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 615, Wed Apr 12 04:52:09 2000 UTC revision 844, Wed Jun 20 20:39:15 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: 2001/06/20 16:40:00 EDT
18    Tag: blume-20010620-cml
19    Description:
20    
21    CML now compiles and works again.
22    
23    Libraries (for version > 110.33):
24    
25      $cml/cml.cm            Main CML library.
26      $cml/basis.cm          CML's version of $/basis.cm.
27      $cml/cml-internal.cm   Internal helper library.
28      $cml/core-cml.cm       Internal helper library.
29      $cml-lib/trace-cml.cm  Tracing facility.
30      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
31    
32    The installer (config/install.sh) has been taught how to properly
33    install this stuff.
34    
35    ----------------------------------------------------------------------
36    Name: Matthias Blume
37    Date: 2001/06/19 17:55:00 EDT
38    Tag: blume-20010619-instantiate
39    Description:
40    
41    This un-breaks the fix for bug 1432.
42    (The bug was originally fixed in 110.9 but I broke it again some
43    time after that.)
44    
45    ----------------------------------------------------------------------
46    Name: Matthias Blume
47    Date: 2001/06/19 17:25:00 EDT
48    Tag: blume-20010619-signals
49    Description:
50    
51    This should (hopefully) fix the long-standing signal handling bug.
52    (The runtime system was constructing a continuation record with an
53    incorrect descriptor which would cause the GC to drop data on the floor...)
54    
55    ----------------------------------------------------------------------
56    Name: Matthias Blume
57    Date: 2001/06/15 15:05:00 EDT
58    Tag: blume-20010615-moresparc
59    Description:
60    
61    Here is a short late-hour update related to Sparc c-calls:
62    
63     -- made handling of double-word arguments a bit smarter
64    
65     -- instruction selection phase tries to collapse certain clumsily
66        constructed ML-Trees; typical example:
67    
68            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
69    
70        This currently has no further impact on SML/NJ since mlriscGen does
71        not seem to generate such patterns in the first place, and c-calls
72        (which did generate them in the beginning) has meanwhile been fixed
73        so as to avoid them as well.
74    
75    ----------------------------------------------------------------------
76    Name: Matthias Blume
77    Date: 2001/06/15 15:05:00 EDT
78    Tag: blume-20010615-sparc
79    Description:
80    
81    The purpose of this update is to provide an implementation of NLFFI
82    on Sparc machines.
83    
84    Here are the changes in detail:
85    
86       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
87       the Sparc implementation of the c-calls API.
88       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
89       for accessing the ML frame.  Thus, we have a real frame pointer and
90       can freely modify %sp without need for an omit-frame-ptr phase.
91       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
92       this case.
93       * ml-nlffigen has been taught to produce code for different architectures
94       and calling conventions.
95       * In a way similar to what was done in the x86 case, the Sparc
96       backend uses its own specific extension to mltree.  (For example,
97       it needs to be able to generate UNIMP instructions which are part
98       of the calling convention.)
99       * ml-nlffi-lib was reorganized to make it more modular (in particular,
100       to make it easier to plug in new machine- and os-dependent parts).
101    
102    There are some other fairly unrelated bug fixes and cleanups as well:
103    
104       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
105       that they properly share their libraries with existing SML/NJ libraries.
106       * I fixed a minor cosmetic bug in CM, supressing certain spurious
107       follow-up error messages.
108       * Updates to CM/CMB documentation.
109    
110    TODO items:
111    
112       * MLRISC should use a different register as its asmTemp on the Sparc.
113         (The current %o2 is a really bad choice because it is part of the
114         calling conventions, so things might interfere in unexpected ways.)
115    
116    ----------------------------------------------------------------------
117    Name: Matthias Blume
118    Date: 2001/06/07
119    Tag: blume-20010607-calls
120    Description:
121    
122    A number of internal changes related to C calls and calling conventions:
123    
124    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
125       number of bytes popped implicitly (by the callee).  In most cases
126       this field is 0 but on x86/win32 it is some non-zero value.  This
127       is information provided for the benefit of the "omit-frameptr" pass.
128    2. The CALL instruction on the x86 carries a similar "pops" field.
129       The instruction selection phase copies its value from the ML-Tree
130       CALL statement.
131    3. On all other architectures, the instruction selection phase checks
132       whether "pops=0" and complains if not.
133    4. The c-calls implementation for x86 now accepts two calling conventions:
134       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
135       up after the call and pops is set to 0.  For "stdcall", the caller
136       does nothing, leaving the cleanup to the callee; pops is set to
137       the number of bytes that were pushed onto the stack.
138    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
139       distinguish between "ccall" and "stdcall".
140    6. The UNIMP instruction has been added to the supported Sparc instruction
141       set. (This is needed for implementing the official C calling convention
142       on this architecture.)
143    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
144       work with the latest CM.
145    
146    ----------------------------------------------------------------------
147    Name: Matthias Blume
148    Date: 2001/06/05 15:10:00 EDT
149    Tag: blume-20010605-cm-index
150    Description:
151    
152    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
153    
154    1. CM can now generate "index files".  These are human-readable files
155       that list on a per-.cm-file basis each toplevel symbol defined or
156       imported.  The location of the index file for
157       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
158       To enable index-file generation, set CM.Control.generate_index to true
159       or export an environment-symbol: export CM_GENERATE_INDEX=true.
160    
161       The CM manual has been updated accordingly.
162    
163    2. I made some slight modifications to the c-calls API in MLRISC.
164    
165         a) There is now a callback to support saving/restoring of
166            dedicated but caller-save registers around the actual call
167            instruction.
168         b) One can optionally specify a comment-annotation for the
169            call instruction.
170    
171    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
172       (For example, the comment annotation shows the C prototype of
173        the function being called.)
174    
175    ----------------------------------------------------------------------
176    Name: Matthias Blume
177    Date: 2001/06/01 13:30:00 EDT
178    Tag: blume-20010601-nlffi-cleanup
179    Description:
180    
181    This is mostly a cleanup of MLFFI stuff:
182    
183       - some signature files have been put into a more exposed place
184       - the ugly 'f type parameter is gone (simplifies types tremendously!)
185       - ml-nlffigen changed accordingly
186       - tutorial updated
187    
188    Other changes:
189    
190       - author's affiliation in CM manual(s) updated
191       - some more recognized keywords added to Allen's sml.sty
192    
193    ----------------------------------------------------------------------
194    Name: Matthias Blume
195    Date: 2001/05/25 15:30:00 EDT
196    Tag: blume-20010525-iptr
197    Description:
198    
199      - put the official 110.33-README (as it appears on the ftp server) under
200        CVS
201      - fixed a small bug related to incomplete pointer types in
202        ml-nlffigen
203      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
204        (it does not need the 'f type parameter)
205    
206    ----------------------------------------------------------------------
207    Name: Matthias Blume
208    Date: 2001/05/23 14:30:00 EDT
209    Tag: Release_110_33
210    Description:
211    
212    New version number (110.33).  New bootfiles.
213    
214    ----------------------------------------------------------------------
215    Name: Matthias Blume
216    Date: 2001/05/22 18:06:00 EDT
217    Tag: blume-20010522-targets
218    Description:
219    
220    Made install.sh use file config/targets.customized if it exists, falling
221    back to config/targets if it doesn't.  This way one can have a customized
222    version of the targets file without touching the "real thing", thus
223    eliminating the constant fear of accidentally checking something bogus
224    back into the CVS repository...  (File config/targets.customized must
225    not be added to the repository!)
226    
227    ----------------------------------------------------------------------
228    Name: Matthias Blume
229    Date: 2001/05/22 16:30:00 EDT
230    Tag: blume-20010522-minitut
231    Description:
232    
233    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
234       struct returns.
235    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
236       incomplete, preliminary documentation for NLFFI.
237    
238    ----------------------------------------------------------------------
239    Name: Matthias Blume
240    Date: 2001/05/14 11:30:00 EDT
241    Tag: blume-20010514-script
242    Description:
243    
244    Some bugs in install script fixed.
245    
246    In addition to that I also made a slight change to the NLFFI API:
247    Functors generated by ml-nlffigen now take the dynamic library as a
248    straight functor argument, not as a suspended one.  (The original
249    functor code used to force the suspension right away anyway, so there
250    was nothing gained by this complication of the interface.)
251    
252    ----------------------------------------------------------------------
253    Name: Matthias Blume
254    Date: 2001/05/11 14:35:00 EDT
255    Tag: blume-20010511-ml-nlffi
256    Description:
257    
258    I finally took the plunge and added my new FFI code to the main
259    repository.  For x86-linux it is now ready for prime-time.
260    
261    There are two new subdirectories of "src":
262    
263      - ml-nlffi-lib:
264           The utility library for programs using the FFI interface.
265           Here is the implementation of $/c.cm and its associated low-level
266           partners $/c-int.cm and $/memory.cm.
267      - ml-nlffigen:
268           A stand-alone program for generating ML glue code from C source
269           code.
270    
271    Building ml-nlffigen requires $/ckit-lib.cm.
272    
273    The config/install.sh script has been updates to do the Right Thing
274    (hopefully).
275    
276    Notice that the source tree for the C-Kit will not be put under "src"
277    but directly under the installation root directory.  (This is the
278    structure that currently exists on the CVS server when you check out
279    module "sml".)  Fortunately, config/install.sh knows about this oddity.
280    
281    Bugs: No documentation yet.
282    
283    ----------------------------------------------------------------------
284    Name: Matthias Blume
285    Date: 2001/05/09 16:35:00 EDT
286    Tag: blume-20010509-cpscontract
287    Description:
288    
289    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
290    wrapper/unwrapper elimination did not decrement usage counts and some
291    dead variables got overlooked by the dead-up logic.)
