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/NOTES/HISTORY
ViewVC logotype

Diff of /sml/trunk/NOTES/HISTORY

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 577, Fri Mar 10 08:07:18 2000 UTC revision 833, Tue May 22 22:06:12 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: 2001/05/22 18:06:00 EDT
18    Tag: blume-20010522-targets
19    Description:
20    
21    Made install.sh use file config/targets.customized if it exists, falling
22    back to config/targets if it doesn't.  This way one can have a customized
23    version of the targets file without touching the "real thing", thus
24    eliminating the constant fear of accidentally checking something bogus
25    back into the CVS repository...  (File config/targets.customized must
26    not be added to the repository!)
27    
28    ----------------------------------------------------------------------
29    Name: Matthias Blume
30    Date: 2001/05/22 16:30:00 EDT
31    Tag: blume-20010522-minitut
32    Description:
33    
34    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
35       struct returns.
36    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
37       incomplete, preliminary documentation for NLFFI.
38    
39    ----------------------------------------------------------------------
40    Name: Matthias Blume
41    Date: 2001/05/14 11:30:00 EDT
42    Tag: blume-20010514-script
43    Description:
44    
45    Some bugs in install script fixed.
46    
47    In addition to that I also made a slight change to the NLFFI API:
48    Functors generated by ml-nlffigen now take the dynamic library as a
49    straight functor argument, not as a suspended one.  (The original
50    functor code used to force the suspension right away anyway, so there
51    was nothing gained by this complication of the interface.)
52    
53    ----------------------------------------------------------------------
54    Name: Matthias Blume
55    Date: 2001/05/11 14:35:00 EDT
56    Tag: blume-20010511-ml-nlffi
57    Description:
58    
59    I finally took the plunge and added my new FFI code to the main
60    repository.  For x86-linux it is now ready for prime-time.
61    
62    There are two new subdirectories of "src":
63    
64      - ml-nlffi-lib:
65           The utility library for programs using the FFI interface.
66           Here is the implementation of $/c.cm and its associated low-level
67           partners $/c-int.cm and $/memory.cm.
68      - ml-nlffigen:
69           A stand-alone program for generating ML glue code from C source
70           code.
71    
72    Building ml-nlffigen requires $/ckit-lib.cm.
73    
74    The config/install.sh script has been updates to do the Right Thing
75    (hopefully).
76    
77    Notice that the source tree for the C-Kit will not be put under "src"
78    but directly under the installation root directory.  (This is the
79    structure that currently exists on the CVS server when you check out
80    module "sml".)  Fortunately, config/install.sh knows about this oddity.
81    
82    Bugs: No documentation yet.
83    
84    ----------------------------------------------------------------------
85    Name: Matthias Blume
86    Date: 2001/05/09 16:35:00 EDT
87    Tag: blume-20010509-cpscontract
88    Description:
89    
90    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
91    wrapper/unwrapper elimination did not decrement usage counts and some
92    dead variables got overlooked by the dead-up logic.)
93    
94    ----------------------------------------------------------------------
95    Name: Lal George
96    Date: 2001/05/08  17:26:09 EDT
97    Tag: george-20010508-omit-frameptr
98    Description:
99    
100    Changes to implement the omit-frame-pointer optimization to support
101    raw C calls. For now, there is only support on the Intel x86, but
102    other architectures will follow as more experience is gained with this.
103    
104    
105    ----------------------------------------------------------------------
106    Name: Matthias Blume
107    Date: 2001/05/07 14:40:00 EDT
108    Tag: blume-20010507-proxies
109    Description:
110    
111    I made into "proxy libraries" all libraries that qualify for such a
112    change.  (A qualifying library is a library that has another library or
113    groups as its sole member and repeats that member's export list
114    verbatim.  A proxy library avoids this repetition by omitting its export
115    list, effectively inheriting the list that its (only) member exports.
116    See the CM manual for more explanation.)
117    The main effect is that explicit export lists for these libraries
118    do not have to be kepts in sync, making maintenance a bit easier.
119    
120    I also added copyright notices to many .cm-files.
121    
122    Last but not least, I made a new set of bootfiles.
123    
124    ----------------------------------------------------------------------
125    Name: Matthias Blume
126    Date: 2001/05/04 17:00:00 EDT
127    Tag: blume-20010504-cm-lsplit
128    Description:
129    
130    0. John merged pending changes to $/smlnj-lib.cm
131    
132    1. Allen's previous change accidentally backed out of one of Lal's
133       earlier changes.  I undid this mistake (re-introducing Lal's change).
134    
135    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
136       within the compiler where applicable.  There is some code simplification
137       because of that.
