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 845, Thu Jun 21 16:25:49 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: 2001/06/21 12:25:00 EDT
18    Tag: blume-20010621-eXene
19    Description:
20    
21    eXene now compiles and (sort of) works again.
22    
23    The library name (for version > 110.33) is $/eXene.cm.
24    
25    I also added an new example in src/eXene/examples/nbody.  See the
26    README file there for details.
27    
28    ----------------------------------------------------------------------
29    Name: Matthias Blume
30    Date: 2001/06/20 16:40:00 EDT
31    Tag: blume-20010620-cml
32    Description:
33    
34    CML now compiles and works again.
35    
36    Libraries (for version > 110.33):
37    
38      $cml/cml.cm            Main CML library.
39      $cml/basis.cm          CML's version of $/basis.cm.
40      $cml/cml-internal.cm   Internal helper library.
41      $cml/core-cml.cm       Internal helper library.
42      $cml-lib/trace-cml.cm  Tracing facility.
43      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
44    
45    The installer (config/install.sh) has been taught how to properly
46    install this stuff.
47    
48    ----------------------------------------------------------------------
49    Name: Matthias Blume
50    Date: 2001/06/19 17:55:00 EDT
51    Tag: blume-20010619-instantiate
52    Description:
53    
54    This un-breaks the fix for bug 1432.
55    (The bug was originally fixed in 110.9 but I broke it again some
56    time after that.)
57    
58    ----------------------------------------------------------------------
59    Name: Matthias Blume
60    Date: 2001/06/19 17:25:00 EDT
61    Tag: blume-20010619-signals
62    Description:
63    
64    This should (hopefully) fix the long-standing signal handling bug.
65    (The runtime system was constructing a continuation record with an
66    incorrect descriptor which would cause the GC to drop data on the floor...)
67    
68    ----------------------------------------------------------------------
69    Name: Matthias Blume
70    Date: 2001/06/15 15:05:00 EDT
71    Tag: blume-20010615-moresparc
72    Description:
73    
74    Here is a short late-hour update related to Sparc c-calls:
75    
76     -- made handling of double-word arguments a bit smarter
77    
78     -- instruction selection phase tries to collapse certain clumsily
79        constructed ML-Trees; typical example:
80    
81            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
82    
83        This currently has no further impact on SML/NJ since mlriscGen does
84        not seem to generate such patterns in the first place, and c-calls
85        (which did generate them in the beginning) has meanwhile been fixed
86        so as to avoid them as well.
87    
88    ----------------------------------------------------------------------
89    Name: Matthias Blume
90    Date: 2001/06/15 15:05:00 EDT
91    Tag: blume-20010615-sparc
92    Description:
93    
94    The purpose of this update is to provide an implementation of NLFFI
95    on Sparc machines.
96    
97    Here are the changes in detail:
98    
99       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
100       the Sparc implementation of the c-calls API.
101       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
102       for accessing the ML frame.  Thus, we have a real frame pointer and
103       can freely modify %sp without need for an omit-frame-ptr phase.
104       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
105       this case.
106       * ml-nlffigen has been taught to produce code for different architectures
107       and calling conventions.
108       * In a way similar to what was done in the x86 case, the Sparc
109       backend uses its own specific extension to mltree.  (For example,
110       it needs to be able to generate UNIMP instructions which are part
111       of the calling convention.)
112       * ml-nlffi-lib was reorganized to make it more modular (in particular,
113       to make it easier to plug in new machine- and os-dependent parts).
114    
115    There are some other fairly unrelated bug fixes and cleanups as well:
116    
117       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
118       that they properly share their libraries with existing SML/NJ libraries.
119       * I fixed a minor cosmetic bug in CM, supressing certain spurious
120       follow-up error messages.
121       * Updates to CM/CMB documentation.
122    
123    TODO items:
124    
125       * MLRISC should use a different register as its asmTemp on the Sparc.
126         (The current %o2 is a really bad choice because it is part of the
127         calling conventions, so things might interfere in unexpected ways.)
128    
129    ----------------------------------------------------------------------
130    Name: Matthias Blume
131    Date: 2001/06/07
132    Tag: blume-20010607-calls
133    Description:
134    
135    A number of internal changes related to C calls and calling conventions:
136    
137    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
138       number of bytes popped implicitly (by the callee).  In most cases
139       this field is 0 but on x86/win32 it is some non-zero value.  This
140       is information provided for the benefit of the "omit-frameptr" pass.
141    2. The CALL instruction on the x86 carries a similar "pops" field.
142       The instruction selection phase copies its value from the ML-Tree
143       CALL statement.
144    3. On all other architectures, the instruction selection phase checks
145       whether "pops=0" and complains if not.
146    4. The c-calls implementation for x86 now accepts two calling conventions:
147       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
148       up after the call and pops is set to 0.  For "stdcall", the caller
149       does nothing, leaving the cleanup to the callee; pops is set to
150       the number of bytes that were pushed onto the stack.
151    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
152       distinguish between "ccall" and "stdcall".
153    6. The UNIMP instruction has been added to the supported Sparc instruction
154       set. (This is needed for implementing the official C calling convention
155       on this architecture.)
156    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
157       work with the latest CM.
158    
159    ----------------------------------------------------------------------
160    Name: Matthias Blume
161    Date: 2001/06/05 15:10:00 EDT
162    Tag: blume-20010605-cm-index
163    Description:
164    
165    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
166    
167    1. CM can now generate "index files".  These are human-readable files
168       that list on a per-.cm-file basis each toplevel symbol defined or
169       imported.  The location of the index file for
170       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
171       To enable index-file generation, set CM.Control.generate_index to true
172       or export an environment-symbol: export CM_GENERATE_INDEX=true.
173    
174       The CM manual has been updated accordingly.
175    
176    2. I made some slight modifications to the c-calls API in MLRISC.
177    
178         a) There is now a callback to support saving/restoring of
179            dedicated but caller-save registers around the actual call
180            instruction.
181         b) One can optionally specify a comment-annotation for the
182            call instruction.
183    
184    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
185       (For example, the comment annotation shows the C prototype of
186        the function being called.)
187    
188    ----------------------------------------------------------------------
189    Name: Matthias Blume
190    Date: 2001/06/01 13:30:00 EDT
191    Tag: blume-20010601-nlffi-cleanup
192    Description:
193    
194    This is mostly a cleanup of MLFFI stuff:
195    
196       - some signature files have been put into a more exposed place
197       - the ugly 'f type parameter is gone (simplifies types tremendously!)
198       - ml-nlffigen changed accordingly
199       - tutorial updated
200    
201    Other changes:
202    
203       - author's affiliation in CM manual(s) updated
204       - some more recognized keywords added to Allen's sml.sty
205    
206    ----------------------------------------------------------------------
207    Name: Matthias Blume
208    Date: 2001/05/25 15:30:00 EDT
209    Tag: blume-20010525-iptr
210    Description:
211    
212      - put the official 110.33-README (as it appears on the ftp server) under
213        CVS
214      - fixed a small bug related to incomplete pointer types in
215        ml-nlffigen
216      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
217        (it does not need the 'f type parameter)
218    
219    ----------------------------------------------------------------------
220    Name: Matthias Blume
221    Date: 2001/05/23 14:30:00 EDT
222    Tag: Release_110_33
223    Description:
224    
225    New version number (110.33).  New bootfiles.
226    
227    ----------------------------------------------------------------------
228    Name: Matthias Blume
229    Date: 2001/05/22 18:06:00 EDT
230    Tag: blume-20010522-targets
231    Description:
232    
233    Made install.sh use file config/targets.customized if it exists, falling
234    back to config/targets if it doesn't.  This way one can have a customized
235    version of the targets file without touching the "real thing", thus
236    eliminating the constant fear of accidentally checking something bogus
237    back into the CVS repository...  (File config/targets.customized must
238    not be added to the repository!)
239    
240    ----------------------------------------------------------------------
241    Name: Matthias Blume
242    Date: 2001/05/22 16:30:00 EDT
243    Tag: blume-20010522-minitut
244    Description:
245    
246    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
247       struct returns.
248    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
249       incomplete, preliminary documentation for NLFFI.
250    
251    ----------------------------------------------------------------------
252    Name: Matthias Blume
253    Date: 2001/05/14 11:30:00 EDT
254    Tag: blume-20010514-script
255    Description:
256    
257    Some bugs in install script fixed.
258    
259    In addition to that I also made a slight change to the NLFFI API:
260    Functors generated by ml-nlffigen now take the dynamic library as a
261    straight functor argument, not as a suspended one.  (The original
262    functor code used to force the suspension right away anyway, so there
263    was nothing gained by this complication of the interface.)
264    
265    ----------------------------------------------------------------------
266    Name: Matthias Blume
267    Date: 2001/05/11 14:35:00 EDT
268    Tag: blume-20010511-ml-nlffi
269    Description:
270    
271    I finally took the plunge and added my new FFI code to the main
272    repository.  For x86-linux it is now ready for prime-time.
273    
274    There are two new subdirectories of "src":
275    
276      - ml-nlffi-lib:
277           The utility library for programs using the FFI interface.
278           Here is the implementation of $/c.cm and its associated low-level
279           partners $/c-int.cm and $/memory.cm.
280      - ml-nlffigen:
281           A stand-alone program for generating ML glue code from C source
282           code.
283    
284    Building ml-nlffigen requires $/ckit-lib.cm.
285    
286    The config/install.sh script has been updates to do the Right Thing
287    (hopefully).
288    
289    Notice that the source tree for the C-Kit will not be put under "src"
290    but directly under the installation root directory.  (This is the
291    structure that currently exists on the CVS server when you check out
292    module "sml".)  Fortunately, config/install.sh knows about this oddity.