292    
293    ----------------------------------------------------------------------
294    Name: Lal George
295    Date: 2001/05/08  17:26:09 EDT
296    Tag: george-20010508-omit-frameptr
297    Description:
298    
299    Changes to implement the omit-frame-pointer optimization to support
300    raw C calls. For now, there is only support on the Intel x86, but
301    other architectures will follow as more experience is gained with this.
302    
303    
304    ----------------------------------------------------------------------
305    Name: Matthias Blume
306    Date: 2001/05/07 14:40:00 EDT
307    Tag: blume-20010507-proxies
308    Description:
309    
310    I made into "proxy libraries" all libraries that qualify for such a
311    change.  (A qualifying library is a library that has another library or
312    groups as its sole member and repeats that member's export list
313    verbatim.  A proxy library avoids this repetition by omitting its export
314    list, effectively inheriting the list that its (only) member exports.
315    See the CM manual for more explanation.)
316    The main effect is that explicit export lists for these libraries
317    do not have to be kepts in sync, making maintenance a bit easier.
318    
319    I also added copyright notices to many .cm-files.
320    
321    Last but not least, I made a new set of bootfiles.
322    
323    ----------------------------------------------------------------------
324    Name: Matthias Blume
325    Date: 2001/05/04 17:00:00 EDT
326    Tag: blume-20010504-cm-lsplit
327    Description:
328    
329    0. John merged pending changes to $/smlnj-lib.cm
330    
331    1. Allen's previous change accidentally backed out of one of Lal's
332       earlier changes.  I undid this mistake (re-introducing Lal's change).
333    
334    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
335       within the compiler where applicable.  There is some code simplification
336       because of that.
337    
338    3. The "split" phase (in FLINT) is now part of the default list of phases.
339       Compiler.Control.LambdaSplitting.* can be used to globally control the
340       lambda-splitting (cross-module-inlining) engine.  In addition to that,
341       it can now also be controlled on a per-source basis: CM has been taught
342       a new tool parameter applicable to ML source files.
343    
344       - To turn lambda-splitting off completely:
345            local open Compiler.Control.LambdaSplitting in
346                val _ = set Off
347            end
348       - To make "no lambda-splitting" the global default (but allow per-source
349         overriding); this is the initial setting:
350            local open Compiler.Control.LambdaSplitting in
351                val _ = set (Default NONE)
352            end
353       - To make "lambda-splitting with aggressiveness a" the global default
354         (and allow per-source overriding):
355            local open Compiler.Control.LambdaSplitting in
356                val _ = set (Default (SOME a))
357            end
358    
359       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
360         write (in the respective .cm-file):
361            a.sml (lambdasplitting:off)
362       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
363            a.sml (lambdasplitting:on)
364       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
365         <a> is a decimal non-negative integer):
366            a.sml (lambdasplitting:<a>)
367       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
368            a.sml (lambdasplitting:infinity)
369       - To use the global default for a.sml:
370            a.sml (lambdasplitting:default)
371         or simply
372            a.sml
373    
374    ----------------------------------------------------------------------
375    Name: Allen Leung
376    Date: 2001/05/04 01:57:00 EDT
377    Tag: leunga-20010504-sync
378    Description:
379    
380      MLRISC features.
381    
382      1. Fix to CMPXCHG instructions.
383      2. Changed RA interface to allow annotations in callbacks.
384      3. Added a new method to the stream interface to allow annotations updates.
385    
386    ----------------------------------------------------------------------
387    Name: Matthias Blume
388    Date: 2001/05/01 11:45:00 EDT
389    Tag: blume-20010501-pcedittmp
390    Description:
391    
392    Changed install.sh to use the current working directory instead of
393    /usr/tmp for a temporary file (pcedittmp).  The previous choice
394    of /usr/tmp caused trouble with MacOS X because of file premission
395    problems.
396    
397    ----------------------------------------------------------------------
398    Name: Matthias Blume
399    Date: 2001/04/20 11:10:00 EDT
400    Tag: blume-20010420-inMLflag
401    Description:
402    
403     - added vp_limitPtrMask to vproc-state.h
404       (for use by the raw-C-calls mechanism to implement proper interrupt
405        handling)
406     - made the ML compiler aware of various data-structure offsets so it
407       can generate code for accessing the vp_inML flag and vp_limitPtrMask
408     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
409       raw C-calls
410    
411    ----------------------------------------------------------------------
412    Name: Lal George
413    Date: 2001/04/20 09:15:28 EDT
414    Tag: george-20010420-macosX
415    Description:
416    
417     - Changes to port to Mac OS X; Darwin.
418    
419     - In the process I found that sqrt was broken on the PPC, because the
420       fsqrt instruction is not implemented.
421    
422    ----------------------------------------------------------------------
423    Name: Matthias Blume
424    Date: 2001/04/18 12:45:00 EDT
425    Tag: blume-20010418-ccalls
426    Description:
427    
428     - fixed two off-by-4 errors in the x86-specific c-calls implementation
429       (this bug prevented structure arguments containing pointers from being
430        passed correctly)
431     - changed the raw-C-call code in mlriscGen.sml in such a way that
432       structure arguments are represented as a pointer to the beginning
433       of the structure (instead of having a series of synthesized arguments,
434       one for each structure member)
435    
436     - made makeml script's verbosity level configurable via environment
437       variable (MAKEML_VERBOSITY)
438    
439     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
440       in rawmem-x86.sml; we are now using the real thing
441    
442    ----------------------------------------------------------------------
443    Name: Matthias Blume
444    Date: 2001/03/22 16:25:00 EST
445    Tag: blume-20010322-bootfiles
446    Description:
447    
448    Created a new set of bootfiles (for your automatic installation convenience).
449    
450    ----------------------------------------------------------------------
451    Name: Matthias Blume
452    Date: 2001/03/22 15:10:00 EST
453    Tag: blume-20010322-rawmem-parcm
454    Description:
455    
456    1. All "raw memory access" primitives for the new FFI are implemented now
457       (at least on the x86).
458    2. Some further cleanup of CM's parallel make mechanism.
459    
460    ----------------------------------------------------------------------
461    Name: Matthias Blume
462    Date: 2001/03/19 17:53:00 EST
463    Tag: blume-20010319-parallel
464    Description:
465    
466    Parallel make (using compile servers) now works again.
467    
468    To this end, CM.stabilize and CMB.make have been modified to work in
469    two passes when compile servers are attached:
470       1. Compile everything, do not perform stabilization; this pass
471          uses compile servers
472       2. Stabilize everything; this pass does not use compile servers
473    If there are no compile servers, the two passes are combined into one
474    (as before).  Splitting the passes increases the inherent parallelism
475    in the dependency graph because the entire graph including all
476    libraries is available at the same time.  This, in turn, improves
477    server utilization.  The downside is that the master process will
478    have to do some extra work after compilation is done (because for
479    technical reasons it must re-read all the binfiles during stabilization).
480    
481    ----------------------------------------------------------------------
482    Name: Matthias Blume
483    Date: 2001/03/16 12:22:00 EST
484    Tag: blume-20010316-bootfiles
485    Description:
486    
487    Created a new set of bootfiles (for your automatic installation convenience).
488    
489    ----------------------------------------------------------------------
490    Name: Matthias Blume
491    Date: 2001/03/16 11:00:00 EST
492    Tag: blume-20010316-MLTREE-fixup
493    Description:
494    
495    This is a minor fixup for an (untagged) earlier commit by Allen.
496    (A file was missing).
497    
498    ----------------------------------------------------------------------
499    Name: Allen Leung
500    Date: Mon Mar  5 18:54:57 EST 2001
501    Tag: leunga-20010305-cut-support
502    
503    1. New support for alternative control-flow in MLTREE.
504       Currently we support
505    
506          FLOW_TO(CALL ...., [k1,...,kn])
507    
508       This is needed for 'cuts to' in C-- and try/handle-like constructs
509       in Moby
510    
511       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
512    
513    2. Register Allocator
514    
515       Changes in interface [from Fermin, John]
516    
517    3. Alpha 8-bit SLL support [Fermin]
518    
519    4. All architectures
520    
521       A new module (ClusterExpandCopies) for expanding parallel copies.
522    
523    ----------------------------------------------------------------------
524    Name: Allen Leung
525    Date: 2001/02/27 23:07:00 EST
526    Tag: leunga-20010227-minor-stuff
527    
528    1. Alpha bug fix for CMOVNE
529    2. Handle mltree COND(..,FCMP ...,...)
530    3. Bug fix in simplifier
531    
532    ----------------------------------------------------------------------
533    Name: Matthias Blume
534    Date: 2001/01/30 17:50:00 EST
535    Tag: blume-20010130-sync
536    Description:
537    
538    This is just a minor update to sync my devel branch with the main brach.
539    The only visible change is the addition of some README files.
540    
541    ----------------------------------------------------------------------
542    Name: Matthias Blume
543    Date: 2001/01/12 23:30:00 JST
544    Tag: blume-20010112-bootfiles
545    Description:
546    
547    Made a new set of bootfiles that goes with the current state of the
548    repository.
549    
550    ----------------------------------------------------------------------
551    Name: Matthias Blume
552    Date: 2001/01/12 21:20:00 JST
553    Tag: blume-20010112-sync
554    Description:
555    
556    I am just flushing out some minor changes that had accumulated in
557    my private branch in order to sync with the main tree.  (This is
558    mainly because I had CVS trouble when trying to merge _into_ my
559    private branch.)
560    
561    Most people should be completely unaffected by this.
562    
563    ----------------------------------------------------------------------
564    Name: Allen Leung
565    Date: Thu Jan 11 21:03:00 EST 2001
566    Tag: leunga-20010111-labexp=mltree
567    Description:
568    
569    1.  Removed the type LabelExp and replace it by MLTree.