138    
139    3. The "split" phase (in FLINT) is now part of the default list of phases.
140       Compiler.Control.LambdaSplitting.* can be used to globally control the
141       lambda-splitting (cross-module-inlining) engine.  In addition to that,
142       it can now also be controlled on a per-source basis: CM has been taught
143       a new tool parameter applicable to ML source files.
144    
145       - To turn lambda-splitting off completely:
146            local open Compiler.Control.LambdaSplitting in
147                val _ = set Off
148            end
149       - To make "no lambda-splitting" the global default (but allow per-source
150         overriding); this is the initial setting:
151            local open Compiler.Control.LambdaSplitting in
152                val _ = set (Default NONE)
153            end
154       - To make "lambda-splitting with aggressiveness a" the global default
155         (and allow per-source overriding):
156            local open Compiler.Control.LambdaSplitting in
157                val _ = set (Default (SOME a))
158            end
159    
160       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
161         write (in the respective .cm-file):
162            a.sml (lambdasplitting:off)
163       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
164            a.sml (lambdasplitting:on)
165       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
166         <a> is a decimal non-negative integer):
167            a.sml (lambdasplitting:<a>)
168       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
169            a.sml (lambdasplitting:infinity)
170       - To use the global default for a.sml:
171            a.sml (lambdasplitting:default)
172         or simply
173            a.sml
174    
175    ----------------------------------------------------------------------
176    Name: Allen Leung
177    Date: 2001/05/04 01:57:00 EDT
178    Tag: leunga-20010504-sync
179    Description:
180    
181      MLRISC features.
182    
183      1. Fix to CMPXCHG instructions.
184      2. Changed RA interface to allow annotations in callbacks.
185      3. Added a new method to the stream interface to allow annotations updates.
186    
187    ----------------------------------------------------------------------
188    Name: Matthias Blume
189    Date: 2001/05/01 11:45:00 EDT
190    Tag: blume-20010501-pcedittmp
191    Description:
192    
193    Changed install.sh to use the current working directory instead of
194    /usr/tmp for a temporary file (pcedittmp).  The previous choice
195    of /usr/tmp caused trouble with MacOS X because of file premission
196    problems.
197    
198    ----------------------------------------------------------------------
199    Name: Matthias Blume
200    Date: 2001/04/20 11:10:00 EDT
201    Tag: blume-20010420-inMLflag
202    Description:
203    
204     - added vp_limitPtrMask to vproc-state.h
205       (for use by the raw-C-calls mechanism to implement proper interrupt
206        handling)
207     - made the ML compiler aware of various data-structure offsets so it
208       can generate code for accessing the vp_inML flag and vp_limitPtrMask
209     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
210       raw C-calls
211    
212    ----------------------------------------------------------------------
213    Name: Lal George
214    Date: 2001/04/20 09:15:28 EDT
215    Tag: george-20010420-macosX
216    Description:
217    
218     - Changes to port to Mac OS X; Darwin.
219    
220     - In the process I found that sqrt was broken on the PPC, because the
221       fsqrt instruction is not implemented.
222    
223    ----------------------------------------------------------------------
224    Name: Matthias Blume
225    Date: 2001/04/18 12:45:00 EDT
226    Tag: blume-20010418-ccalls
227    Description:
228    
229     - fixed two off-by-4 errors in the x86-specific c-calls implementation
230       (this bug prevented structure arguments containing pointers from being
231        passed correctly)
232     - changed the raw-C-call code in mlriscGen.sml in such a way that
233       structure arguments are represented as a pointer to the beginning
234       of the structure (instead of having a series of synthesized arguments,
235       one for each structure member)
236    
237     - made makeml script's verbosity level configurable via environment
238       variable (MAKEML_VERBOSITY)
239    
240     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
241       in rawmem-x86.sml; we are now using the real thing
242    
243    ----------------------------------------------------------------------
244    Name: Matthias Blume
245    Date: 2001/03/22 16:25:00 EST
246    Tag: blume-20010322-bootfiles
247    Description:
248    
249    Created a new set of bootfiles (for your automatic installation convenience).
250    
251    ----------------------------------------------------------------------
252    Name: Matthias Blume
253    Date: 2001/03/22 15:10:00 EST
254    Tag: blume-20010322-rawmem-parcm
255    Description:
256    
257    1. All "raw memory access" primitives for the new FFI are implemented now
258       (at least on the x86).
259    2. Some further cleanup of CM's parallel make mechanism.
