
1754    
1755    - Removed all(?) remaining RCS Log entries from sources.
1756    
1757    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1758      anchored paths).
1759    
1760    ----------------------------------------------------------------------
1761    Name: Matthias Blume
1762    Date: 2000/06/07 17:00:00 JST
1763    Tag: blume-20000607-no-implicit-anchors
1764    Description:
1765    
1766    1. This update changes the default setting for
1767    CM.Control.implicit_anchors from true to false.  This means that
1768    implicit anchors are no longer permitted by default.  I also tried to
1769    make sure that nothing else still relies on implicit anchors.
1770    (This is the next step on the schedule towards a CM that does not even
1771    have the notion of implicit anchors anymore.)
1772    
1773    2. More CM manual updates.
1774    
1775    3. I managed to track down and fix the pickling bug I mentioned last
1776    time.  Because of the previously existing workaround, this entails no
1777    immediate practical changes.
1778    
1779    ----------------------------------------------------------------------
1780    Name: Matthias Blume
1781    Date: 2000/06/06 11:15:00 JST
1782    Tag: blume-20000606-lazierpickle
1783    Description:
1784    
1785    !!!! NEW BOOT FILES !!!!
1786    
1787    * The main purpose of this update is to make library pickles lazier in
1788    order to reduce the initial space penalty for autoloading a library.
1789    As a result, it is now possible to have $smlnj/compiler.cm
1790    pre-registered.  This should take care of the many complaints or
1791    inquiries about missing structure Compiler.  This required changes to
1792    CM's internal data structures and small tweaks to some algorithms.
1793    
1794    As a neat additional effect, it is no longer necessary (for the sake
1795    of lean heap image files) to distinguish between a "minimal" CM and a
1796    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1797    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1798    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1799    
1800    To make the life of compiler-hackers easier, "makeml" now also
1801    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1802    words, after you bootstrap a new sml for the first time, you will not
1803    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1804    around you will still have to do it, though.)
1805    
1806    * A second change consists of major updates to the CM manual.  There
1807    are now several appendices with summary information and also a full
1808    specification of the CM description file syntax.
1809    
1810    * In directory src/system I added the script "allcross".  This script
1811    invokes sml and cross-compiles the compiler for all supported
1812    architectures.  (Useful when providing a new set of boot files.)
1813    
1814    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1815    added a small tweak to pickle-util.sml to work around this problem,
1816    but it is not a proper fix yet.  I will investigate further.  (The
1817    effect of the bug was an inflation of library pickle size.)
1818    
1819    * Version number increased to 110.28.1 (to avoid compatibility problems).
1820    
1821    ----------------------------------------------------------------------
1822    Name: Allen Leung
1823    Date: 2000/05/25 17:28 EDT
1824    Tag: leunga-20000525-ra
1825    Description:
1826    
1827      Fixed a bug in freezing phase of the register allocator.
1828    
1829    ----------------------------------------------------------------------
1830    Name: Allen Leung
1831    Date: 2000/05/15 22:53 EDT
1832    Tag: leunga-20000515-alpha-x86-ra
1833    Description:
1834    
1835      1. Alpha
1836    
1837          Slight cleanup.  Removed the instruction SGNXL
1838    
1839      2. X86
1840    
1841          Added the following instructions to the instruction set:
1842    
1843            ROLx, RORx,
1844            BTx, BTSx, BTLx, BTRx,
1845            XCHGx, and variants with the LOCK prefix
1846    
1847      3. Register Allocation
1848    
1849          The module ra-rewrite-with-renaming has been improved.
1850    
1851      These have no effect on SML/NJ.
1852    
1853    ----------------------------------------------------------------------
1854    Name: Matthias Blume
1855    Date: 2000/05/15 16:20:00 JST
1856    Tag: blume-20000515-lightrebuild
1857    Description:
1858    
1859    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1860       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1861       defined.  In effect, the command will not build any cross-compiler
1862       backends and therefore finish more quickly.
1863    
1864       The "fixpt" script also takes a "-light" switch to be able to use
1865       this new facility while compiling for a fixpoint.