293    
294    Bugs: No documentation yet.
295    
296    ----------------------------------------------------------------------
297    Name: Matthias Blume
298    Date: 2001/05/09 16:35:00 EDT
299    Tag: blume-20010509-cpscontract
300    Description:
301    
302    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
303    wrapper/unwrapper elimination did not decrement usage counts and some
304    dead variables got overlooked by the dead-up logic.)
305    
306    ----------------------------------------------------------------------
307    Name: Lal George
308    Date: 2001/05/08  17:26:09 EDT
309    Tag: george-20010508-omit-frameptr
310    Description:
311    
312    Changes to implement the omit-frame-pointer optimization to support
313    raw C calls. For now, there is only support on the Intel x86, but
314    other architectures will follow as more experience is gained with this.
315    
316    
317    ----------------------------------------------------------------------
318    Name: Matthias Blume
319    Date: 2001/05/07 14:40:00 EDT
320    Tag: blume-20010507-proxies
321    Description:
322    
323    I made into "proxy libraries" all libraries that qualify for such a
324    change.  (A qualifying library is a library that has another library or
325    groups as its sole member and repeats that member's export list
326    verbatim.  A proxy library avoids this repetition by omitting its export
327    list, effectively inheriting the list that its (only) member exports.
328    See the CM manual for more explanation.)
329    The main effect is that explicit export lists for these libraries
330    do not have to be kepts in sync, making maintenance a bit easier.
331    
332    I also added copyright notices to many .cm-files.
333    
334    Last but not least, I made a new set of bootfiles.
335    
336    ----------------------------------------------------------------------
337    Name: Matthias Blume
338    Date: 2001/05/04 17:00:00 EDT
339    Tag: blume-20010504-cm-lsplit
340    Description:
341    
342    0. John merged pending changes to $/smlnj-lib.cm
343    
344    1. Allen's previous change accidentally backed out of one of Lal's
345       earlier changes.  I undid this mistake (re-introducing Lal's change).
346    
347    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
348       within the compiler where applicable.  There is some code simplification
349       because of that.
350    
351    3. The "split" phase (in FLINT) is now part of the default list of phases.
352       Compiler.Control.LambdaSplitting.* can be used to globally control the
353       lambda-splitting (cross-module-inlining) engine.  In addition to that,
354       it can now also be controlled on a per-source basis: CM has been taught
355       a new tool parameter applicable to ML source files.
356    
357       - To turn lambda-splitting off completely:
358            local open Compiler.Control.LambdaSplitting in
359                val _ = set Off
360            end
361       - To make "no lambda-splitting" the global default (but allow per-source
362         overriding); this is the initial setting:
363            local open Compiler.Control.LambdaSplitting in
364                val _ = set (Default NONE)
365            end
366       - To make "lambda-splitting with aggressiveness a" the global default
367         (and allow per-source overriding):
368            local open Compiler.Control.LambdaSplitting in
369                val _ = set (Default (SOME a))
370            end
371    
372       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
373         write (in the respective .cm-file):
374            a.sml (lambdasplitting:off)
375       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
376            a.sml (lambdasplitting:on)
377       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
378         <a> is a decimal non-negative integer):
379            a.sml (lambdasplitting:<a>)
380       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
381            a.sml (lambdasplitting:infinity)
382       - To use the global default for a.sml:
383            a.sml (lambdasplitting:default)
384         or simply
385            a.sml
386    
387    ----------------------------------------------------------------------
388    Name: Allen Leung
389    Date: 2001/05/04 01:57:00 EDT
390    Tag: leunga-20010504-sync
391    Description:
392    
393      MLRISC features.
394    
395      1. Fix to CMPXCHG instructions.
396      2. Changed RA interface to allow annotations in callbacks.
397      3. Added a new method to the stream interface to allow annotations updates.
398    
399    ----------------------------------------------------------------------
400    Name: Matthias Blume
401    Date: 2001/05/01 11:45:00 EDT
402    Tag: blume-20010501-pcedittmp
403    Description:
404    
405    Changed install.sh to use the current working directory instead of
406    /usr/tmp for a temporary file (pcedittmp).  The previous choice
407    of /usr/tmp caused trouble with MacOS X because of file premission
408    problems.
409    
410    ----------------------------------------------------------------------
411    Name: Matthias Blume
412    Date: 2001/04/20 11:10:00 EDT
413    Tag: blume-20010420-inMLflag
414    Description:
415    
416     - added vp_limitPtrMask to vproc-state.h
417       (for use by the raw-C-calls mechanism to implement proper interrupt
418        handling)
419     - made the ML compiler aware of various data-structure offsets so it
420       can generate code for accessing the vp_inML flag and vp_limitPtrMask
421     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
422       raw C-calls
423    
424    ----------------------------------------------------------------------
425    Name: Lal George
426    Date: 2001/04/20 09:15:28 EDT
427    Tag: george-20010420-macosX
428    Description:
429    
430     - Changes to port to Mac OS X; Darwin.
431    
432     - In the process I found that sqrt was broken on the PPC, because the
433       fsqrt instruction is not implemented.
434    
435    ----------------------------------------------------------------------
436    Name: Matthias Blume
437    Date: 2001/04/18 12:45:00 EDT
438    Tag: blume-20010418-ccalls
439    Description:
440    
441     - fixed two off-by-4 errors in the x86-specific c-calls implementation
442       (this bug prevented structure arguments containing pointers from being
443        passed correctly)
444     - changed the raw-C-call code in mlriscGen.sml in such a way that
445       structure arguments are represented as a pointer to the beginning
446       of the structure (instead of having a series of synthesized arguments,
447       one for each structure member)
448    
449     - made makeml script's verbosity level configurable via environment
450       variable (MAKEML_VERBOSITY)
451    
452     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
453       in rawmem-x86.sml; we are now using the real thing
454    
455    ----------------------------------------------------------------------
456    Name: Matthias Blume
457    Date: 2001/03/22 16:25:00 EST
458    Tag: blume-20010322-bootfiles
459    Description:
460    
461    Created a new set of bootfiles (for your automatic installation convenience).
462    
463    ----------------------------------------------------------------------
464    Name: Matthias Blume
465    Date: 2001/03/22 15:10:00 EST
466    Tag: blume-20010322-rawmem-parcm
467    Description:
468    
469    1. All "raw memory access" primitives for the new FFI are implemented now
470       (at least on the x86).
471    2. Some further cleanup of CM's parallel make mechanism.
472    
473    ----------------------------------------------------------------------
474    Name: Matthias Blume
475    Date: 2001/03/19 17:53:00 EST
476    Tag: blume-20010319-parallel
477    Description:
478    
479    Parallel make (using compile servers) now works again.
480    
481    To this end, CM.stabilize and CMB.make have been modified to work in
482    two passes when compile servers are attached:
483       1. Compile everything, do not perform stabilization; this pass
484          uses compile servers
485       2. Stabilize everything; this pass does not use compile servers
486    If there are no compile servers, the two passes are combined into one
487    (as before).  Splitting the passes increases the inherent parallelism
488    in the dependency graph because the entire graph including all
489    libraries is available at the same time.  This, in turn, improves
490    server utilization.  The downside is that the master process will
491    have to do some extra work after compilation is done (because for
492    technical reasons it must re-read all the binfiles during stabilization).
493    
494    ----------------------------------------------------------------------
495    Name: Matthias Blume
496    Date: 2001/03/16 12:22:00 EST
497    Tag: blume-20010316-bootfiles
498    Description:
499    
500    Created a new set of bootfiles (for your automatic installation convenience).
501    
502    ----------------------------------------------------------------------
503    Name: Matthias Blume
504    Date: 2001/03/16 11:00:00 EST
505    Tag: blume-20010316-MLTREE-fixup
506    Description:
507    
508    This is a minor fixup for an (untagged) earlier commit by Allen.
509    (A file was missing).
510    
511    ----------------------------------------------------------------------
512    Name: Allen Leung
513    Date: Mon Mar  5 18:54:57 EST 2001
514    Tag: leunga-20010305-cut-support
515    
516    1. New support for alternative control-flow in MLTREE.
517       Currently we support
518    
519          FLOW_TO(CALL ...., [k1,...,kn])
520    
521       This is needed for 'cuts to' in C-- and try/handle-like constructs
522       in Moby
523    
524       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
525    
526    2. Register Allocator
527    
528       Changes in interface [from Fermin, John]
529    
530    3. Alpha 8-bit SLL support [Fermin]
531    
532    4. All architectures
533    
534       A new module (ClusterExpandCopies) for expanding parallel copies.
535    
536    ----------------------------------------------------------------------
537    Name: Allen Leung
538    Date: 2001/02/27 23:07:00 EST
539    Tag: leunga-20010227-minor-stuff
540    
541    1. Alpha bug fix for CMOVNE
542    2. Handle mltree COND(..,FCMP ...,...)
543    3. Bug fix in simplifier
544    
545    ----------------------------------------------------------------------
546    Name: Matthias Blume
547    Date: 2001/01/30 17:50:00 EST
548    Tag: blume-20010130-sync
549    Description:
550    
551    This is just a minor update to sync my devel branch with the main brach.
552    The only visible change is the addition of some README files.
553    
554    ----------------------------------------------------------------------
555    Name: Matthias Blume
556    Date: 2001/01/12 23:30:00 JST
557    Tag: blume-20010112-bootfiles
558    Description:
559    
560    Made a new set of bootfiles that goes with the current state of the
561    repository.
562    
563    ----------------------------------------------------------------------
564    Name: Matthias Blume
565    Date: 2001/01/12 21:20:00 JST
566    Tag: blume-20010112-sync
567    Description:
568    
569    I am just flushing out some minor changes that had accumulated in
570    my private branch in order to sync with the main tree.  (This is
571    mainly because I had CVS trouble when trying to merge _into_ my
572    private branch.)