570    2.  Rewritten mltree-simplify with the pattern matcher tool.
571    3.  There were some bugs in alpha code generator which would break
572        64-bit code generation.
573    4.  Redo the tools to generate code with the
574    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
575        are now generated by perl scripts.
576    
577    ----------------------------------------------------------------------
578    Name: Matthias Blume
579    Date: 2001/01/10 21:55:00 JST
580    Tag: blume-20010110-rcc
581    Description:
582    
583    The RCC stuff now seems to work (but only on the x86).
584    This required hacking of the c-calls interface (and -implementation) in
585    MLRISC.
586    
587    Normal compiler users should be unaffected.
588    
589    ----------------------------------------------------------------------
590    Name: Matthias Blume
591    Date: 2001/01/09 01:20:00 JST
592    Tag: blume-20010109-rcc
593    Description:
594    
595    This is a fairly big patch, flushing out a large number of pending
596    changes that I made to my development copy over the last couple of days.
597    
598    Of practical relevance at this moment is a workaround for a pickling
599    bug that Allen ran into the other day.  The cause of the bug itself is
600    still unknown and it might be hard to fix it properly, but the
601    workaround has some merits of its own (namely somewhat reducing pickling
602    overhead for certain libraries).  Therefore, I think this solution should
603    be satisfactory at this time.
604    
605    The rest of the changes (i.e., the vast majority) has to do with my
606    ongoing efforts of providing direct support for C function calls from
607    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
608    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
609    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
610    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
611    handle RCC.
612    
613    The last part (the changes to mlriscGen.sml) are still known to be
614    wrong on the x86 and not implemented on all other architectures.  But
615    the infrastructure is in place. I had to change a few functor
616    signatures in the backend to be able to route the CCalls interface
617    from MLRISC there, and I had to specialize the mltree type (on the
618    x86) to include the necessary extensions. (The extensions themselves
619    were already there and redy to go in MLRISC/x86).
620    
621    Everything should be very happy as soon as someone helps me with
622    mlriscGen.sml...
623    
624    In any case, nothing of this should matter to anyone as long as the
625    new primop is not being used (which is going to be the case unless you
626    find it where I hid it :). The rest of the compiler is completely
627    unaffected.
628    
629    ----------------------------------------------------------------------
630    Name: Matthias Blume
631    Date: 2001/01/05 00:30:00 JST
632    Tag: blume-20010105-primops
633    Description:
634    
635    Added some experimental support for work that I am doing right now.
636    These changes mostly concern added primops, but there is also a new
637    experimental C library in the runtime system (but currently not enabled
638    anywhere except on Linux/X86).
639    
640    In the course of adding primops (and playing with them), I discovered that
641    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
642    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
643    To recover, I made type information in INL_PRIM mandatory and changed
644    prim.sml as well as built-in.sml accordingly.  The InLine structure now
645    has complete, correct type information (i.e., no bottom types).
646    
647    Since all these changes mean that we need new binfiles, I also bumped the
648    version number to 110.32.1.
649    
650    ----------------------------------------------------------------------
651    Name: Matthias Blume
652    Date: 2000/12/30 22:10:00 JST
653    Tag: blume-20001230-various
654    Description:
655    
656    Added proxy libraries for MLRISC and let MLRISC libraries refer
657    to each other using path anchors.  (See CM manual for explanation.)
658    
659    Updated CM documentation.
660    
661    Fixed some bugs in CM.
662    
663    Implemented "proxy" libraries (= syntactic sugar for CM).
664    
665    Added "-quiet" option to makeml and changed runtime system accordingly.
666    
667    Added cleanup handler for exportML to reset timers and compiler stats.
668    
669    ----------------------------------------------------------------------
670    Name: Lal George
671    Date: 2000/12/22 22:22:58 EST 2000
672    Tag: Release_110_32
673    Description:
674    
675            Infinite precision used throughout MLRISC.
676            see MLRISC/mltree/machine-int.sig
677    
678    ----------------------------------------------------------------------
679    Name: Matthias Blume
680    Date: 2000/12/22 23:16:00 JST
681    Tag: blume-20001222-warn
682    Description:
683    
684    Corrected wording and formatting of some CM warning message which I
685    broke in my previous patch.
686    
687    ----------------------------------------------------------------------
688    Name: Matthias Blume
689    Date: 2000/12/22 21:20:00 JST
690    Tag: blume-20001222-anchorenv
691    Description:
692    
693    Fixed CM's handling of anchor environments in connection with CMB.make.
694    
695    ----------------------------------------------------------------------
696    Name: Matthias Blume
697    Date: 2000/12/22 13:15:00 JST
698    Tag: blume-20001222-cleanup
699    Description:
700    
701    Removed src/cm/ffi which does not (and did not) belong here.
702    
703    ----------------------------------------------------------------------
704    Name: Matthias Blume
705    Date: 2000/12/21 23:55:00 JST
706    Tag: blume-20001221-exn
707    Description:
708    
709    Probably most important: CM no longer silently swallows all exceptions
710    in the compiler.
711    Plus: some other minor CM changes.  For example, CM now reports some
712    sizes for generated binfiles (code, data, envpickle, lambdapickle).
713    
714    ----------------------------------------------------------------------
715    Name: Matthias Blume
716    Date: 2000/12/15 00:01:05 JST
717    Tag: blume-20001215-dirtool
718    Description:
719    
720    - "dir" tool added.
721    - improvements and cleanup to Tools structure
722    - documentation updates
723    
724    ----------------------------------------------------------------------
725    Name: Allen Leung
726    Date: Thu Dec 14 03:45:24 EST 2000
727    Description:
728    Tag:  leunga-20001214-int-inf
729    Description:
730    
731       In IntInf, added these standard functions, which are missing from our
732    implementation:
733    
734        andb : int * int -> int
735        xorb : int * int -> int
736        orb  : int * int -> int
737        notb : int -> int
738         <<   : int * word -> int
739        ~>>  : int * word -> int
740    
741       Not tested, I hope they are correct.
742    
743    ----------------------------------------------------------------------
744    Name: Allen Leung
745    Date: Fri Dec  8 19:23:26 EST 2000
746    Description:
747    Tag:  leunga-20001208-nowhere
748    Description:
749    
750      Slight improvements to the 'nowhere' tool to handle OR-patterns,
751    to generate better error messages etc.  Plus a brief manual.
752    
753    ----------------------------------------------------------------------
754    Name: Lal George
755    Date: 2000/12/08 09:54:02 EST 2000
756    Tag: Release_110_31
757    Description:
758    
759    - Version 110.31
760    ----------------------------------------------------------------------
761    Name: Allen Leung
762    Date: Thu Dec  7 22:01:04 EST 2000
763    Tag:  leunga-20001207-cell-monster-hack
764    Description:
765    
766    Major MLRISC internal changes.  Affect all clients.
767    Summary:
768    
769    1.  Type CELLS.cell = int is now replaced by a datatype.
770        As a result, the old regmap is now gone.  Almost all interfaces
771        in MLRISC change as a consequence.
772    
773    2.  A new brand version of machine description tool (v3.0) that generates
774        modules expecting the new interface.  The old version is removed.
775    
776    3.  The RA interface has been further abstracted into two new functors.
777        RISC_RA and X86RA.  These functors have much simpler interfaces.
778        [See also directory MLRISC/demo.]
779    
780    4.  Some other new source->source code generation tools are available:
781    
782        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
783        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
784           I use this tool to generate the peephole optimizers---with the new
785           cell type changes, peephole rules are becoming difficult to write
786           without conditional pattern matching.
787    
788    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
789        cover the entire MLRISC source tree so many things broke.
790    
791    6.  CM files have been moved to the subdirectory MLRISC/cm.
792        They are moved because there are a lot of them and they clutter up the
793        root dir.
794    
795    7.  More detailed documentation to come...
796    
797        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
798        a makeml -rebuild first.  This is because of other other
799        changes that Matthias has made (see below).
800    
801    
802    ----------------------------------------------------------------------
803    Name: Matthias Blume
804    Date: 2000/11/30 23:12:00 JST
805    Tag: blume-20001130-filereorg
806    Description:
807    
808    Some manual updates and some file reorganizations in CM.
809    
810    ----------------------------------------------------------------------
811    Name: Matthias Blume
812    Date: 2000/11/24 17:45:00 JST
813    Tag: blume-20001124-link
814    Description:
815    
816    Drastically improved link traversal code for the case that the dynamic
817    value was already loaded at bootstrap time.  As a result, CM and CMB
818    now both load blazingly fast -- even on a very slow machine.  Also,
819    memory consumption has been further reduced by this.
820    
821    Warning: The format of the PIDMAP file has changed.  THerefore, to
822    bootstrap you have to do this:
823    
824    1. Run CMB.make
825    2. Make a symbolic link for the boot directory:
826         ln -s sml.boot.ARCH-OS xxx
827    3. "Rebuild" the boot directory:
828         ./makeml -boot xxx -rebuild sml ; rm xxx
829    4. Boot normally:
830          ./makeml
831    
832    ----------------------------------------------------------------------
833    Name: Matthias Blume
834    Date: 2000/11/21 21:20:00 JST
835    Tag: blume-20001121-tools
836    Description:
837    
838    Continued hacking on autoloading problem -- with success this time.
839    Also changed tool-plugin mechanism.  See new CM manual.
840    
841    ----------------------------------------------------------------------
842    Name: Matthias Blume
843    Date: 2000/11/19 14:30:00 JST
844    Tag:  blume-20001119-autoload
845    Description:
846    
847    Some hacking to make autoloading faster.  Success for CMB, no success
848    so far for CM.  There is a reduced structure CM' that autoloads faster.