260    
261    ----------------------------------------------------------------------
262    Name: Matthias Blume
263    Date: 2001/03/19 17:53:00 EST
264    Tag: blume-20010319-parallel
265    Description:
266    
267    Parallel make (using compile servers) now works again.
268    
269    To this end, CM.stabilize and CMB.make have been modified to work in
270    two passes when compile servers are attached:
271       1. Compile everything, do not perform stabilization; this pass
272          uses compile servers
273       2. Stabilize everything; this pass does not use compile servers
274    If there are no compile servers, the two passes are combined into one
275    (as before).  Splitting the passes increases the inherent parallelism
276    in the dependency graph because the entire graph including all
277    libraries is available at the same time.  This, in turn, improves
278    server utilization.  The downside is that the master process will
279    have to do some extra work after compilation is done (because for
280    technical reasons it must re-read all the binfiles during stabilization).
281    
282    ----------------------------------------------------------------------
283    Name: Matthias Blume
284    Date: 2001/03/16 12:22:00 EST
285    Tag: blume-20010316-bootfiles
286    Description:
287    
288    Created a new set of bootfiles (for your automatic installation convenience).
289    
290    ----------------------------------------------------------------------
291    Name: Matthias Blume
292    Date: 2001/03/16 11:00:00 EST
293    Tag: blume-20010316-MLTREE-fixup
294    Description:
295    
296    This is a minor fixup for an (untagged) earlier commit by Allen.
297    (A file was missing).
298    
299    ----------------------------------------------------------------------
300    Name: Allen Leung
301    Date: Mon Mar  5 18:54:57 EST 2001
302    Tag: leunga-20010305-cut-support
303    
304    1. New support for alternative control-flow in MLTREE.
305       Currently we support
306    
307          FLOW_TO(CALL ...., [k1,...,kn])
308    
309       This is needed for 'cuts to' in C-- and try/handle-like constructs
310       in Moby
311    
312       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
313    
314    2. Register Allocator
315    
316       Changes in interface [from Fermin, John]
317    
318    3. Alpha 8-bit SLL support [Fermin]
319    
320    4. All architectures
321    
322       A new module (ClusterExpandCopies) for expanding parallel copies.
323    
324    ----------------------------------------------------------------------
325    Name: Allen Leung
326    Date: 2001/02/27 23:07:00 EST
327    Tag: leunga-20010227-minor-stuff
328    
329    1. Alpha bug fix for CMOVNE
330    2. Handle mltree COND(..,FCMP ...,...)
331    3. Bug fix in simplifier
332    
333    ----------------------------------------------------------------------
334    Name: Matthias Blume
335    Date: 2001/01/30 17:50:00 EST
336    Tag: blume-20010130-sync
337    Description:
338    
339    This is just a minor update to sync my devel branch with the main brach.
340    The only visible change is the addition of some README files.
341    
342    ----------------------------------------------------------------------
343    Name: Matthias Blume
344    Date: 2001/01/12 23:30:00 JST
345    Tag: blume-20010112-bootfiles
346    Description:
347    
348    Made a new set of bootfiles that goes with the current state of the
349    repository.
350    
351    ----------------------------------------------------------------------
352    Name: Matthias Blume
353    Date: 2001/01/12 21:20:00 JST
354    Tag: blume-20010112-sync
355    Description:
356    
357    I am just flushing out some minor changes that had accumulated in
358    my private branch in order to sync with the main tree.  (This is
359    mainly because I had CVS trouble when trying to merge _into_ my
360    private branch.)
361    
362    Most people should be completely unaffected by this.
363    
364    ----------------------------------------------------------------------
365    Name: Allen Leung
366    Date: Thu Jan 11 21:03:00 EST 2001
367    Tag: leunga-20010111-labexp=mltree
368    Description:
369    
370    1.  Removed the type LabelExp and replace it by MLTree.
371    2.  Rewritten mltree-simplify with the pattern matcher tool.
372    3.  There were some bugs in alpha code generator which would break
373        64-bit code generation.
374    4.  Redo the tools to generate code with the
375    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
376        are now generated by perl scripts.
377    
378    ----------------------------------------------------------------------
379    Name: Matthias Blume
380    Date: 2001/01/10 21:55:00 JST
381    Tag: blume-20010110-rcc
382    Description:
383    
384    The RCC stuff now seems to work (but only on the x86).
385    This required hacking of the c-calls interface (and -implementation) in
386    MLRISC.
387    
388    Normal compiler users should be unaffected.
389    
390    ----------------------------------------------------------------------
391    Name: Matthias Blume
392    Date: 2001/01/09 01:20:00 JST
393    Tag: blume-20010109-rcc
394    Description:
395    
396    This is a fairly big patch, flushing out a large number of pending
397    changes that I made to my development copy over the last couple of days.