573    
574    Most people should be completely unaffected by this.
575    
576    ----------------------------------------------------------------------
577    Name: Allen Leung
578    Date: Thu Jan 11 21:03:00 EST 2001
579    Tag: leunga-20010111-labexp=mltree
580    Description:
581    
582    1.  Removed the type LabelExp and replace it by MLTree.
583    2.  Rewritten mltree-simplify with the pattern matcher tool.
584    3.  There were some bugs in alpha code generator which would break
585        64-bit code generation.
586    4.  Redo the tools to generate code with the
587    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
588        are now generated by perl scripts.
589    
590    ----------------------------------------------------------------------
591    Name: Matthias Blume
592    Date: 2001/01/10 21:55:00 JST
593    Tag: blume-20010110-rcc
594    Description:
595    
596    The RCC stuff now seems to work (but only on the x86).
597    This required hacking of the c-calls interface (and -implementation) in
598    MLRISC.
599    
600    Normal compiler users should be unaffected.
601    
602    ----------------------------------------------------------------------
603    Name: Matthias Blume
604    Date: 2001/01/09 01:20:00 JST
605    Tag: blume-20010109-rcc
606    Description:
607    
608    This is a fairly big patch, flushing out a large number of pending
609    changes that I made to my development copy over the last couple of days.
610    
611    Of practical relevance at this moment is a workaround for a pickling
612    bug that Allen ran into the other day.  The cause of the bug itself is
613    still unknown and it might be hard to fix it properly, but the
614    workaround has some merits of its own (namely somewhat reducing pickling
615    overhead for certain libraries).  Therefore, I think this solution should
616    be satisfactory at this time.
617    
618    The rest of the changes (i.e., the vast majority) has to do with my
619    ongoing efforts of providing direct support for C function calls from
620    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
621    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
622    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
623    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
624    handle RCC.
625    
626    The last part (the changes to mlriscGen.sml) are still known to be
627    wrong on the x86 and not implemented on all other architectures.  But
628    the infrastructure is in place. I had to change a few functor
629    signatures in the backend to be able to route the CCalls interface
630    from MLRISC there, and I had to specialize the mltree type (on the
631    x86) to include the necessary extensions. (The extensions themselves
632    were already there and redy to go in MLRISC/x86).
633    
634    Everything should be very happy as soon as someone helps me with
635    mlriscGen.sml...
636    
637    In any case, nothing of this should matter to anyone as long as the
638    new primop is not being used (which is going to be the case unless you
639    find it where I hid it :). The rest of the compiler is completely
640    unaffected.
641    
642    ----------------------------------------------------------------------
643    Name: Matthias Blume
644    Date: 2001/01/05 00:30:00 JST
645    Tag: blume-20010105-primops
646    Description:
647    
648    Added some experimental support for work that I am doing right now.
649    These changes mostly concern added primops, but there is also a new
650    experimental C library in the runtime system (but currently not enabled
651    anywhere except on Linux/X86).
652    
653    In the course of adding primops (and playing with them), I discovered that
654    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
655    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
656    To recover, I made type information in INL_PRIM mandatory and changed
657    prim.sml as well as built-in.sml accordingly.  The InLine structure now
658    has complete, correct type information (i.e., no bottom types).
659    
660    Since all these changes mean that we need new binfiles, I also bumped the
661    version number to 110.32.1.
662    
663    ----------------------------------------------------------------------
664    Name: Matthias Blume
665    Date: 2000/12/30 22:10:00 JST
666    Tag: blume-20001230-various
667    Description:
668    
669    Added proxy libraries for MLRISC and let MLRISC libraries refer
670    to each other using path anchors.  (See CM manual for explanation.)
671    
672    Updated CM documentation.
673    
674    Fixed some bugs in CM.
675    
676    Implemented "proxy" libraries (= syntactic sugar for CM).
677    
678    Added "-quiet" option to makeml and changed runtime system accordingly.
679    
680    Added cleanup handler for exportML to reset timers and compiler stats.
681    
682    ----------------------------------------------------------------------
683    Name: Lal George
684    Date: 2000/12/22 22:22:58 EST 2000
685    Tag: Release_110_32
686    Description:
687    
688            Infinite precision used throughout MLRISC.
689            see MLRISC/mltree/machine-int.sig
690    
691    ----------------------------------------------------------------------
692    Name: Matthias Blume
693    Date: 2000/12/22 23:16:00 JST
694    Tag: blume-20001222-warn
695    Description:
696    
697    Corrected wording and formatting of some CM warning message which I
698    broke in my previous patch.
699    
700    ----------------------------------------------------------------------
701    Name: Matthias Blume
702    Date: 2000/12/22 21:20:00 JST
703    Tag: blume-20001222-anchorenv
704    Description:
705    
706    Fixed CM's handling of anchor environments in connection with CMB.make.
707    
708    ----------------------------------------------------------------------
709    Name: Matthias Blume
710    Date: 2000/12/22 13:15:00 JST
711    Tag: blume-20001222-cleanup
712    Description:
713    
714    Removed src/cm/ffi which does not (and did not) belong here.
715    
716    ----------------------------------------------------------------------
717    Name: Matthias Blume
718    Date: 2000/12/21 23:55:00 JST
719    Tag: blume-20001221-exn
720    Description:
721    
722    Probably most important: CM no longer silently swallows all exceptions
723    in the compiler.
724    Plus: some other minor CM changes.  For example, CM now reports some
725    sizes for generated binfiles (code, data, envpickle, lambdapickle).
726    
727    ----------------------------------------------------------------------
728    Name: Matthias Blume
729    Date: 2000/12/15 00:01:05 JST
730    Tag: blume-20001215-dirtool
731    Description:
732    
733    - "dir" tool added.
734    - improvements and cleanup to Tools structure
735    - documentation updates
736    
737    ----------------------------------------------------------------------
738    Name: Allen Leung
739    Date: Thu Dec 14 03:45:24 EST 2000
740    Description:
741    Tag:  leunga-20001214-int-inf
742    Description:
743    
744       In IntInf, added these standard functions, which are missing from our
745    implementation:
746    
747        andb : int * int -> int
748        xorb : int * int -> int
749        orb  : int * int -> int
750        notb : int -> int
751         <<   : int * word -> int
752        ~>>  : int * word -> int
753    
754       Not tested, I hope they are correct.
755    
756    ----------------------------------------------------------------------
757    Name: Allen Leung
758    Date: Fri Dec  8 19:23:26 EST 2000
759    Description:
760    Tag:  leunga-20001208-nowhere
761    Description:
762    
763      Slight improvements to the 'nowhere' tool to handle OR-patterns,
764    to generate better error messages etc.  Plus a brief manual.
765    
766    ----------------------------------------------------------------------
767    Name: Lal George
768    Date: 2000/12/08 09:54:02 EST 2000
769    Tag: Release_110_31
770    Description:
771    
772    - Version 110.31
773    ----------------------------------------------------------------------
774    Name: Allen Leung
775    Date: Thu Dec  7 22:01:04 EST 2000
776    Tag:  leunga-20001207-cell-monster-hack
777    Description:
778    
779    Major MLRISC internal changes.  Affect all clients.
780    Summary:
781    
782    1.  Type CELLS.cell = int is now replaced by a datatype.
783        As a result, the old regmap is now gone.  Almost all interfaces
784        in MLRISC change as a consequence.
785    
786    2.  A new brand version of machine description tool (v3.0) that generates
787        modules expecting the new interface.  The old version is removed.
788    
789    3.  The RA interface has been further abstracted into two new functors.
790        RISC_RA and X86RA.  These functors have much simpler interfaces.
791        [See also directory MLRISC/demo.]
792    
793    4.  Some other new source->source code generation tools are available:
794    
795        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
796        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
797           I use this tool to generate the peephole optimizers---with the new
798           cell type changes, peephole rules are becoming difficult to write
799           without conditional pattern matching.
800    
801    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
802        cover the entire MLRISC source tree so many things broke.
803    
804    6.  CM files have been moved to the subdirectory MLRISC/cm.
805        They are moved because there are a lot of them and they clutter up the
806        root dir.
807    
808    7.  More detailed documentation to come...
809    
810        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
811        a makeml -rebuild first.  This is because of other other
812        changes that Matthias has made (see below).
813    
814    
815    ----------------------------------------------------------------------
816    Name: Matthias Blume
817    Date: 2000/11/30 23:12:00 JST
818    Tag: blume-20001130-filereorg
819    Description:
820    
821    Some manual updates and some file reorganizations in CM.
822    
823    ----------------------------------------------------------------------
824    Name: Matthias Blume
825    Date: 2000/11/24 17:45:00 JST
826    Tag: blume-20001124-link
827    Description:
828    
829    Drastically improved link traversal code for the case that the dynamic
830    value was already loaded at bootstrap time.  As a result, CM and CMB
831    now both load blazingly fast -- even on a very slow machine.  Also,
832    memory consumption has been further reduced by this.
833    
834    Warning: The format of the PIDMAP file has changed.  THerefore, to
835    bootstrap you have to do this:
836    
837    1. Run CMB.make
838    2. Make a symbolic link for the boot directory:
839         ln -s sml.boot.ARCH-OS xxx
840    3. "Rebuild" the boot directory:
841         ./makeml -boot xxx -rebuild sml ; rm xxx
842    4. Boot normally:
843          ./makeml
844    
845    ----------------------------------------------------------------------
846    Name: Matthias Blume
847    Date: 2000/11/21 21:20:00 JST
848    Tag: blume-20001121-tools
849    Description:
850    
851    Continued hacking on autoloading problem -- with success this time.
852    Also changed tool-plugin mechanism.  See new CM manual.
853    
854    ----------------------------------------------------------------------
855    Name: Matthias Blume
856    Date: 2000/11/19 14:30:00 JST
857    Tag:  blume-20001119-autoload
858    Description:
859    
860    Some hacking to make autoloading faster.  Success for CMB, no success
861    so far for CM.  There is a reduced structure CM' that autoloads faster.