849    (This is a temporary, non-documented hack to be eliminated again when
850    the general problem is solved.)
851    
852    ----------------------------------------------------------------------
853    Name: Matthias Blume
854    Date: 2000/11/17 14:10:00 JST
855    Tag: blume-20001117-pickle-lib
856    Description:
857    
858    1. Eliminated comp-lib.cm
859    2. Made pickle-lib.cm
860    3. Eliminated all uses of intset.sml (from comp-lib.cm)
861    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
862       equivalent constructs from smlnj-lib.cm (INtHashTable).
863    5. Point 4. also goes for those uses of intmap.* in MLRISC.
864       Duplicated intmap modules thrown out.
865    6. Hunted down all duplicated SCC code and replaced it with
866       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
867    7. Rewrote Feedback module.
868    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
869       should be thrown out and equivalent modules from smlnj-lib.cm
870       should be used (IntRedBlackSet, IntListSet, ...).
871    
872    Confirmed that compiler compiles to fixpoint.
873    
874    ----------------------------------------------------------------------
875    Name: Allen Leung
876    Date: 2000/11/10 18:00:00
877    Tag: leunga-20001110-new-x86-fp
878    
879    A new x86 floating point code generator has been added.
880    By default this is turned off.  To turn this on, do:
881    
882        CM.autoload "$smlnj/compiler.cm";
883        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
884    
885    Changes:
886    
887    1.  Changed FTAN to FPTAN so that the assembly output is correct.
888    2.  Changed the extension callback for FTANGENT to generate:
889    
890              fptan
891              fstp  %st(0)
892        instead of
893              fptan
894              fstpl ftempmem
895    
896    3.  Numerous assembly fixes for x86.
897    
898    5.  Cleaned up the machine code output module x86/x86MC.sml and added
899        support for a whole bunch of instructions and addressing modes:
900    
901          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
902          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
903          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
904          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
905          fxch %st(n)
906          fld %st(n)
907          fst %st(n)
908          fst mem
909          fstp %st(n)
910          fucom %st(n)
911          fucomp %st(n)
912    
913        All these are now generated when the fast fp mode is turned on.
914    
915    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
916    
917    ----------------------------------------------------------------------
918    Name: Matthias Blume
919    Date: 2000/11/09 11:20:00 JST
920    Tag: blume-20001109-scc
921    Description:
922    
923    Eliminated some code duplication:
924    
925    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
926       (Otherwise the functor is useless.)
927    2. Used GraphSCCFn where SCCUtilFun was used previously.
928    3. Got rid of SCCUtilFun (in comp-lib.cm).
929    
930    ----------------------------------------------------------------------
931    Name: Lal George
932    Date: 2000/11/06 09:02:21 EST 2000
933    Tag: Release_110_30
934    Description:
935    
936    - Version 110.30
937    ----------------------------------------------------------------------
938    Name: Matthias Blume
939    Date: 2000/11/04 14:45:00
940    Tag: blume-20001104-mlbuild
941    Description:
942    
943    - Made ml-build faster on startup.
944    - Documentation fixes.
945    
946    ----------------------------------------------------------------------
947    Name: Matthias Blume
948    Date: 2000/11/02 17:00:00 JST
949    Tag: blume-20001102-condcomp
950    Description:
951    
952    - Small tweaks to pickler -- new BOOTFILES!
953    - Version bumped to 110.29.2.
954    - Added conditional compilation facility to init.cmi (see comment there).
955    ----------------------------------------------------------------------
956    Name: Allen Leung
957    Date: 2000/10/23 19:31:00
958    Tag: leunga-20001023-demo-ra
959    
960    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
961    2. Test programs for the graph library updated
962    3. Some new MLRISC demo programs added
963    
964    ----------------------------------------------------------------------
965    Name: Matthias Blume
966    Date: 2000/08/31 22:15:00 JST
967    Tag: blume-20001017-errmsg
968    Description:
969    
970    More error message grief: Where there used to be no messages, there
971    now were some that had bogus error regions.  Fixed.
972    
973    ----------------------------------------------------------------------
974    Name: Matthias Blume
975    Date: 2000/08/31 17:30:00 JST
976    Tag: blume-20001017-v110p29p1
977    Description:
978    
979    I made a version 110.29.1 with new bootfiles.
980    
981    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
982              CM documentation changes and a small bugfix in CM's error reporting.
983    
984    ----------------------------------------------------------------------
985    Name: Lal George
986    Date: 2000/09/27 14:42:35 EDT
987    Tag: george-20000927-nodestatus
988    Description:
989    
990    Changed the type of the nodestatus, so that:
991    
992            SPILLED(~1)             is now SPILLED
993            SPILLED(m) where m>=0   is now MEMREG(m)
994            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
995    
996    ----------------------------------------------------------------------
997    Name: Matthias Blume
998    Date: 2000/09/07 14:45:00 JST
999    Tag: blume-20000907-cmerrmsg
1000    Description:
1001    
1002    Small tweak to CM to avoid getting ML syntax error messages twice.
1003    
1004    ----------------------------------------------------------------------
1005    Name: Matthias Blume
1006    Date: 2000/08/31 18:00:00 JST
1007    Tag: blume-20000831-cvsbootfiles
1008    Description:
1009    
1010    New URL for boot files (because the 110.29 files on the BL server do
1011    now work correctly with my updated install scripts for yacc and lex).
1012    
1013    ----------------------------------------------------------------------
1014    Name: Matthias Blume
1015    Date: 2000/08/08 12:33:00 JST
1016    Tag: blume-20000808-manual
1017    Description:
1018    
1019    Tiny update to CM manual.
1020    
1021    ----------------------------------------------------------------------
1022    Name: Allen Leung
1023    Date: 2000/08/7 19:31:00
1024    Tag: leunga-20000807-a-whole-bunch-of-stuff
1025    
1026      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
1027    and Mobdy.
1028    
1029    1.  x86
1030    
1031       a.  Fixes to peephole module by John and Dan.
1032       b.  Assembly fix to SETcc by Allen.
1033       c.  Fix to c-call by John.
1034       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
1035       e.  Instruction selection optimization to SETcc as suggested by John.
1036    
1037           For example,
1038    
1039            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
1040    
1041           should generate:
1042    
1043            MOVL a, x
1044            SUBL b, x
1045            SHRL 31, x
1046    
1047    2.  IR stuff
1048    
1049         A bunch of new DJ-graph related algorithms added.  These
1050         speed up SSA construction.
1051    
1052    3.  SSA + Scheduling
1053    
1054         Added code for SSA and scheduling to the repository
1055    
1056    ----------------------------------------------------------------------
1057    Name: Lal George
1058    Date: 2000/07/27 11:53:14 EDT
1059    
1060    Tag: lal-20000727-linux-ppc
1061    Description:
1062    
1063     Made changes to support Linux PPC.
1064     p.s. I have confirmation that the 110.29 boot files work fine.
1065    
1066    ----------------------------------------------------------------------
1067    Name: Matthias Blume
1068    Date: 2000/07/27 17:40:00 JST
1069    Tag: blume-20000727-scripts
1070    Description:
1071    
1072    !!!! WARNING !!!!
1073    You must recompile the runtime system!
1074    !!!! WARNING !!!!
1075    
1076    This is basically another round of script-enhancements:
1077    
1078    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
1079       and undefine CM preprocessor symbols.
1080    
1081    2. ml-build avoids generating a new heap image if it finds that the
1082       existing one is still ok.  (The condition is that no ML file had to
1083       be recompiled and all ML files are found to be older that the heap
1084       file.)
1085    
1086       To make this work smoothly, I also hacked the runtime system as
1087       well as SMLofNJ.SysInfo to get access to the heap image suffix
1088       (.sparc-solaris, ...) that is currently being used.
1089    
1090       Moreover, the signature of CM.mk_standalone has changed.  See the
1091       CM manual.
1092    
1093    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
1094       CM manual for details.)
1095    
1096    4. More CM manual updates:
1097        - all of the above has been documented.
1098        - there is now a section describing the (CM-related) command line
1099          arguments that are accepted by the "sml" command
1100    
1101    ----------------------------------------------------------------------
1102    Name: Matthias Blume
1103    Date: 2000/07/25 16:20:00 JST
1104    Tag: blume-20000725-makedepend
1105    Description:
1106    
1107    Added a script called ml-makedepend.  This can be used in makefiles
1108    for Unix' make in a way very similar to the "makedepend" command for
1109    C.
1110    
1111    The script internally uses function CM.sources.
1112    
1113    Synopsis:
1114    
1115        ml-makedepend [-f makefile] cmfile targetname
1116    
1117    The default for the makefile is "makefile" (or "Makefile" should
1118    "makefile" not exist).
1119    
1120    ml-makedepend adds a cmfile/targetname-specific section to this
1121    makefile (after removing the previous version of this section).  The
1122    section contains a single dependency specification with targetname on
1123    the LHS (targetname is an arbitrary name), and a list of files derived
1124    from the cmfile on the RHS.  Some of the files on the RHS are
1125    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
1126    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
1127    path names.  The makefile writer is responsible for making sure that
1128    these variables have correct at the time "make" is invoked.
1129    
1130    ----------------------------------------------------------------------
1131    Name: Matthias Blume
1132    Date: 2000/07/22 23:30:00 JST
1133    Tag: blume-20000722-urlupdate
1134    Description:
1135    
1136    Changed BOOT and config/srcarchiveurl to point to BL server:
1137    
1138        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
1139    
1140    ----------------------------------------------------------------------
1141    Name: Matthias Blume
1142    Date: 2000/07/18 18:00:00 JST
1143    Tag: blume-20000718-Version_110_29
1144    Description:
1145    
1146    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
1147    
1148    2. Updated config/version to 110.29
1149    
1150    3. Updated config/srcarchiveurl
1151    
1152    3. New boot files!