1866    
1867    2. I replaced all mentions of anchored paths in group owner specifications
1868       with simple relative paths (usually starting with "..").
1869       The rationale is that a library's internal workings should not be
1870       compromised by the lack of some anchor.  (An anchor is necessary
1871       for someone who wants to refer to the library by an anchored path,
1872       but it should not be necessary to build the same library in the first
1873       place.)
1874    
1875    3. I changed the way CM's tool mechanism determines the shell command
1876       string used for things like ml-yacc etc. so that it does not break
1877       when CM.Control.implicit_anchors is turned off.
1878    
1879    ----------------------------------------------------------------------
1880    Name: Matthias Blume
1881    Date: 2000/05/12 18:20:00 JST
1882    Tag: blume-20000512-ml-build
1883    Description:
1884    
1885    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1886    getting installed properly (by config/install.sh).
1887    
1888    ----------------------------------------------------------------------
1889    Name: Matthias Blume
1890    Date: 2000/05/12 17:30:00 JST
1891    Tag: blume-20000512-anchors
1892    Description:
1893    
1894    !!! NEW BOOT FILES !!!
1895    
1896    This change is in preparation of fading out support for "implicitly
1897    anchored path names".  I went through all sources and used the
1898    explicit (and relatively new) $-notation.  See system/README and the
1899    CM manual for more info on this.
1900    
1901    I also modified the anchoring scheme for some things such as "smlnj",
1902    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1903    anchors are more expressive: anchor name and first arc do not have to
1904    coincide.  This entails the following user-visible change:
1905    
1906    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1907    particular, when you fire up sml with a command-line argument, say,
1908    e.g.:
1909    
1910       sml '$smlnj/cmb.cm'
1911    
1912    At the ML toplevel prompt:
1913    
1914       CM.autoload "$smlnj/cmb.cm";
1915    
1916    There is also a new controller in CM.Control that can be used to turn
1917    off all remaining support for implicit anchors by saying:
1918    
1919        CM.autoload "$smlnj/
1920        #set CM.Control.implicit_anchors false;
1921    
1922    This causes CM to reject implicitly anchored paths.  This is (for the
1923    time being) less permissive than the "final" version where there will
1924    be no more such implicit anchors and relative paths will be just that:
1925    relative.
1926    
1927    The next step (version after next version?) will be to make the
1928    default for CM.Control.implicit_anchors false.  After the dust has
1929    settled, I can then produce the "final" version of this...
1930    
1931    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1932    
1933  ----------------------------------------------------------------------  ----------------------------------------------------------------------
1934  Name: Matthias Blume  Name: Matthias Blume
1935  Date: 2000/05/11 16:30:00 JST  Date: 2000/05/11 16:30:00 JST
# Line 390  Line 2309 
2309    
2310  3.  Assembly  3.  Assembly
2311    
2312      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
2313      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
2314      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
2315    
# Line 413  Line 2332 
2332    
2333  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
2334  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
2335  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
2336  legal surface-syntax symbol).  legal surface-syntax symbol).
2337    
2338  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 423  Line 2342 
2342  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
2343    
2344  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2345  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
2346  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
2347  more details.  more details.
2348    
# Line 538  Line 2457 
2457     (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
2458     used).     used).
2459    
2460     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
2461     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
2462     Makefiles or Shell scripts.     Makefiles or Shell scripts.
2463    
2464     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 551  Line 2470 
2470    
2471     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
2472    
2473     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
2474     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
2475    
2476  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 685  Line 2604 
2604    
2605       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
2606       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
2607       testl is perferable to andl.       testl is preferable to andl.
2608    
2609  7. RA (x86 only)  7. RA (x86 only)
2610    
# Line 867  Line 2786 
2786    
2787  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2788  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2789  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
2790  explanation.  explanation.
2791    
2792  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2793  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2794  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2795  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 914  Line 2833 
2833  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
2834  <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,
2835  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
2836  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2837  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2838  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2839    
# Line 964  Line 2883 
2883    it from that remote directory.    it from that remote directory.
2884    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2885    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2886    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2887    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2888    
2889    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.642  
changed lines
  Added in v.846

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