862    (This is a temporary, non-documented hack to be eliminated again when
863    the general problem is solved.)
864    
865    ----------------------------------------------------------------------
866    Name: Matthias Blume
867    Date: 2000/11/17 14:10:00 JST
868    Tag: blume-20001117-pickle-lib
869    Description:
870    
871    1. Eliminated comp-lib.cm
872    2. Made pickle-lib.cm
873    3. Eliminated all uses of intset.sml (from comp-lib.cm)
874    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
875       equivalent constructs from smlnj-lib.cm (INtHashTable).
876    5. Point 4. also goes for those uses of intmap.* in MLRISC.
877       Duplicated intmap modules thrown out.
878    6. Hunted down all duplicated SCC code and replaced it with
879       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
880    7. Rewrote Feedback module.
881    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
882       should be thrown out and equivalent modules from smlnj-lib.cm
883       should be used (IntRedBlackSet, IntListSet, ...).
884    
885    Confirmed that compiler compiles to fixpoint.
886    
887    ----------------------------------------------------------------------
888    Name: Allen Leung
889    Date: 2000/11/10 18:00:00
890    Tag: leunga-20001110-new-x86-fp
891    
892    A new x86 floating point code generator has been added.
893    By default this is turned off.  To turn this on, do:
894    
895        CM.autoload "$smlnj/compiler.cm";
896        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
897    
898    Changes:
899    
900    1.  Changed FTAN to FPTAN so that the assembly output is correct.
901    2.  Changed the extension callback for FTANGENT to generate:
902    
903              fptan
904              fstp  %st(0)
905        instead of
906              fptan
907              fstpl ftempmem
908    
909    3.  Numerous assembly fixes for x86.
910    
911    5.  Cleaned up the machine code output module x86/x86MC.sml and added
912        support for a whole bunch of instructions and addressing modes:
913    
914          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
915          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
916          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
917          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
918          fxch %st(n)
919          fld %st(n)
920          fst %st(n)
921          fst mem
922          fstp %st(n)
923          fucom %st(n)
924          fucomp %st(n)
925    
926        All these are now generated when the fast fp mode is turned on.
927    
928    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
929    
930    ----------------------------------------------------------------------
931    Name: Matthias Blume
932    Date: 2000/11/09 11:20:00 JST
933    Tag: blume-20001109-scc
934    Description:
935    
936    Eliminated some code duplication:
937    
938    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
939       (Otherwise the functor is useless.)
940    2. Used GraphSCCFn where SCCUtilFun was used previously.
941    3. Got rid of SCCUtilFun (in comp-lib.cm).
942    
943    ----------------------------------------------------------------------
944    Name: Lal George
945    Date: 2000/11/06 09:02:21 EST 2000
946    Tag: Release_110_30
947    Description:
948    
949    - Version 110.30
950    ----------------------------------------------------------------------
951    Name: Matthias Blume
952    Date: 2000/11/04 14:45:00
953    Tag: blume-20001104-mlbuild
954    Description:
955    
956    - Made ml-build faster on startup.
957    - Documentation fixes.
958    
959    ----------------------------------------------------------------------
960    Name: Matthias Blume
961    Date: 2000/11/02 17:00:00 JST
962    Tag: blume-20001102-condcomp
963    Description:
964    
965    - Small tweaks to pickler -- new BOOTFILES!
966    - Version bumped to 110.29.2.
967    - Added conditional compilation facility to init.cmi (see comment there).
968    ----------------------------------------------------------------------
969    Name: Allen Leung
970    Date: 2000/10/23 19:31:00
971    Tag: leunga-20001023-demo-ra
972    
973    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
974    2. Test programs for the graph library updated
975    3. Some new MLRISC demo programs added
976    
977    ----------------------------------------------------------------------
978    Name: Matthias Blume
979    Date: 2000/08/31 22:15:00 JST
980    Tag: blume-20001017-errmsg
981    Description:
982    
983    More error message grief: Where there used to be no messages, there
984    now were some that had bogus error regions.  Fixed.
985    
986    ----------------------------------------------------------------------
987    Name: Matthias Blume
988    Date: 2000/08/31 17:30:00 JST
989    Tag: blume-20001017-v110p29p1
990    Description:
991    
992    I made a version 110.29.1 with new bootfiles.
993    
994    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
995              CM documentation changes and a small bugfix in CM's error reporting.
996    
997    ----------------------------------------------------------------------
998    Name: Lal George
999    Date: 2000/09/27 14:42:35 EDT
1000    Tag: george-20000927-nodestatus
1001    Description:
1002    
1003    Changed the type of the nodestatus, so that:
1004    
1005            SPILLED(~1)             is now SPILLED
1006            SPILLED(m) where m>=0   is now MEMREG(m)
1007            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
1008    
1009    ----------------------------------------------------------------------
1010    Name: Matthias Blume
1011    Date: 2000/09/07 14:45:00 JST
1012    Tag: blume-20000907-cmerrmsg
1013    Description:
1014    
1015    Small tweak to CM to avoid getting ML syntax error messages twice.
1016    
1017    ----------------------------------------------------------------------
1018    Name: Matthias Blume
1019    Date: 2000/08/31 18:00:00 JST
1020    Tag: blume-20000831-cvsbootfiles
1021    Description:
1022    
1023    New URL for boot files (because the 110.29 files on the BL server do
1024    now work correctly with my updated install scripts for yacc and lex).
1025    
1026    ----------------------------------------------------------------------
1027    Name: Matthias Blume
1028    Date: 2000/08/08 12:33:00 JST
1029    Tag: blume-20000808-manual
1030    Description:
1031    
1032    Tiny update to CM manual.
1033    
1034    ----------------------------------------------------------------------
1035    Name: Allen Leung
1036    Date: 2000/08/7 19:31:00
1037    Tag: leunga-20000807-a-whole-bunch-of-stuff
1038    
1039      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
1040    and Mobdy.
1041    
1042    1.  x86
1043    
1044       a.  Fixes to peephole module by John and Dan.
1045       b.  Assembly fix to SETcc by Allen.
1046       c.  Fix to c-call by John.
1047       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
1048       e.  Instruction selection optimization to SETcc as suggested by John.
1049    
1050           For example,
1051    
1052            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
1053    
1054           should generate:
1055    
1056            MOVL a, x
1057            SUBL b, x
1058            SHRL 31, x
1059    
1060    2.  IR stuff
1061    
1062         A bunch of new DJ-graph related algorithms added.  These
1063         speed up SSA construction.
1064    
1065    3.  SSA + Scheduling
1066    
1067         Added code for SSA and scheduling to the repository
1068    
1069    ----------------------------------------------------------------------
1070    Name: Lal George
1071    Date: 2000/07/27 11:53:14 EDT
1072    
1073    Tag: lal-20000727-linux-ppc
1074    Description:
1075    
1076     Made changes to support Linux PPC.
1077     p.s. I have confirmation that the 110.29 boot files work fine.
1078    
1079    ----------------------------------------------------------------------
1080    Name: Matthias Blume
1081    Date: 2000/07/27 17:40:00 JST
1082    Tag: blume-20000727-scripts
1083    Description:
1084    
1085    !!!! WARNING !!!!
1086    You must recompile the runtime system!
1087    !!!! WARNING !!!!
1088    
1089    This is basically another round of script-enhancements:
1090    
1091    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
1092       and undefine CM preprocessor symbols.
1093    
1094    2. ml-build avoids generating a new heap image if it finds that the
1095       existing one is still ok.  (The condition is that no ML file had to
1096       be recompiled and all ML files are found to be older that the heap
1097       file.)
1098    
1099       To make this work smoothly, I also hacked the runtime system as
1100       well as SMLofNJ.SysInfo to get access to the heap image suffix
1101       (.sparc-solaris, ...) that is currently being used.
1102    
1103       Moreover, the signature of CM.mk_standalone has changed.  See the
1104       CM manual.
1105    
1106    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
1107       CM manual for details.)
1108    
1109    4. More CM manual updates:
1110        - all of the above has been documented.
1111        - there is now a section describing the (CM-related) command line
1112          arguments that are accepted by the "sml" command
1113    
1114    ----------------------------------------------------------------------
1115    Name: Matthias Blume
1116    Date: 2000/07/25 16:20:00 JST
1117    Tag: blume-20000725-makedepend
1118    Description:
1119    
1120    Added a script called ml-makedepend.  This can be used in makefiles
1121    for Unix' make in a way very similar to the "makedepend" command for
1122    C.
1123    
1124    The script internally uses function CM.sources.
1125    
1126    Synopsis:
1127    
1128        ml-makedepend [-f makefile] cmfile targetname
1129    
1130    The default for the makefile is "makefile" (or "Makefile" should
1131    "makefile" not exist).
1132    
1133    ml-makedepend adds a cmfile/targetname-specific section to this
1134    makefile (after removing the previous version of this section).  The
1135    section contains a single dependency specification with targetname on
1136    the LHS (targetname is an arbitrary name), and a list of files derived
1137    from the cmfile on the RHS.  Some of the files on the RHS are
1138    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
1139    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
1140    path names.  The makefile writer is responsible for making sure that
1141    these variables have correct at the time "make" is invoked.
1142    
1143    ----------------------------------------------------------------------
1144    Name: Matthias Blume
1145    Date: 2000/07/22 23:30:00 JST
1146    Tag: blume-20000722-urlupdate
1147    Description:
1148    
1149    Changed BOOT and config/srcarchiveurl to point to BL server:
1150    
1151        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
1152    
1153    ----------------------------------------------------------------------
1154    Name: Matthias Blume
1155    Date: 2000/07/18 18:00:00 JST
1156    Tag: blume-20000718-Version_110_29
1157    Description:
1158    
1159    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
1160    
1161    2. Updated config/version to 110.29
1162    
1163    3. Updated config/srcarchiveurl
1164    
1165    3. New boot files!