398    
399    Of practical relevance at this moment is a workaround for a pickling
400    bug that Allen ran into the other day.  The cause of the bug itself is
401    still unknown and it might be hard to fix it properly, but the
402    workaround has some merits of its own (namely somewhat reducing pickling
403    overhead for certain libraries).  Therefore, I think this solution should
404    be satisfactory at this time.
405    
406    The rest of the changes (i.e., the vast majority) has to do with my
407    ongoing efforts of providing direct support for C function calls from
408    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
409    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
410    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
411    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
412    handle RCC.
413    
414    The last part (the changes to mlriscGen.sml) are still known to be
415    wrong on the x86 and not implemented on all other architectures.  But
416    the infrastructure is in place. I had to change a few functor
417    signatures in the backend to be able to route the CCalls interface
418    from MLRISC there, and I had to specialize the mltree type (on the
419    x86) to include the necessary extensions. (The extensions themselves
420    were already there and redy to go in MLRISC/x86).
421    
422    Everything should be very happy as soon as someone helps me with
423    mlriscGen.sml...
424    
425    In any case, nothing of this should matter to anyone as long as the
426    new primop is not being used (which is going to be the case unless you
427    find it where I hid it :). The rest of the compiler is completely
428    unaffected.
429    
430    ----------------------------------------------------------------------
431    Name: Matthias Blume
432    Date: 2001/01/05 00:30:00 JST
433    Tag: blume-20010105-primops
434    Description:
435    
436    Added some experimental support for work that I am doing right now.
437    These changes mostly concern added primops, but there is also a new
438    experimental C library in the runtime system (but currently not enabled
439    anywhere except on Linux/X86).
440    
441    In the course of adding primops (and playing with them), I discovered that
442    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
443    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
444    To recover, I made type information in INL_PRIM mandatory and changed
445    prim.sml as well as built-in.sml accordingly.  The InLine structure now
446    has complete, correct type information (i.e., no bottom types).
447    
448    Since all these changes mean that we need new binfiles, I also bumped the
449    version number to 110.32.1.
450    
451    ----------------------------------------------------------------------
452    Name: Matthias Blume
453    Date: 2000/12/30 22:10:00 JST
454    Tag: blume-20001230-various
455    Description:
456    
457    Added proxy libraries for MLRISC and let MLRISC libraries refer
458    to each other using path anchors.  (See CM manual for explanation.)
459    
460    Updated CM documentation.
461    
462    Fixed some bugs in CM.
463    
464    Implemented "proxy" libraries (= syntactic sugar for CM).
465    
466    Added "-quiet" option to makeml and changed runtime system accordingly.
467    
468    Added cleanup handler for exportML to reset timers and compiler stats.
469    
470    ----------------------------------------------------------------------
471    Name: Lal George
472    Date: 2000/12/22 22:22:58 EST 2000
473    Tag: Release_110_32
474    Description:
475    
476            Infinite precision used throughout MLRISC.
477            see MLRISC/mltree/machine-int.sig
478    
479    ----------------------------------------------------------------------
480    Name: Matthias Blume
481    Date: 2000/12/22 23:16:00 JST
482    Tag: blume-20001222-warn
483    Description:
484    
485    Corrected wording and formatting of some CM warning message which I
486    broke in my previous patch.
487    
488    ----------------------------------------------------------------------
489    Name: Matthias Blume
490    Date: 2000/12/22 21:20:00 JST
491    Tag: blume-20001222-anchorenv
492    Description:
493    
494    Fixed CM's handling of anchor environments in connection with CMB.make.
495    
496    ----------------------------------------------------------------------
497    Name: Matthias Blume
498    Date: 2000/12/22 13:15:00 JST
499    Tag: blume-20001222-cleanup
500    Description:
501    
502    Removed src/cm/ffi which does not (and did not) belong here.
503    
504    ----------------------------------------------------------------------
505    Name: Matthias Blume
506    Date: 2000/12/21 23:55:00 JST
507    Tag: blume-20001221-exn
508    Description:
509    
510    Probably most important: CM no longer silently swallows all exceptions
511    in the compiler.
512    Plus: some other minor CM changes.  For example, CM now reports some
513    sizes for generated binfiles (code, data, envpickle, lambdapickle).
514    
515    ----------------------------------------------------------------------
516    Name: Matthias Blume
517    Date: 2000/12/15 00:01:05 JST
518    Tag: blume-20001215-dirtool
519    Description:
520    
521    - "dir" tool added.