1153       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
1154    
1155    ----------------------------------------------------------------------
1156    Name: Matthias Blume
1157    Date: 2000/07/11 13:58:00 JST
1158    Tag: blume-20000711-doctypo
1159    Description:
1160    
1161    Fixed a few typos in CM manual.
1162    
1163    ----------------------------------------------------------------------
1164    Name: Allen Leung
1165    Date: 2000/06/15 00:38:00
1166    Tag: leunga-20000704-sparc-x86
1167    
1168    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
1169    2. fix to x86 RET bug [found by Dan Grossman]
1170    3. sparc assembly bug fix for ticc instructions [found by Fermin]
1171    
1172       Affects c-- and moby only
1173    
1174    ----------------------------------------------------------------------
1175    Name: Matthias Blume
1176    Date: 2000/07/04 15:26:00
1177    Tag: blume-20000704-trigger
1178    Description:
1179    
1180    1. Improvements to CM manual.
1181    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
1182       of getting a back-trace.  The function, when called, raises an
1183       internal exception which explicitly carries the full back-trace history,
1184       so it is unaffected by any intervening handle-raise pairs ("trivial"
1185       or not).  The interactive loop will print that history once it arrives
1186       at top level.
1187       Short of having all exceptions implicitly carry the full history, the
1188       recommended way of using this facility is:
1189         - compile your program with instrumentation "on"
1190         - run it, when it raises an exception, look at the history
1191         - if the history is "cut off" because of some handler, go and modify
1192           your program so that it explicitly calls BTrace.trigger
1193         - recompile (still instrumented), and rerun; look at the full history
1194    
1195    ----------------------------------------------------------------------
1196    Name: Matthias Blume
1197    Date: 2000/07/03 15:36:00 JST
1198    Tag: blume-20000702-manual
1199    Description:
1200    
1201    Small corrections and updates to CM manual.
1202    
1203    ----------------------------------------------------------------------
1204    Name: Matthias Blume
1205    Date: 2000/06/29 16:04:00 JST
1206    Tag: blume-20000629-yacctool
1207    Description:
1208    
1209    Changes:
1210    
1211    1. Class "mlyacc" now takes separate arguments to pass options to
1212       generated .sml- and .sig-files independently.
1213    2. Corresponding CM manual updates.
1214    3. BTrace module now also reports call sites.  (However, for loop clusters
1215       it only shows from where the cluster was entered.)  There are associated
1216       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1217    
1218    ----------------------------------------------------------------------
1219    Name: Matthias Blume
1220    Date: 2000/06/27 16:51:00 JST
1221    Tag: blume-20000627-noweb
1222    Description:
1223    
1224    Changes:
1225    
1226     1. Implemented "subdir" and "witness" options for noweb tool.
1227        This caused some slight internal changes in CM's tool implementation.
1228     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
1229        some remaining issues from earlier path anchor changes.
1230     3. Updated CM manual accordingly.
1231    
1232     4. Changed implementation of back-tracing so that I now consider it
1233        ready for prime-time.
1234    
1235        In particular, you don't have to explicitly trigger the back-trace
1236        anymore.  Instead, if you are running BTrace-instrumented code and
1237        there is an uncaught exception (regardless of whether or not it was
1238        raised in instrumented code), the top-level evalloop will print
1239        the back-trace.
1240    
1241        Features:
1242    
1243          - Instrumented and uninstrumented code work together seemlessly.
1244            (Of course, uninstrumented code is never mentioned in actual
1245             back-traces.)
1246    
1247          - Asymptotic time- and space-complexity of instrumented code is
1248            equal to that of uninstrumented code.  (This means that
1249            tail-recursion is preserved by the instrumentation phase.)
1250    
1251          - Modules whose code has been instrumented in different sessions
1252            work together without problem.
1253    
1254          - There is no penalty whatsoever on uninstrumented code.
1255    
1256          - There is no penalty on "raise" expressions, even in
1257            instrumented code.
1258    
1259        A potential bug (or perhaps it is a feature, too):
1260    
1261          A back-trace reaches no further than the outermost instrumented
1262          non-trivial "raise".  Here, a "trivial" raise is one that is the
1263          sole RHS of a "handle" rule.  Thus, back-traces reach trough
1264    
1265               <exp> handle e => raise e
1266    
1267          and even
1268    
1269               <exp> handle Foo => raise Bar
1270    
1271          and, of course, through
1272    
1273               <exp> handle Foo => ...
1274    
1275         if the exception was not Foo.
1276    
1277         Back-traces always reach right through any un-instrumented code
1278         including any of its "handle" expressions, trivial or not.
1279    
1280       To try this out, do the following:
1281    
1282         - Erase all existing binfiles for your program.
1283           (You may keep binfiles for those modules where you think you
1284            definitely don't need back-tracing.)
1285         - Turn on back-trace instrumentation:
1286              SMLofNJ.Internals.BTrace.mode (SOME true);
1287         - Recompile your program.  (I.e., run "CM.make" or "use".)
1288         - You may now turn instrumentation off again (if you want):
1289              SMLofNJ.Internals.BTrace.mode (SOME false);
1290         - Run your program as usual.  If it raises an exception that
1291           reaches the interactive toplevel, then a back-trace will
1292           automatically be printed.  After that, the toplevel loop
1293           will print the exception history as usual.
1294    
1295    ----------------------------------------------------------------------
1296    Name: Matthias Blume
1297    Date: 2000/06/26 09:56:46 JST
1298    Tag: blume-20000626-setup
1299    Description:
1300    
1301    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1302          ML code before and after compiling a file (e.g., to set compiler
1303          flags)
1304    
1305    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1306              - associated changes to btrace.sml (BTrace instrumentation pass)
1307              - cleaner implementation of btimp.sml (BTrace tracing and report
1308                module)
1309    
1310    CM manual: * new path encoding documented
1311               * description of setup-parameter to "sml" added
1312    
1313    The biggest user-visible change to back-tracing is that it is no
1314    longer necessary to compile all traced modules within the same
1315    session.  (This was a real limitation.)
1316    
1317    ----------------------------------------------------------------------
1318    Name: Matthias Blume
1319    Date: 2000/06/24 12:40:00 JST
1320    Tag: blume-20000624-startup
1321    Description:
1322    
1323    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1324    bit too often -- to put it mildly. :)
1325    
1326    ----------------------------------------------------------------------
1327    Name: Matthias Blume
1328    Date: 2000/06/23 18:20:00 JST
1329    Tag: blume-20000623-btrace
1330    Description:
1331    
1332    This updates adds a backtrace facility to aid programmers in debugging
1333    their programs.  This involves the following changes:
1334    
1335    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1336       keeping track of the current call stack.  When programs are compiled
1337       in a special mode, the compiler will insert calls to these hooks
1338       into the user program.
1339       "Hook" means that it is possible for different implementations of
1340       back-tracing to register themselves (at different times).
1341    
1342    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1343       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1344       It can be turned on using this call:
1345         SMLofNJ.Internals.BTrace.mode (SOME true);
1346       Turning it off again:
1347         SMLofNJ.Internals.BTrace.mode (SOME false);
1348       Querying the current status:
1349         SMLofNJ.Internals.BTrace.mode NONE;
1350       Annotated programs are about twice as big as normal ones, and they
1351       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1352       where all hooks do nothing).  The slowdown with a plugin that is
1353       actually useful (such as the one supplied by default) is even greater,
1354       but in the case of the default plugin it is still only an constant
1355       factor (amortized).
1356    
1357    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1358       with a sub-structure BTrace for controlling back-tracing.  In particular,
1359       the above-mentioned function "mode" controls whether the annotation
1360       phase is invoked by the compiler.  Another important function is
1361       "trigger": when called it aborts the current execution and causes
1362       the top-level loop to print a full back-trace.
1363    
1364    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1365       for back-tracing.  It keeps track of the dynamic call stack and in
1366       addition to that it keeps a partial history at each "level" of that
1367       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1368       d and b (at separate times, dynamically), then the report will show:
1369    
1370       GOTO   d
1371             /c
1372       GOTO  \b
1373       CALL   a
1374    
1375       This shows that there was an initial non-tail call of a, then a
1376       tail-call to b or c, looping behavior in a cluster of functions that
1377       consist of b and c, and then a goto from that cluster (i.e., either from
1378       b or from c) to d.
1379    
1380       Note that (depending on the user program) the amount of information
1381       that the back-trace module has to keep track of at each level is bounded
1382       by a constant.  Thus, the whole implementation has the same asymptotical
1383       complexity as the original program (both in space and in time).
1384    
1385    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1386       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1387       which is raised by the "trigger" function mentioned above.
1388    
1389    Notes on usage:
1390    
1391    - Annotated code works well together with unannotated code:
1392    Unannotated calls simply do not show up at all in the backtrace.
1393    
1394    - It is not a good idea to let modules that were annotated during
1395    different sessions run at the same time.  This is because the compiler
1396    chooses small integers to identify individual functions, and there
1397    will be clashes if different modules were compiled in separate sessions.
1398    (Nothing will crash, and you will even be told about the clashes, but
1399    back-trace information will in general not be useful.)
1400    
1401    - Back-tracing can be confused by callcc and capture.
1402    
1403    - The only way of getting a back-trace right now is to explicitly
1404    invoke the "trigger" function from your user program.  Eventually, we
1405    should make every exception carry back-trace information (if
1406    available).  But since this creates more overhead at "raise"-time
1407    (similar to the current exnHistory overhead), I have not yet
1408    implemented this.  (The implementation will be rather easy.)  With
1409    exceptions carrying back-trace information, this facility will be even
1410    more useful because users don't need to modify their programs...