1166       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
1167    
1168    ----------------------------------------------------------------------
1169    Name: Matthias Blume
1170    Date: 2000/07/11 13:58:00 JST
1171    Tag: blume-20000711-doctypo
1172    Description:
1173    
1174    Fixed a few typos in CM manual.
1175    
1176    ----------------------------------------------------------------------
1177    Name: Allen Leung
1178    Date: 2000/06/15 00:38:00
1179    Tag: leunga-20000704-sparc-x86
1180    
1181    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
1182    2. fix to x86 RET bug [found by Dan Grossman]
1183    3. sparc assembly bug fix for ticc instructions [found by Fermin]
1184    
1185       Affects c-- and moby only
1186    
1187    ----------------------------------------------------------------------
1188    Name: Matthias Blume
1189    Date: 2000/07/04 15:26:00
1190    Tag: blume-20000704-trigger
1191    Description:
1192    
1193    1. Improvements to CM manual.
1194    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
1195       of getting a back-trace.  The function, when called, raises an
1196       internal exception which explicitly carries the full back-trace history,
1197       so it is unaffected by any intervening handle-raise pairs ("trivial"
1198       or not).  The interactive loop will print that history once it arrives
1199       at top level.
1200       Short of having all exceptions implicitly carry the full history, the
1201       recommended way of using this facility is:
1202         - compile your program with instrumentation "on"
1203         - run it, when it raises an exception, look at the history
1204         - if the history is "cut off" because of some handler, go and modify
1205           your program so that it explicitly calls BTrace.trigger
1206         - recompile (still instrumented), and rerun; look at the full history
1207    
1208    ----------------------------------------------------------------------
1209    Name: Matthias Blume
1210    Date: 2000/07/03 15:36:00 JST
1211    Tag: blume-20000702-manual
1212    Description:
1213    
1214    Small corrections and updates to CM manual.
1215    
1216    ----------------------------------------------------------------------
1217    Name: Matthias Blume
1218    Date: 2000/06/29 16:04:00 JST
1219    Tag: blume-20000629-yacctool
1220    Description:
1221    
1222    Changes:
1223    
1224    1. Class "mlyacc" now takes separate arguments to pass options to
1225       generated .sml- and .sig-files independently.
1226    2. Corresponding CM manual updates.
1227    3. BTrace module now also reports call sites.  (However, for loop clusters
1228       it only shows from where the cluster was entered.)  There are associated
1229       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1230    
1231    ----------------------------------------------------------------------
1232    Name: Matthias Blume
1233    Date: 2000/06/27 16:51:00 JST
1234    Tag: blume-20000627-noweb
1235    Description:
1236    
1237    Changes:
1238    
1239     1. Implemented "subdir" and "witness" options for noweb tool.
1240        This caused some slight internal changes in CM's tool implementation.
1241     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
1242        some remaining issues from earlier path anchor changes.
1243     3. Updated CM manual accordingly.
1244    
1245     4. Changed implementation of back-tracing so that I now consider it
1246        ready for prime-time.
1247    
1248        In particular, you don't have to explicitly trigger the back-trace
1249        anymore.  Instead, if you are running BTrace-instrumented code and
1250        there is an uncaught exception (regardless of whether or not it was
1251        raised in instrumented code), the top-level evalloop will print
1252        the back-trace.
1253    
1254        Features:
1255    
1256          - Instrumented and uninstrumented code work together seemlessly.
1257            (Of course, uninstrumented code is never mentioned in actual
1258             back-traces.)
1259    
1260          - Asymptotic time- and space-complexity of instrumented code is
1261            equal to that of uninstrumented code.  (This means that
1262            tail-recursion is preserved by the instrumentation phase.)
1263    
1264          - Modules whose code has been instrumented in different sessions
1265            work together without problem.
1266    
1267          - There is no penalty whatsoever on uninstrumented code.
1268    
1269          - There is no penalty on "raise" expressions, even in
1270            instrumented code.
1271    
1272        A potential bug (or perhaps it is a feature, too):
1273    
1274          A back-trace reaches no further than the outermost instrumented
1275          non-trivial "raise".  Here, a "trivial" raise is one that is the
1276          sole RHS of a "handle" rule.  Thus, back-traces reach trough
1277    
1278               <exp> handle e => raise e
1279    
1280          and even
1281    
1282               <exp> handle Foo => raise Bar
1283    
1284          and, of course, through
1285    
1286               <exp> handle Foo => ...
1287    
1288         if the exception was not Foo.
1289    
1290         Back-traces always reach right through any un-instrumented code
1291         including any of its "handle" expressions, trivial or not.
1292    
1293       To try this out, do the following:
1294    
1295         - Erase all existing binfiles for your program.
1296           (You may keep binfiles for those modules where you think you
1297            definitely don't need back-tracing.)
1298         - Turn on back-trace instrumentation:
1299              SMLofNJ.Internals.BTrace.mode (SOME true);
1300         - Recompile your program.  (I.e., run "CM.make" or "use".)
1301         - You may now turn instrumentation off again (if you want):
1302              SMLofNJ.Internals.BTrace.mode (SOME false);
1303         - Run your program as usual.  If it raises an exception that
1304           reaches the interactive toplevel, then a back-trace will
1305           automatically be printed.  After that, the toplevel loop
1306           will print the exception history as usual.
1307    
1308    ----------------------------------------------------------------------
1309    Name: Matthias Blume
1310    Date: 2000/06/26 09:56:46 JST
1311    Tag: blume-20000626-setup
1312    Description:
1313    
1314    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1315          ML code before and after compiling a file (e.g., to set compiler
1316          flags)
1317    
1318    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1319              - associated changes to btrace.sml (BTrace instrumentation pass)
1320              - cleaner implementation of btimp.sml (BTrace tracing and report
1321                module)
1322    
1323    CM manual: * new path encoding documented
1324               * description of setup-parameter to "sml" added
1325    
1326    The biggest user-visible change to back-tracing is that it is no
1327    longer necessary to compile all traced modules within the same
1328    session.  (This was a real limitation.)
1329    
1330    ----------------------------------------------------------------------
1331    Name: Matthias Blume
1332    Date: 2000/06/24 12:40:00 JST
1333    Tag: blume-20000624-startup
1334    Description:
1335    
1336    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1337    bit too often -- to put it mildly. :)
1338    
1339    ----------------------------------------------------------------------
1340    Name: Matthias Blume
1341    Date: 2000/06/23 18:20:00 JST
1342    Tag: blume-20000623-btrace
1343    Description:
1344    
1345    This updates adds a backtrace facility to aid programmers in debugging
1346    their programs.  This involves the following changes:
1347    
1348    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1349       keeping track of the current call stack.  When programs are compiled
1350       in a special mode, the compiler will insert calls to these hooks
1351       into the user program.
1352       "Hook" means that it is possible for different implementations of
1353       back-tracing to register themselves (at different times).
1354    
1355    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1356       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1357       It can be turned on using this call:
1358         SMLofNJ.Internals.BTrace.mode (SOME true);
1359       Turning it off again:
1360         SMLofNJ.Internals.BTrace.mode (SOME false);
1361       Querying the current status:
1362         SMLofNJ.Internals.BTrace.mode NONE;
1363       Annotated programs are about twice as big as normal ones, and they
1364       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1365       where all hooks do nothing).  The slowdown with a plugin that is
1366       actually useful (such as the one supplied by default) is even greater,
1367       but in the case of the default plugin it is still only an constant
1368       factor (amortized).
1369    
1370    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1371       with a sub-structure BTrace for controlling back-tracing.  In particular,
1372       the above-mentioned function "mode" controls whether the annotation
1373       phase is invoked by the compiler.  Another important function is
1374       "trigger": when called it aborts the current execution and causes
1375       the top-level loop to print a full back-trace.
1376    
1377    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1378       for back-tracing.  It keeps track of the dynamic call stack and in
1379       addition to that it keeps a partial history at each "level" of that
1380       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1381       d and b (at separate times, dynamically), then the report will show:
1382    
1383       GOTO   d
1384             /c
1385       GOTO  \b
1386       CALL   a
1387    
1388       This shows that there was an initial non-tail call of a, then a
1389       tail-call to b or c, looping behavior in a cluster of functions that
1390       consist of b and c, and then a goto from that cluster (i.e., either from
1391       b or from c) to d.
1392    
1393       Note that (depending on the user program) the amount of information
1394       that the back-trace module has to keep track of at each level is bounded
1395       by a constant.  Thus, the whole implementation has the same asymptotical
1396       complexity as the original program (both in space and in time).
1397    
1398    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1399       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1400       which is raised by the "trigger" function mentioned above.
1401    
1402    Notes on usage:
1403    
1404    - Annotated code works well together with unannotated code:
1405    Unannotated calls simply do not show up at all in the backtrace.
1406    
1407    - It is not a good idea to let modules that were annotated during
1408    different sessions run at the same time.  This is because the compiler
1409    chooses small integers to identify individual functions, and there
1410    will be clashes if different modules were compiled in separate sessions.
1411    (Nothing will crash, and you will even be told about the clashes, but
1412    back-trace information will in general not be useful.)
1413    
1414    - Back-tracing can be confused by callcc and capture.
1415    
1416    - The only way of getting a back-trace right now is to explicitly
1417    invoke the "trigger" function from your user program.  Eventually, we
1418    should make every exception carry back-trace information (if
1419    available).  But since this creates more overhead at "raise"-time
1420    (similar to the current exnHistory overhead), I have not yet
1421    implemented this.  (The implementation will be rather easy.)  With
1422    exceptions carrying back-trace information, this facility will be even
1423    more useful because users don't need to modify their programs...
1424    
1425    - While it is possible to compile the compiler with back-trace
1426    annotations turned on (I did it to get some confidence in
1427    correctness), you must make absolutely sure that core.sml and
1428    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1429    be compiled with annotation because there is no core access yet, but
1430    if you compile btimp.sml with annotation, then the system will go into
1431    an infinite recursion and crash.)
1432    Since CM currently does not know about BTrace, the only way to turn
1433    annotations on and off for different modules of the compiler is to
1434    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1435    this is awkward and clumsy.
1436    
1437    Sample sessions:
1438    
1439    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1440    - SMLofNJ.Internals.BTrace.mode (SOME true);
1441    [autoloading]
1442    [autoloading done]
1443    val it = false : bool
1444    - structure X = struct
1445    -     fun main n = let
1446    -         fun a (x, 0) = d x
1447    -           | a (x, n) = b (x, n - 1)
1448    -         and b (x, n) = c (x, n)
1449    -         and c (x, n) = a (x, n)
1450    -         and d x = e (x, 3)
1451    -         and e (x, 0) = f x
1452    -           | e (x, n) = e (x, n - 1)
1453    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1454    -           | f n = n * g (n - 1)
1455    -         and g n = a (n, 3)
1456    -     in
1457    -         f n
1458    -     end
1459    - end;
1460    structure X : sig val main : int -> int end
1461    - X.main 3;
1462    *** BACK-TRACE ***
1463    GOTO   stdIn:4.2-13.20: X.main[2].f
1464    GOTO-( stdIn:4.2-13.20: X.main[2].e
1465    GOTO   stdIn:4.2-13.20: X.main[2].d
1466         / stdIn:4.2-13.20: X.main[2].a
1467         | stdIn:4.2-13.20: X.main[2].b
1468    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1469    CALL   stdIn:4.2-13.20: X.main[2].g
1470    GOTO   stdIn:4.2-13.20: X.main[2].f
1471    GOTO-( stdIn:4.2-13.20: X.main[2].e
1472    GOTO   stdIn:4.2-13.20: X.main[2].d
1473         / stdIn:4.2-13.20: X.main[2].a
1474         | stdIn:4.2-13.20: X.main[2].b
1475    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1476    CALL   stdIn:4.2-13.20: X.main[2].g
1477    GOTO   stdIn:4.2-13.20: X.main[2].f
1478    GOTO-( stdIn:4.2-13.20: X.main[2].e
1479    GOTO   stdIn:4.2-13.20: X.main[2].d
1480         / stdIn:4.2-13.20: X.main[2].a
1481         | stdIn:4.2-13.20: X.main[2].b
1482    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1483    CALL   stdIn:4.2-13.20: X.main[2].g
1484    GOTO   stdIn:4.2-13.20: X.main[2].f
1485    CALL   stdIn:2.15-17.4: X.main[2]
1486    -
1487    
1488    (Note that because of a FLINt bug the above code currently does not
1489    compile without BTrace turned on.)
1490    
1491    Here is another example, using my modified Tiger compiler:
1492    
1493    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1494    - SMLofNJ.Internals.BTrace.mode (SOME true);
1495    [autoloading]
1496    [autoloading done]
1497    val it = false : bool
1498    - CM.make "sources.cm";
1499    [autoloading]
1500    ...
1501    [autoloading done]
1502    [scanning sources.cm]
1503    [parsing (sources.cm):parse.sml]
1504    [creating directory CM/SKEL ...]
1505    [parsing (sources.cm):tiger.lex.sml]
1506    ...
1507    [wrote CM/sparc-unix/semant.sml]
1508    [compiling (sources.cm):main.sml]
1509    [wrote CM/sparc-unix/main.sml]
1510    [New bindings added.]
1511    val it = true : bool
1512    - Main.compile ("../testcases/merge.tig", "foo.out");
1513    *** BACK-TRACE ***
1514    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
1515    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1516    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1517    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1518    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1519    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1520    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
1521         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
1522    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
1523    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1524    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
1525    CALL   main.sml:1.18-118.4: Main.compile[2]
1526    -
1527    
1528    ----------------------------------------------------------------------
1529    Name: Matthias Blumen
1530    Date: 2000/06/21 18:00:00 JST
1531    Tag: blume-20000621-manual
1532    Description:
1533    
1534    CM manual update: Path environments documented.
1535    
1536    ----------------------------------------------------------------------
1537    Name: Matthias Blume
1538    Date: 2000/06/19 13:40:00
1539    Tag: blume-20000619-manual
1540    Description:
1541    
1542    CM manual and system/README update.  This only covers the fact that
1543    there are no more implicit anchors.  (Path environments and the "bind"
1544    option to "cm" have yet to be documented.)
1545    
1546    ----------------------------------------------------------------------
1547    Name: Matthias Blume
1548    Date: 2000/06/19 11:05:00 JST
1549    Tag: blume-20000619-chdir-bugfix
1550    Description:
1551    
1552    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1553    
1554    ----------------------------------------------------------------------
1555    Name: Matthias Blume
1556    Date: 2000/06/18 22:00:10 JST
1557    Tag: blume-20000618-implicit-anchors-really-gone
1558    Description:
1559    
1560    I updates the previous HISTORY entry where I forgot to mention that
1561    implicit anchors are no longer with us.
1562    
1563    The current update also gets rid of the (now useless) controller
1564    CM.Control.implicit_anchors.
1565    
1566    ----------------------------------------------------------------------
1567    Name: Matthias Blume
1568    Date: 2000/06/16 17:30:00 JST
1569    Tag: blume-20000616-anchorenv
1570    Description:
1571    
1572    This patch implements the long anticipated (just kidding :) "anchor
1573    environment" mechanism.  In the course of doing this, I also
1574    re-implemented CM's internal "SrcPath" module from scratch.  The new
1575    one should be more robust in certain boundary cases.  In any case, it
1576    is a lot cleaner than its predecessor (IMHO).
1577    
1578    This time, although there is yet another boot file format change, I
1579    kept the unpickler backward-compatible.  As a result, no new bootfiles
1580    are necessary and bootstrapping is straightforward.  (You cannot read
1581    new bootfiles into an old system, but the other way around is no
1582    problem.)
1583    
1584    Visible changes:
1585    
1586    ** 0. Implicit path anchors (without the leading $-symbol) are no
1587    longer recognized at all. This means that such path names are not
1588    illegal either.  For example, the name basis.cm simply refers to a
1589    local file called "basis.cm" (i.e, the name is an ordinary path
1590    relative to .cm-files directory).  Or, to put it differently, only
1591    names that start with $ are anchored paths.
1592    
1593    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1594    vanished.
1595    
1596    John (Reppy) had critizised this as soon as I originally proposed and
1597    implemented it, but at that time I did not really deeply believe
1598    him. :) Now I came full-circle because I need the $<singlearc> syntax
1599    in another place where it cannot be seen as an abbreviation for
1600    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1601    seems to mean (i.e., it "expands" into the corresponding anchor
1602    value).
1603    
1604    However, when paths are used as members in CM description files, it
1605    continues to be true that there must be at least another arc after the
1606    anchor.  This is now enforced separately during semantic analysis
1607    (i.e., from a lexical/syntactical point of view, the notation is ok.)
1608    
1609    ** 2. The "cm" class now accepts an option "bind".  The option's value
1610    is a sub-option list of precisely two items -- one labeled "anchor"
1611    and the other one labeled "value".  As you might expect, "anchor" is
1612    used to specify an anchor name to be bound, and "value" specifies what
1613    the anchor is being bound to.
1614    
1615    The value must be a directory name and can be given in either standard
1616    syntax (including the possibility that it is itself an anchored path)
1617    or native syntax.
1618    
1619    Examples:
1620    
1621       foo.cm (bind:(anchor:bar value:$mystuff/bar))
1622       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1623    
1624    and so on.
1625    
1626    The meaning of this is that the .cm-file will be processed with an
1627    augmented anchor environment where the given anchor(s) is/are bound to
1628    the given values(s).
1629    
1630    The rationale for having this feature is this: Suppose you are trying
1631    to use two different (already stable) libraries a.cm and b.cm (that
1632    you perhaps didn't write yourself).  Further, suppose each of these
1633    two libraries internally uses its own auxiliary library $aux/lib.cm.
1634    Normally you would now have a problem because the anchor "lib" can not
1635    be bound to more than one value globally.  Therefore, the project that
1636    uses both a.cm and b.cm must locally redirect the anchor to some other
1637    place:
1638    
1639       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1640       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1641    
1642    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1643    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
1644    
1645    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1646    you when you do so at the time of CM.stabilize).  Therefore, you can
1647    also use an anchored path as the value:
1648    
1649      a.cm (bind:(anchor:lib value:$a-lib))
1650      b.cm (bind:(anchor:lib value:$b-lib))
1651    
1652    Now you can globally configure (using the usual CM.Anchor.anchor or
1653    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1654    itself is always locally bound, setting it globally is no longer
1655    meaningful or necessary (but it does not hurt either).  In fact, "lib"
1656    can still be used as a global anchor for separate purposes.  As a
1657    matter of fact, one can locally define "lib" in terms of a global
1658    "lib":
1659    
1660      a.cm (bind:(anchor:lib value:$lib/a))
1661      b.cm (bind:(anchor:lib value:$lib/b))
1662    
1663    ** 3: The encoding of path names has changed.  This affects the way
1664    path names are shown in CM's progress report and also the internal
1665    protocol encoding used for parallel make.