1411    
1412    - While it is possible to compile the compiler with back-trace
1413    annotations turned on (I did it to get some confidence in
1414    correctness), you must make absolutely sure that core.sml and
1415    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1416    be compiled with annotation because there is no core access yet, but
1417    if you compile btimp.sml with annotation, then the system will go into
1418    an infinite recursion and crash.)
1419    Since CM currently does not know about BTrace, the only way to turn
1420    annotations on and off for different modules of the compiler is to
1421    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1422    this is awkward and clumsy.
1423    
1424    Sample sessions:
1425    
1426    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1427    - SMLofNJ.Internals.BTrace.mode (SOME true);
1428    [autoloading]
1429    [autoloading done]
1430    val it = false : bool
1431    - structure X = struct
1432    -     fun main n = let
1433    -         fun a (x, 0) = d x
1434    -           | a (x, n) = b (x, n - 1)
1435    -         and b (x, n) = c (x, n)
1436    -         and c (x, n) = a (x, n)
1437    -         and d x = e (x, 3)
1438    -         and e (x, 0) = f x
1439    -           | e (x, n) = e (x, n - 1)
1440    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1441    -           | f n = n * g (n - 1)
1442    -         and g n = a (n, 3)
1443    -     in
1444    -         f n
1445    -     end
1446    - end;
1447    structure X : sig val main : int -> int end
1448    - X.main 3;
1449    *** BACK-TRACE ***
1450    GOTO   stdIn:4.2-13.20: X.main[2].f
1451    GOTO-( stdIn:4.2-13.20: X.main[2].e
1452    GOTO   stdIn:4.2-13.20: X.main[2].d
1453         / stdIn:4.2-13.20: X.main[2].a
1454         | stdIn:4.2-13.20: X.main[2].b
1455    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1456    CALL   stdIn:4.2-13.20: X.main[2].g
1457    GOTO   stdIn:4.2-13.20: X.main[2].f
1458    GOTO-( stdIn:4.2-13.20: X.main[2].e
1459    GOTO   stdIn:4.2-13.20: X.main[2].d
1460         / stdIn:4.2-13.20: X.main[2].a
1461         | stdIn:4.2-13.20: X.main[2].b
1462    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1463    CALL   stdIn:4.2-13.20: X.main[2].g
1464    GOTO   stdIn:4.2-13.20: X.main[2].f
1465    GOTO-( stdIn:4.2-13.20: X.main[2].e
1466    GOTO   stdIn:4.2-13.20: X.main[2].d
1467         / stdIn:4.2-13.20: X.main[2].a
1468         | stdIn:4.2-13.20: X.main[2].b
1469    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1470    CALL   stdIn:4.2-13.20: X.main[2].g
1471    GOTO   stdIn:4.2-13.20: X.main[2].f
1472    CALL   stdIn:2.15-17.4: X.main[2]
1473    -
1474    
1475    (Note that because of a FLINt bug the above code currently does not
1476    compile without BTrace turned on.)
1477    
1478    Here is another example, using my modified Tiger compiler:
1479    
1480    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1481    - SMLofNJ.Internals.BTrace.mode (SOME true);
1482    [autoloading]
1483    [autoloading done]
1484    val it = false : bool
1485    - CM.make "sources.cm";
1486    [autoloading]
1487    ...
1488    [autoloading done]
1489    [scanning sources.cm]
1490    [parsing (sources.cm):parse.sml]
1491    [creating directory CM/SKEL ...]
1492    [parsing (sources.cm):tiger.lex.sml]
1493    ...
1494    [wrote CM/sparc-unix/semant.sml]
1495    [compiling (sources.cm):main.sml]
1496    [wrote CM/sparc-unix/main.sml]
1497    [New bindings added.]
1498    val it = true : bool
1499    - Main.compile ("../testcases/merge.tig", "foo.out");
1500    *** BACK-TRACE ***
1501    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
1502    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1503    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1504    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1505    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1506    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1507    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
1508         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
1509    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
1510    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1511    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
1512    CALL   main.sml:1.18-118.4: Main.compile[2]
1513    -
1514    
1515    ----------------------------------------------------------------------
1516    Name: Matthias Blumen
1517    Date: 2000/06/21 18:00:00 JST
1518    Tag: blume-20000621-manual
1519    Description:
1520    
1521    CM manual update: Path environments documented.
1522    
1523    ----------------------------------------------------------------------
1524    Name: Matthias Blume
1525    Date: 2000/06/19 13:40:00
1526    Tag: blume-20000619-manual
1527    Description:
1528    
1529    CM manual and system/README update.  This only covers the fact that
1530    there are no more implicit anchors.  (Path environments and the "bind"
1531    option to "cm" have yet to be documented.)
1532    
1533    ----------------------------------------------------------------------
1534    Name: Matthias Blume
1535    Date: 2000/06/19 11:05:00 JST
1536    Tag: blume-20000619-chdir-bugfix
1537    Description:
1538    
1539    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1540    
1541    ----------------------------------------------------------------------
1542    Name: Matthias Blume
1543    Date: 2000/06/18 22:00:10 JST
1544    Tag: blume-20000618-implicit-anchors-really-gone
1545    Description:
1546    
1547    I updates the previous HISTORY entry where I forgot to mention that
1548    implicit anchors are no longer with us.
1549    
1550    The current update also gets rid of the (now useless) controller
1551    CM.Control.implicit_anchors.
1552    
1553    ----------------------------------------------------------------------
1554    Name: Matthias Blume
1555    Date: 2000/06/16 17:30:00 JST
1556    Tag: blume-20000616-anchorenv
1557    Description:
1558    
1559    This patch implements the long anticipated (just kidding :) "anchor
1560    environment" mechanism.  In the course of doing this, I also
1561    re-implemented CM's internal "SrcPath" module from scratch.  The new
1562    one should be more robust in certain boundary cases.  In any case, it
1563    is a lot cleaner than its predecessor (IMHO).
1564    
1565    This time, although there is yet another boot file format change, I
1566    kept the unpickler backward-compatible.  As a result, no new bootfiles
1567    are necessary and bootstrapping is straightforward.  (You cannot read
1568    new bootfiles into an old system, but the other way around is no
1569    problem.)
1570    
1571    Visible changes:
1572    
1573    ** 0. Implicit path anchors (without the leading $-symbol) are no
1574    longer recognized at all. This means that such path names are not
1575    illegal either.  For example, the name basis.cm simply refers to a
1576    local file called "basis.cm" (i.e, the name is an ordinary path
1577    relative to .cm-files directory).  Or, to put it differently, only
1578    names that start with $ are anchored paths.
1579    
1580    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1581    vanished.
1582    
1583    John (Reppy) had critizised this as soon as I originally proposed and
1584    implemented it, but at that time I did not really deeply believe
1585    him. :) Now I came full-circle because I need the $<singlearc> syntax
1586    in another place where it cannot be seen as an abbreviation for
1587    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1588    seems to mean (i.e., it "expands" into the corresponding anchor
1589    value).
1590    
1591    However, when paths are used as members in CM description files, it
1592    continues to be true that there must be at least another arc after the
1593    anchor.  This is now enforced separately during semantic analysis
1594    (i.e., from a lexical/syntactical point of view, the notation is ok.)
1595    
1596    ** 2. The "cm" class now accepts an option "bind".  The option's value
1597    is a sub-option list of precisely two items -- one labeled "anchor"
1598    and the other one labeled "value".  As you might expect, "anchor" is
1599    used to specify an anchor name to be bound, and "value" specifies what
1600    the anchor is being bound to.
1601    
1602    The value must be a directory name and can be given in either standard
1603    syntax (including the possibility that it is itself an anchored path)
1604    or native syntax.
1605    
1606    Examples:
1607    
1608       foo.cm (bind:(anchor:bar value:$mystuff/bar))
1609       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1610    
1611    and so on.
1612    
1613    The meaning of this is that the .cm-file will be processed with an
1614    augmented anchor environment where the given anchor(s) is/are bound to
1615    the given values(s).
1616    
1617    The rationale for having this feature is this: Suppose you are trying
1618    to use two different (already stable) libraries a.cm and b.cm (that
1619    you perhaps didn't write yourself).  Further, suppose each of these
1620    two libraries internally uses its own auxiliary library $aux/lib.cm.
1621    Normally you would now have a problem because the anchor "lib" can not
1622    be bound to more than one value globally.  Therefore, the project that
1623    uses both a.cm and b.cm must locally redirect the anchor to some other
1624    place:
1625    
1626       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1627       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1628    
1629    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1630    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
1631    
1632    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1633    you when you do so at the time of CM.stabilize).  Therefore, you can
1634    also use an anchored path as the value:
1635    
1636      a.cm (bind:(anchor:lib value:$a-lib))
1637      b.cm (bind:(anchor:lib value:$b-lib))
1638    
1639    Now you can globally configure (using the usual CM.Anchor.anchor or
1640    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1641    itself is always locally bound, setting it globally is no longer
1642    meaningful or necessary (but it does not hurt either).  In fact, "lib"
1643    can still be used as a global anchor for separate purposes.  As a
1644    matter of fact, one can locally define "lib" in terms of a global
1645    "lib":
1646    
1647      a.cm (bind:(anchor:lib value:$lib/a))
1648      b.cm (bind:(anchor:lib value:$lib/b))
1649    
1650    ** 3: The encoding of path names has changed.  This affects the way
1651    path names are shown in CM's progress report and also the internal
1652    protocol encoding used for parallel make.
1653    
1654    The encoding now uses one or more ':'-separated segments.  Each
1655    segments corresponds to a file that has been specified relative to the
1656    file given by its preceding segment.  The first segment is either
1657    relative to the CWD, absolute, or anchored.  Each segment itself is
1658    basically a Unix pathname; all segments but the first are relative.