1666    
1667    The encoding now uses one or more ':'-separated segments.  Each
1668    segments corresponds to a file that has been specified relative to the
1669    file given by its preceding segment.  The first segment is either
1670    relative to the CWD, absolute, or anchored.  Each segment itself is
1671    basically a Unix pathname; all segments but the first are relative.
1672    
1673    Example:
1674    
1675       $foo/bar/baz.cm:a/b/c.sml
1676    
1677    This path denotes the file bar/a/b/c.sml relative to the directory
1678    denoted by anchor "foo".  Notice that the encoding also includes
1679    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1680    paths are resolved relative to the .cm-files directory, so baz.cm must
1681    be ignored to get the "real" pathname.
1682    
1683    To make this fact more obvious, CM puts the names of such "virtual
1684    arcs" into parentheses when they appear in progress reports. (No
1685    parentheses will appear in the internal protocol encoding.)  Thus,
1686    what you really see is:
1687    
1688      $foo/bar/(baz.cm):a/b/c.sml
1689    
1690    I find this notation to be much more informative than before.
1691    
1692    Another new feature of the encoding is that special characters
1693    including parentheses, colons, (back)slashes, and white space are
1694    written as \ddd (where ddd is the decimal encoding of the character).
1695    
1696    *** The CM manual still needs to be updated.
1697    
1698    ----------------------------------------------------------------------
1699    Name: Allen Leung
1700    Date: 2000/06/15 00:38:00
1701    Tag: leunga-20000615-x86-peephole
1702    
1703    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1704    
1705    ----------------------------------------------------------------------
1706    Name: Matthias Blume
1707    Date: 2000/06/12 11:40:00
1708    Tag: blume-20000612-parmakefix
1709    Description:
1710    
1711    More cleanup after changing the file naming scheme: This time I
1712    repaired the parallel make mechanism for CMB.make which I broke earlier.
1713    
1714    ----------------------------------------------------------------------
1715    Name: Allen Leung
1716    Date: 2000/06/09 01:25:00
1717    Tag: leunga-20000609-various
1718    
1719    None of these things should affect normal SML/NJ operations
1720    
1721    1. Peephole improvements provided by Fermin (c--)
1722    2. New annotation DEFUSE for adding extra dependence (moby)
1723    3. New X86 LOCK instructions (moby)
1724    4. New machine description language for reservation tables (scheduling)
1725    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1726       trees etc.)
1727    6. I've changed the CM files so that they can work with versions
1728       110.0.6, 110.25 and 110.28
1729    
1730    ----------------------------------------------------------------------
1731    Name: Matthias Blume
1732    Date: 2000/06/09 12:40:00
1733    Tag: blume-20000609-log
1734    Description:
1735    
1736    - Removed all(?) remaining RCS Log entries from sources.
1737    
1738    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1739      anchored paths).
1740    
1741    ----------------------------------------------------------------------
1742    Name: Matthias Blume
1743    Date: 2000/06/07 17:00:00 JST
1744    Tag: blume-20000607-no-implicit-anchors
1745    Description:
1746    
1747    1. This update changes the default setting for
1748    CM.Control.implicit_anchors from true to false.  This means that
1749    implicit anchors are no longer permitted by default.  I also tried to
1750    make sure that nothing else still relies on implicit anchors.
1751    (This is the next step on the schedule towards a CM that does not even
1752    have the notion of implicit anchors anymore.)
1753    
1754    2. More CM manual updates.
1755    
1756    3. I managed to track down and fix the pickling bug I mentioned last
1757    time.  Because of the previously existing workaround, this entails no
1758    immediate practical changes.
1759    
1760    ----------------------------------------------------------------------
1761    Name: Matthias Blume
1762    Date: 2000/06/06 11:15:00 JST
1763    Tag: blume-20000606-lazierpickle
1764    Description:
1765    
1766    !!!! NEW BOOT FILES !!!!
1767    
1768    * The main purpose of this update is to make library pickles lazier in
1769    order to reduce the initial space penalty for autoloading a library.
1770    As a result, it is now possible to have $smlnj/compiler.cm
1771    pre-registered.  This should take care of the many complaints or
1772    inquiries about missing structure Compiler.  This required changes to
1773    CM's internal data structures and small tweaks to some algorithms.
1774    
1775    As a neat additional effect, it is no longer necessary (for the sake
1776    of lean heap image files) to distinguish between a "minimal" CM and a
1777    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1778    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1779    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1780    
1781    To make the life of compiler-hackers easier, "makeml" now also
1782    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1783    words, after you bootstrap a new sml for the first time, you will not
1784    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1785    around you will still have to do it, though.)
1786    
1787    * A second change consists of major updates to the CM manual.  There
1788    are now several appendices with summary information and also a full
1789    specification of the CM description file syntax.
1790    
1791    * In directory src/system I added the script "allcross".  This script
1792    invokes sml and cross-compiles the compiler for all supported
1793    architectures.  (Useful when providing a new set of boot files.)
1794    
1795    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1796    added a small tweak to pickle-util.sml to work around this problem,
1797    but it is not a proper fix yet.  I will investigate further.  (The
1798    effect of the bug was an inflation of library pickle size.)
1799    
1800    * Version number increased to 110.28.1 (to avoid compatibility problems).
1801    
1802    ----------------------------------------------------------------------
1803    Name: Allen Leung
1804    Date: 2000/05/25 17:28 EDT
1805    Tag: leunga-20000525-ra
1806    Description:
1807    
1808      Fixed a bug in freezing phase of the register allocator.
1809    
1810    ----------------------------------------------------------------------
1811    Name: Allen Leung
1812    Date: 2000/05/15 22:53 EDT
1813    Tag: leunga-20000515-alpha-x86-ra
1814    Description:
1815    
1816      1. Alpha
1817    
1818          Slight cleanup.  Removed the instruction SGNXL
1819    
1820      2. X86
1821    
1822          Added the following instructions to the instruction set:
1823    
1824            ROLx, RORx,
1825            BTx, BTSx, BTLx, BTRx,
1826            XCHGx, and variants with the LOCK prefix
1827    
1828      3. Register Allocation
1829    
1830          The module ra-rewrite-with-renaming has been improved.
1831    
1832      These have no effect on SML/NJ.
1833    
1834    ----------------------------------------------------------------------
1835    Name: Matthias Blume
1836    Date: 2000/05/15 16:20:00 JST
1837    Tag: blume-20000515-lightrebuild
1838    Description:
1839    
1840    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1841       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1842       defined.  In effect, the command will not build any cross-compiler
1843       backends and therefore finish more quickly.
1844    
1845       The "fixpt" script also takes a "-light" switch to be able to use
1846       this new facility while compiling for a fixpoint.
1847    
1848    2. I replaced all mentions of anchored paths in group owner specifications
1849       with simple relative paths (usually starting with "..").
1850       The rationale is that a library's internal workings should not be
1851       compromised by the lack of some anchor.  (An anchor is necessary
1852       for someone who wants to refer to the library by an anchored path,
1853       but it should not be necessary to build the same library in the first
1854       place.)
1855    
1856    3. I changed the way CM's tool mechanism determines the shell command
1857       string used for things like ml-yacc etc. so that it does not break
1858       when CM.Control.implicit_anchors is turned off.
1859    
1860    ----------------------------------------------------------------------
1861    Name: Matthias Blume
1862    Date: 2000/05/12 18:20:00 JST
1863    Tag: blume-20000512-ml-build
1864    Description:
1865    
1866    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1867    getting installed properly (by config/install.sh).
1868    
1869    ----------------------------------------------------------------------
1870    Name: Matthias Blume
1871    Date: 2000/05/12 17:30:00 JST
1872    Tag: blume-20000512-anchors
1873    Description:
1874    
1875    !!! NEW BOOT FILES !!!
1876    
1877    This change is in preparation of fading out support for "implicitly
1878    anchored path names".  I went through all sources and used the
1879    explicit (and relatively new) $-notation.  See system/README and the
1880    CM manual for more info on this.
1881    
1882    I also modified the anchoring scheme for some things such as "smlnj",
1883    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1884    anchors are more expressive: anchor name and first arc do not have to
1885    coincide.  This entails the following user-visible change:
1886    
1887    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1888    particular, when you fire up sml with a command-line argument, say,
1889    e.g.:
1890    
1891       sml '$smlnj/cmb.cm'
1892    
1893    At the ML toplevel prompt:
1894    
1895       CM.autoload "$smlnj/cmb.cm";
1896    
1897    There is also a new controller in CM.Control that can be used to turn
1898    off all remaining support for implicit anchors by saying:
1899    
1900        CM.autoload "$smlnj/
1901        #set CM.Control.implicit_anchors false;
1902    
1903    This causes CM to reject implicitly anchored paths.  This is (for the
1904    time being) less permissive than the "final" version where there will
1905    be no more such implicit anchors and relative paths will be just that:
1906    relative.
1907    
1908    The next step (version after next version?) will be to make the
1909    default for CM.Control.implicit_anchors false.  After the dust has
1910    settled, I can then produce the "final" version of this...
1911    
1912    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1913    
1914    ----------------------------------------------------------------------
1915    Name: Matthias Blume
1916    Date: 2000/05/11 16:30:00 JST
1917    Tag: blume-20000511-sources
1918    Description:
1919    
1920    The main change is that I added function CM.sources as a generalized
1921    version of the earlier CM.makedepend.  This entails the following
1922    additional changes:
1923    
1924      - CM.makedepend has been dropped.
1925    
1926      - CM manual has been updated.
1927    
1928      - TOOLS signature and API have been changed.
1929    
1930    ----------------------------------------------------------------------
1931    Name: Allen Leung
1932    Date: 2000/05/10 21:17 EDT
1933    Tag: leunga-20000510-moby-c--ssa
1934    Description:
1935    
1936      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1937    None of these affect SML/NJ.