1659    
1660    Example:
1661    
1662       $foo/bar/baz.cm:a/b/c.sml
1663    
1664    This path denotes the file bar/a/b/c.sml relative to the directory
1665    denoted by anchor "foo".  Notice that the encoding also includes
1666    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1667    paths are resolved relative to the .cm-files directory, so baz.cm must
1668    be ignored to get the "real" pathname.
1669    
1670    To make this fact more obvious, CM puts the names of such "virtual
1671    arcs" into parentheses when they appear in progress reports. (No
1672    parentheses will appear in the internal protocol encoding.)  Thus,
1673    what you really see is:
1674    
1675      $foo/bar/(baz.cm):a/b/c.sml
1676    
1677    I find this notation to be much more informative than before.
1678    
1679    Another new feature of the encoding is that special characters
1680    including parentheses, colons, (back)slashes, and white space are
1681    written as \ddd (where ddd is the decimal encoding of the character).
1682    
1683    *** The CM manual still needs to be updated.
1684    
1685    ----------------------------------------------------------------------
1686    Name: Allen Leung
1687    Date: 2000/06/15 00:38:00
1688    Tag: leunga-20000615-x86-peephole
1689    
1690    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1691    
1692    ----------------------------------------------------------------------
1693    Name: Matthias Blume
1694    Date: 2000/06/12 11:40:00
1695    Tag: blume-20000612-parmakefix
1696    Description:
1697    
1698    More cleanup after changing the file naming scheme: This time I
1699    repaired the parallel make mechanism for CMB.make which I broke earlier.
1700    
1701    ----------------------------------------------------------------------
1702    Name: Allen Leung
1703    Date: 2000/06/09 01:25:00
1704    Tag: leunga-20000609-various
1705    
1706    None of these things should affect normal SML/NJ operations
1707    
1708    1. Peephole improvements provided by Fermin (c--)
1709    2. New annotation DEFUSE for adding extra dependence (moby)
1710    3. New X86 LOCK instructions (moby)
1711    4. New machine description language for reservation tables (scheduling)
1712    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1713       trees etc.)
1714    6. I've changed the CM files so that they can work with versions
1715       110.0.6, 110.25 and 110.28
1716    
1717    ----------------------------------------------------------------------
1718    Name: Matthias Blume
1719    Date: 2000/06/09 12:40:00
1720    Tag: blume-20000609-log
1721    Description:
1722    
1723    - Removed all(?) remaining RCS Log entries from sources.
1724    
1725    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1726      anchored paths).
1727    
1728    ----------------------------------------------------------------------
1729    Name: Matthias Blume
1730    Date: 2000/06/07 17:00:00 JST
1731    Tag: blume-20000607-no-implicit-anchors
1732    Description:
1733    
1734    1. This update changes the default setting for
1735    CM.Control.implicit_anchors from true to false.  This means that
1736    implicit anchors are no longer permitted by default.  I also tried to
1737    make sure that nothing else still relies on implicit anchors.
1738    (This is the next step on the schedule towards a CM that does not even
1739    have the notion of implicit anchors anymore.)
1740    
1741    2. More CM manual updates.
1742    
1743    3. I managed to track down and fix the pickling bug I mentioned last
1744    time.  Because of the previously existing workaround, this entails no
1745    immediate practical changes.
1746    
1747    ----------------------------------------------------------------------
1748    Name: Matthias Blume
1749    Date: 2000/06/06 11:15:00 JST
1750    Tag: blume-20000606-lazierpickle
1751    Description:
1752    
1753    !!!! NEW BOOT FILES !!!!
1754    
1755    * The main purpose of this update is to make library pickles lazier in
1756    order to reduce the initial space penalty for autoloading a library.
1757    As a result, it is now possible to have $smlnj/compiler.cm
1758    pre-registered.  This should take care of the many complaints or
1759    inquiries about missing structure Compiler.  This required changes to
1760    CM's internal data structures and small tweaks to some algorithms.
1761    
1762    As a neat additional effect, it is no longer necessary (for the sake
1763    of lean heap image files) to distinguish between a "minimal" CM and a
1764    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1765    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1766    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1767    
1768    To make the life of compiler-hackers easier, "makeml" now also
1769    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1770    words, after you bootstrap a new sml for the first time, you will not
1771    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1772    around you will still have to do it, though.)
1773    
1774    * A second change consists of major updates to the CM manual.  There
1775    are now several appendices with summary information and also a full
1776    specification of the CM description file syntax.
1777    
1778    * In directory src/system I added the script "allcross".  This script
1779    invokes sml and cross-compiles the compiler for all supported
1780    architectures.  (Useful when providing a new set of boot files.)
1781    
1782    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1783    added a small tweak to pickle-util.sml to work around this problem,
1784    but it is not a proper fix yet.  I will investigate further.  (The
1785    effect of the bug was an inflation of library pickle size.)
1786    
1787    * Version number increased to 110.28.1 (to avoid compatibility problems).
1788    
1789    ----------------------------------------------------------------------
1790    Name: Allen Leung
1791    Date: 2000/05/25 17:28 EDT
1792    Tag: leunga-20000525-ra
1793    Description:
1794    
1795      Fixed a bug in freezing phase of the register allocator.
1796    
1797    ----------------------------------------------------------------------
1798    Name: Allen Leung
1799    Date: 2000/05/15 22:53 EDT
1800    Tag: leunga-20000515-alpha-x86-ra
1801    Description:
1802    
1803      1. Alpha
1804    
1805          Slight cleanup.  Removed the instruction SGNXL
1806    
1807      2. X86
1808    
1809          Added the following instructions to the instruction set:
1810    
1811            ROLx, RORx,
1812            BTx, BTSx, BTLx, BTRx,
1813            XCHGx, and variants with the LOCK prefix
1814    
1815      3. Register Allocation
1816    
1817          The module ra-rewrite-with-renaming has been improved.
1818    
1819      These have no effect on SML/NJ.
1820    
1821    ----------------------------------------------------------------------
1822    Name: Matthias Blume
1823    Date: 2000/05/15 16:20:00 JST
1824    Tag: blume-20000515-lightrebuild
1825    Description:
1826    
1827    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1828       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1829       defined.  In effect, the command will not build any cross-compiler
1830       backends and therefore finish more quickly.
1831    
1832       The "fixpt" script also takes a "-light" switch to be able to use
1833       this new facility while compiling for a fixpoint.
1834    
1835    2. I replaced all mentions of anchored paths in group owner specifications
1836       with simple relative paths (usually starting with "..").
1837       The rationale is that a library's internal workings should not be
1838       compromised by the lack of some anchor.  (An anchor is necessary
1839       for someone who wants to refer to the library by an anchored path,
1840       but it should not be necessary to build the same library in the first
1841       place.)
1842    
1843    3. I changed the way CM's tool mechanism determines the shell command
1844       string used for things like ml-yacc etc. so that it does not break
1845       when CM.Control.implicit_anchors is turned off.
1846    
1847    ----------------------------------------------------------------------
1848    Name: Matthias Blume
1849    Date: 2000/05/12 18:20:00 JST
1850    Tag: blume-20000512-ml-build
1851    Description:
1852    
1853    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1854    getting installed properly (by config/install.sh).
1855    
1856    ----------------------------------------------------------------------
1857    Name: Matthias Blume
1858    Date: 2000/05/12 17:30:00 JST
1859    Tag: blume-20000512-anchors
1860    Description:
1861    
1862    !!! NEW BOOT FILES !!!
1863    
1864    This change is in preparation of fading out support for "implicitly
1865    anchored path names".  I went through all sources and used the
1866    explicit (and relatively new) $-notation.  See system/README and the
1867    CM manual for more info on this.
1868    
1869    I also modified the anchoring scheme for some things such as "smlnj",
1870    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1871    anchors are more expressive: anchor name and first arc do not have to
1872    coincide.  This entails the following user-visible change:
1873    
1874    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1875    particular, when you fire up sml with a command-line argument, say,
1876    e.g.:
1877    
1878       sml '$smlnj/cmb.cm'
1879    
1880    At the ML toplevel prompt:
1881    
1882       CM.autoload "$smlnj/cmb.cm";
1883    
1884    There is also a new controller in CM.Control that can be used to turn
1885    off all remaining support for implicit anchors by saying:
1886    
1887        CM.autoload "$smlnj/
1888        #set CM.Control.implicit_anchors false;
1889    
1890    This causes CM to reject implicitly anchored paths.  This is (for the
1891    time being) less permissive than the "final" version where there will
1892    be no more such implicit anchors and relative paths will be just that:
1893    relative.
1894    
1895    The next step (version after next version?) will be to make the
1896    default for CM.Control.implicit_anchors false.  After the dust has
1897    settled, I can then produce the "final" version of this...
1898    
1899    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1900    
1901    ----------------------------------------------------------------------
1902    Name: Matthias Blume
1903    Date: 2000/05/11 16:30:00 JST
1904    Tag: blume-20000511-sources
1905    Description:
1906    
1907    The main change is that I added function CM.sources as a generalized
1908    version of the earlier CM.makedepend.  This entails the following
1909    additional changes:
1910    
1911      - CM.makedepend has been dropped.
1912    
1913      - CM manual has been updated.
1914    
1915      - TOOLS signature and API have been changed.
1916    
1917    ----------------------------------------------------------------------
1918    Name: Allen Leung
1919    Date: 2000/05/10 21:17 EDT
1920    Tag: leunga-20000510-moby-c--ssa
1921    Description:
1922    
1923      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1924    None of these affect SML/NJ.
1925    
1926    1. Register Allocation
1927    
1928        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1929           This module tries to remove local (i.e. basic block level) redundancies
1930           during spilling.
1931    
1932        b. A new framework for performing region based register allocation.
1933           Not yet entirely functional.
1934    
1935    2. X86
1936    
1937       a. DefUse for POP was missing the stack pointer [found by Lal]
1938       b. Reload for CALL was incorrect in X86Spill [found by John]
1939       c. Various fixes in X86Spill so that it can be used correctly for
1940          the new spilling module.
1941    
1942    3. SSA/IR
1943    
1944       a. New module ir/dj-dataflow.sml implements elimination based
1945          data flow analysis.
1946    
1947    4. MLRiscGen
1948    
1949       a. Fix for gc type annotation
1950    
1951    5. MDGen
1952    
1953       Various fixes for machine description -> ml code translation.  For ssa
1954       only.
1955    
1956    ----------------------------------------------------------------------
1957    Name: Allen Leung
1958    Date: 2000/05/08 22:17 EDT
1959    Tag: leunga-20000508-labexp
1960    Description:
1961    
1962      Fermin has found a few assembly problems with constant expressions
1963      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1964      which choke on dumb assemblers.  This is his fix.
1965    
1966    ----------------------------------------------------------------------
1967    Name: Dave MacQueen
1968    Date: 2000/04/09 14:00 EDT
1969    Tag: dbm-20000502-Version_110_28
1970    Description:
1971    
1972    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1973    
1974    2. Updated config/version to 110.28
1975    
1976    3. Updated config/srcarchiveurl
1977    
1978    3. New boot files!
1979       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1980    
1981    ----------------------------------------------------------------------
1982    Name: Matthias Blume
1983    Date: 2000/05/01 19:05:00 JST
1984    Tag: blume-20000501-noweb
1985    Description:
1986    
1987    A new noweb tool has been added.  The existing system is entirely
1988    unaffected by this, but some CM users have asked for renewed noweb
1989    support.  Everything is documented in the CM manual.
1990    
1991    New (plugin) libraries:
1992    
1993       noweb-tool.cm
1994       nw-ext.cm
1995    
1996    ----------------------------------------------------------------------
1997    Name: Dave MacQueen
1998    Date: 2000/04/30 12:40PM EDT
1999    Tag: dbm-20000430-bug_fixes
2000    Description:
2001    
2002    1. Fix for bug 1498
2003       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
2004       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
2005         added toRealArray function
2006       smlnj/src/compiler/MiscUtil/print/ppobj.sml
2007         added check for tag Obj.RealArray to array printing case in ppObj
2008    
2009    2. Fix for bug 1510
2010       smlnj/src/compiler/Semant/types/typesutil.sml
2011         fixed definition of dummyargs (used by equalTycon) so that
2012         dummy args are distinct types
2013    
2014    ----------------------------------------------------------------------
2015    Name: Matthias Blume
2016    Date: 2000/04/30 01:00:00 JST
2017    Tag: blume-20000430-versions
2018    Description:
2019    
2020    1. CM version numbering added.  This is an implementation of Lal's
2021       proposal for adding version numbers and version checking to .cm
2022       files.  Lal said that his proposal was just that -- a proposal.
2023       For the time being I went ahead and implemented it so that people
2024       can comment on it.  Everything is completely backward-compatible
2025       (except for the stable library format, i.e., new bootfiles!).
2026    
2027       As usual, see the CM manual for details.
2028    
2029    2. An alternative syntax for anchored paths has been implemented.
2030       Dave has recently voiced the same concerns that I had when I did
2031       this, so there should be some support.  My take is that eventually
2032       I will let support for the current syntax (where anchors are
2033       "implicit") fade out in favor of the new, explicit syntax.
2034       In order to be backward-compatible, both old and new syntax are
2035       currently supported.
2036    
2037       Again, see the CM manual for details.
2038    
2039    3. Parallel make is trying to be slightly smarter:  When the master
2040       process finds a "bottleneck", i.e., when there is only one
2041       compilation unit that can be compiled and everybody else is
2042       waiting on it, then it will simply compile it directly instead
2043       of clumsily telling one of the slaves to do it.
2044    
2045    4. Support for "unsharing" added.  This is necessary in order to be
2046       able to have two different versions of the same library running
2047       at the same time (e.g., for trying out a new MLRISC while still
2048       having the old MLRISC linked into the current compiler, etc.)
2049       See the CM manual.
2050    
2051    5. Simple "makedepend" functionality added for generating Makefile
2052       dependency information.  (This is rather crude at the moment.
2053       Expect some changes here in the future.)
2054    
2055    6. ".fun" added as a recognized suffix for ML files. Also documented
2056       explicitly in the manual that the fallback behavior (unknown suffix
2057       -> ML file) is not an official feature!
2058    
2059    7. Small changes to the pickler for stable libraries.
2060    
2061    8. Several internal changes to CM (for cleanup/improvement).
2062    
2063    
2064    !!!! NEW BINFILES !!!!
2065    
2066    ----------------------------------------------------------------------
2067    Name: Matthias Blume
2068    Date: 2000/04/28 17:30:00 JST
2069    Tag: blume-20000428-pathconfig
2070    Description:
2071    
2072    1. I changed config/install.sh to remove duplicate entries from the
2073       lib/pathconfig file at the end.  Moreover, the final version of
2074       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
2075       in src/system/installml.
2076    
2077    2. The config/install.sh script now consistently uses relative
2078       pathnames in lib/pathconfig whenever the anchor is in the lib
2079       directory.  (So far this was true for the libraries that come
2080       pre-compiled and bundled as part of the bootfiles but not for
2081       libraries that are compiled by the script itself.)
2082    
2083    ----------------------------------------------------------------------
2084    Name: Matthias Blume
2085    Date: 2000/04/26 13:10:00 JST
2086    Tag: blume-20000426-fun_suffix
2087    Description:
2088    
2089    Added ".fun" as a recognized file name suffix (for ML code).
2090    
2091    ----------------------------------------------------------------------
2092    Name: Allen Leung
2093    Date: 2000/04/25 17:00:00 EST
2094    Tag: leunga-20000425-alpha-ra
2095    Description:
2096    
2097    1. Alpha
2098    
2099        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
2100    in C--.
2101    
2102    2. RA
2103    
2104       Added a flag "ra-dump-size" to print out the size of the flowgraph
2105       and the interference graph.
2106    
2107    ----------------------------------------------------------------------
2108    Name: Dave MacQueen
2109    Date: 2000/04/25/
2110    Tag: dbm-20000425-mlyacc_doc_examples
2111    Description:
2112      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
2113      Updated all three examples in src/ml-yacc/examples to run
2114      under 110.* using CM.make.
2115    
2116    ----------------------------------------------------------------------
2117    Name: Allen Leung
2118    Date: 2000/04/20 23:04:00 EST
2119    Tag: leunga-20000420-ssa-c---stuff
2120    Description:
2121    
2122      This update synchronizes my repository with Yale's.  Most of these
2123    changes, however, do not affect SML/NJ at all (the RA is an exception).
2124    
2125    1. Register Allocator
2126    
2127       a. An improvement in the interference graph construction:
2128          Given a copy
2129    
2130                s <- t
2131    
2132          no interference edge between s and t is added for this definition of s.
2133    
2134       b. I've added two new spill heuristic modules that Fermin and I developed
2135          (in the new library RA.cm). These are unused in SML/NJ but maybe
2136          useful for others (Moby?)
2137    
2138    2. X86
2139    
2140       a. Various fixes in the backend provided by Fermin [C--] and Lal.
2141    
2142    3. Alpha
2143    
2144       a. Added the BSR instruction and code generation that goes with it [C--]
2145       b. Other fixes too numerous to recount provided by Fermin [C--]
2146    
2147    4. Regmaps
2148    
2149       a. The regmaps are not initialized with the identity physical bindings
2150          at creation time.  This is unneeded.
2151    
2152    5. MLRISC Optimizations
2153    
2154       a. The DJ-Graph module can now compute the iterated dominance frontiers
2155          intersects with liveness incrementally in linear time! Woohoo!
2156          This is now used in my new SSA construction algorithm.
2157    
2158       b. THe branch reorganization module is now smarter about linear chains of
2159          basic blocks.
2160    
2161    
2162  ----------------------------------------------------------------------  ----------------------------------------------------------------------
2163  Name: Matthias Blume  Name: Matthias Blume
2164  Date: 2000/04/12 13:52:00 JST  Date: 2000/04/12 13:52:00 JST
# Line 129  Line 2277 
2277    
2278  3.  Assembly  3.  Assembly
2279    
2280      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
2281      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
2282      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
2283    
# Line 152  Line 2300 
2300    
2301  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
2302  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
2303  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
2304  legal surface-syntax symbol).  legal surface-syntax symbol).
2305    
2306  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 162  Line 2310 
2310  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
2311    
2312  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2313  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
2314  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
2315  more details.  more details.
2316    
# Line 277  Line 2425 
2425     (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
2426     used).     used).
2427    
2428     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
2429     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
2430     Makefiles or Shell scripts.     Makefiles or Shell scripts.
2431    
2432     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 290  Line 2438 
2438    
2439     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
2440    
2441     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
2442     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
2443    
2444  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 424  Line 2572 
2572    
2573       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
2574       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
2575       testl is perferable to andl.       testl is preferable to andl.
2576    
2577  7. RA (x86 only)  7. RA (x86 only)
2578    
# Line 606  Line 2754 
2754    
2755  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2756  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2757  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
2758  explanation.  explanation.
2759    
2760  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2761  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2762  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2763  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 653  Line 2801 
2801  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
2802  <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,
2803  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
2804  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2805  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2806  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2807    
# Line 703  Line 2851 
2851    it from that remote directory.    it from that remote directory.
2852    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2853    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2854    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2855    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2856    
2857    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.615  
changed lines
  Added in v.844

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