1938    
1939    1. Register Allocation
1940    
1941        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1942           This module tries to remove local (i.e. basic block level) redundancies
1943           during spilling.
1944    
1945        b. A new framework for performing region based register allocation.
1946           Not yet entirely functional.
1947    
1948    2. X86
1949    
1950       a. DefUse for POP was missing the stack pointer [found by Lal]
1951       b. Reload for CALL was incorrect in X86Spill [found by John]
1952       c. Various fixes in X86Spill so that it can be used correctly for
1953          the new spilling module.
1954    
1955    3. SSA/IR
1956    
1957       a. New module ir/dj-dataflow.sml implements elimination based
1958          data flow analysis.
1959    
1960    4. MLRiscGen
1961    
1962       a. Fix for gc type annotation
1963    
1964    5. MDGen
1965    
1966       Various fixes for machine description -> ml code translation.  For ssa
1967       only.
1968    
1969    ----------------------------------------------------------------------
1970    Name: Allen Leung
1971    Date: 2000/05/08 22:17 EDT
1972    Tag: leunga-20000508-labexp
1973    Description:
1974    
1975      Fermin has found a few assembly problems with constant expressions
1976      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1977      which choke on dumb assemblers.  This is his fix.
1978    
1979    ----------------------------------------------------------------------
1980    Name: Dave MacQueen
1981    Date: 2000/04/09 14:00 EDT
1982    Tag: dbm-20000502-Version_110_28
1983    Description:
1984    
1985    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1986    
1987    2. Updated config/version to 110.28
1988    
1989    3. Updated config/srcarchiveurl
1990    
1991    3. New boot files!
1992       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1993    
1994    ----------------------------------------------------------------------
1995    Name: Matthias Blume
1996    Date: 2000/05/01 19:05:00 JST
1997    Tag: blume-20000501-noweb
1998    Description:
1999    
2000    A new noweb tool has been added.  The existing system is entirely
2001    unaffected by this, but some CM users have asked for renewed noweb
2002    support.  Everything is documented in the CM manual.
2003    
2004    New (plugin) libraries:
2005    
2006       noweb-tool.cm
2007       nw-ext.cm
2008    
2009    ----------------------------------------------------------------------
2010    Name: Dave MacQueen
2011    Date: 2000/04/30 12:40PM EDT
2012    Tag: dbm-20000430-bug_fixes
2013    Description:
2014    
2015    1. Fix for bug 1498
2016       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
2017       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
2018         added toRealArray function
2019       smlnj/src/compiler/MiscUtil/print/ppobj.sml
2020         added check for tag Obj.RealArray to array printing case in ppObj
2021    
2022    2. Fix for bug 1510
2023       smlnj/src/compiler/Semant/types/typesutil.sml
2024         fixed definition of dummyargs (used by equalTycon) so that
2025         dummy args are distinct types
2026    
2027    ----------------------------------------------------------------------
2028    Name: Matthias Blume
2029    Date: 2000/04/30 01:00:00 JST
2030    Tag: blume-20000430-versions
2031    Description:
2032    
2033    1. CM version numbering added.  This is an implementation of Lal's
2034       proposal for adding version numbers and version checking to .cm
2035       files.  Lal said that his proposal was just that -- a proposal.
2036       For the time being I went ahead and implemented it so that people
2037       can comment on it.  Everything is completely backward-compatible
2038       (except for the stable library format, i.e., new bootfiles!).
2039    
2040       As usual, see the CM manual for details.
2041    
2042    2. An alternative syntax for anchored paths has been implemented.
2043       Dave has recently voiced the same concerns that I had when I did
2044       this, so there should be some support.  My take is that eventually
2045       I will let support for the current syntax (where anchors are
2046       "implicit") fade out in favor of the new, explicit syntax.
2047       In order to be backward-compatible, both old and new syntax are
2048       currently supported.
2049    
2050       Again, see the CM manual for details.
2051    
2052    3. Parallel make is trying to be slightly smarter:  When the master
2053       process finds a "bottleneck", i.e., when there is only one
2054       compilation unit that can be compiled and everybody else is
2055       waiting on it, then it will simply compile it directly instead
2056       of clumsily telling one of the slaves to do it.
2057    
2058    4. Support for "unsharing" added.  This is necessary in order to be
2059       able to have two different versions of the same library running
2060       at the same time (e.g., for trying out a new MLRISC while still
2061       having the old MLRISC linked into the current compiler, etc.)
2062       See the CM manual.
2063    
2064    5. Simple "makedepend" functionality added for generating Makefile
2065       dependency information.  (This is rather crude at the moment.
2066       Expect some changes here in the future.)
2067    
2068    6. ".fun" added as a recognized suffix for ML files. Also documented
2069       explicitly in the manual that the fallback behavior (unknown suffix
2070       -> ML file) is not an official feature!
2071    
2072    7. Small changes to the pickler for stable libraries.
2073    
2074    8. Several internal changes to CM (for cleanup/improvement).
2075    
2076    
2077    !!!! NEW BINFILES !!!!
2078    
2079    ----------------------------------------------------------------------
2080    Name: Matthias Blume
2081    Date: 2000/04/28 17:30:00 JST
2082    Tag: blume-20000428-pathconfig
2083    Description:
2084    
2085    1. I changed config/install.sh to remove duplicate entries from the
2086       lib/pathconfig file at the end.  Moreover, the final version of
2087       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
2088       in src/system/installml.
2089    
2090    2. The config/install.sh script now consistently uses relative
2091       pathnames in lib/pathconfig whenever the anchor is in the lib
2092       directory.  (So far this was true for the libraries that come
2093       pre-compiled and bundled as part of the bootfiles but not for
2094       libraries that are compiled by the script itself.)
2095    
2096    ----------------------------------------------------------------------
2097    Name: Matthias Blume
2098    Date: 2000/04/26 13:10:00 JST
2099    Tag: blume-20000426-fun_suffix
2100    Description:
2101    
2102    Added ".fun" as a recognized file name suffix (for ML code).
2103    
2104    ----------------------------------------------------------------------
2105    Name: Allen Leung
2106    Date: 2000/04/25 17:00:00 EST
2107    Tag: leunga-20000425-alpha-ra
2108    Description:
2109    
2110    1. Alpha
2111    
2112        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
2113    in C--.
2114    
2115    2. RA
2116    
2117       Added a flag "ra-dump-size" to print out the size of the flowgraph
2118       and the interference graph.
2119    
2120    ----------------------------------------------------------------------
2121    Name: Dave MacQueen
2122    Date: 2000/04/25/
2123    Tag: dbm-20000425-mlyacc_doc_examples
2124    Description:
2125      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
2126      Updated all three examples in src/ml-yacc/examples to run
2127      under 110.* using CM.make.
2128    
2129    ----------------------------------------------------------------------
2130    Name: Allen Leung
2131    Date: 2000/04/20 23:04:00 EST
2132    Tag: leunga-20000420-ssa-c---stuff
2133    Description:
2134    
2135      This update synchronizes my repository with Yale's.  Most of these
2136    changes, however, do not affect SML/NJ at all (the RA is an exception).
2137    
2138    1. Register Allocator
2139    
2140       a. An improvement in the interference graph construction:
2141          Given a copy
2142    
2143                s <- t
2144    
2145          no interference edge between s and t is added for this definition of s.
2146    
2147       b. I've added two new spill heuristic modules that Fermin and I developed
2148          (in the new library RA.cm). These are unused in SML/NJ but maybe
2149          useful for others (Moby?)
2150    
2151    2. X86
2152    
2153       a. Various fixes in the backend provided by Fermin [C--] and Lal.
2154    
2155    3. Alpha
2156    
2157       a. Added the BSR instruction and code generation that goes with it [C--]
2158       b. Other fixes too numerous to recount provided by Fermin [C--]
2159    
2160    4. Regmaps
2161    
2162       a. The regmaps are not initialized with the identity physical bindings
2163          at creation time.  This is unneeded.
2164    
2165    5. MLRISC Optimizations
2166    
2167       a. The DJ-Graph module can now compute the iterated dominance frontiers
2168          intersects with liveness incrementally in linear time! Woohoo!
2169          This is now used in my new SSA construction algorithm.
2170    
2171       b. THe branch reorganization module is now smarter about linear chains of
2172          basic blocks.
2173    
2174    
2175  ----------------------------------------------------------------------  ----------------------------------------------------------------------
2176  Name: Matthias Blume  Name: Matthias Blume
2177  Date: 2000/04/12 13:52:00 JST  Date: 2000/04/12 13:52:00 JST
# Line 129  Line 2290 
2290    
2291  3.  Assembly  3.  Assembly
2292    
2293      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
2294      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
2295      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
2296    
# Line 152  Line 2313 
2313    
2314  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
2315  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
2316  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
2317  legal surface-syntax symbol).  legal surface-syntax symbol).
2318    
2319  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 2323 
2323  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
2324    
2325  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2326  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
2327  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
2328  more details.  more details.
2329    
# Line 277  Line 2438 
2438     (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
2439     used).     used).
2440    
2441     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
2442     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
2443     Makefiles or Shell scripts.     Makefiles or Shell scripts.
2444    
2445     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 290  Line 2451 
2451    
2452     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
2453    
2454     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
2455     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
2456    
2457  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 2585 
2585    
2586       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
2587       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
2588       testl is perferable to andl.       testl is preferable to andl.
2589    
2590  7. RA (x86 only)  7. RA (x86 only)
2591    
# Line 606  Line 2767 
2767    
2768  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2769  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2770  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
2771  explanation.  explanation.
2772    
2773  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2774  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2775  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2776  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 2814 
2814  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
2815  <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,
2816  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
2817  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2818  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2819  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2820    
# Line 703  Line 2864 
2864    it from that remote directory.    it from that remote directory.
2865    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2866    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2867    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2868    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2869    
2870    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.845

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