522    - improvements and cleanup to Tools structure
523    - documentation updates
524    
525    ----------------------------------------------------------------------
526    Name: Allen Leung
527    Date: Thu Dec 14 03:45:24 EST 2000
528    Description:
529    Tag:  leunga-20001214-int-inf
530    Description:
531    
532       In IntInf, added these standard functions, which are missing from our
533    implementation:
534    
535        andb : int * int -> int
536        xorb : int * int -> int
537        orb  : int * int -> int
538        notb : int -> int
539         <<   : int * word -> int
540        ~>>  : int * word -> int
541    
542       Not tested, I hope they are correct.
543    
544    ----------------------------------------------------------------------
545    Name: Allen Leung
546    Date: Fri Dec  8 19:23:26 EST 2000
547    Description:
548    Tag:  leunga-20001208-nowhere
549    Description:
550    
551      Slight improvements to the 'nowhere' tool to handle OR-patterns,
552    to generate better error messages etc.  Plus a brief manual.
553    
554    ----------------------------------------------------------------------
555    Name: Lal George
556    Date: 2000/12/08 09:54:02 EST 2000
557    Tag: Release_110_31
558    Description:
559    
560    - Version 110.31
561    ----------------------------------------------------------------------
562    Name: Allen Leung
563    Date: Thu Dec  7 22:01:04 EST 2000
564    Tag:  leunga-20001207-cell-monster-hack
565    Description:
566    
567    Major MLRISC internal changes.  Affect all clients.
568    Summary:
569    
570    1.  Type CELLS.cell = int is now replaced by a datatype.
571        As a result, the old regmap is now gone.  Almost all interfaces
572        in MLRISC change as a consequence.
573    
574    2.  A new brand version of machine description tool (v3.0) that generates
575        modules expecting the new interface.  The old version is removed.
576    
577    3.  The RA interface has been further abstracted into two new functors.
578        RISC_RA and X86RA.  These functors have much simpler interfaces.
579        [See also directory MLRISC/demo.]
580    
581    4.  Some other new source->source code generation tools are available:
582    
583        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
584        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
585           I use this tool to generate the peephole optimizers---with the new
586           cell type changes, peephole rules are becoming difficult to write
587           without conditional pattern matching.
588    
589    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
590        cover the entire MLRISC source tree so many things broke.
591    
592    6.  CM files have been moved to the subdirectory MLRISC/cm.
593        They are moved because there are a lot of them and they clutter up the
594        root dir.
595    
596    7.  More detailed documentation to come...
597    
598        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
599        a makeml -rebuild first.  This is because of other other
600        changes that Matthias has made (see below).
601    
602    
603    ----------------------------------------------------------------------
604    Name: Matthias Blume
605    Date: 2000/11/30 23:12:00 JST
606    Tag: blume-20001130-filereorg
607    Description:
608    
609    Some manual updates and some file reorganizations in CM.
610    
611    ----------------------------------------------------------------------
612    Name: Matthias Blume
613    Date: 2000/11/24 17:45:00 JST
614    Tag: blume-20001124-link
615    Description:
616    
617    Drastically improved link traversal code for the case that the dynamic
618    value was already loaded at bootstrap time.  As a result, CM and CMB
619    now both load blazingly fast -- even on a very slow machine.  Also,
620    memory consumption has been further reduced by this.
621    
622    Warning: The format of the PIDMAP file has changed.  THerefore, to
623    bootstrap you have to do this:
624    
625    1. Run CMB.make
626    2. Make a symbolic link for the boot directory:
627         ln -s sml.boot.ARCH-OS xxx
628    3. "Rebuild" the boot directory:
629         ./makeml -boot xxx -rebuild sml ; rm xxx
630    4. Boot normally:
631          ./makeml
632    
633    ----------------------------------------------------------------------
634    Name: Matthias Blume
635    Date: 2000/11/21 21:20:00 JST
636    Tag: blume-20001121-tools
637    Description:
638    
639    Continued hacking on autoloading problem -- with success this time.
640    Also changed tool-plugin mechanism.  See new CM manual.
641    
642    ----------------------------------------------------------------------
643    Name: Matthias Blume
644    Date: 2000/11/19 14:30:00 JST
645    Tag:  blume-20001119-autoload
646    Description:
647    
648    Some hacking to make autoloading faster.  Success for CMB, no success
649    so far for CM.  There is a reduced structure CM' that autoloads faster.
650    (This is a temporary, non-documented hack to be eliminated again when
651    the general problem is solved.)
652    
653    ----------------------------------------------------------------------
654    Name: Matthias Blume
655    Date: 2000/11/17 14:10:00 JST
656    Tag: blume-20001117-pickle-lib
657    Description:
658    
659    1. Eliminated comp-lib.cm
660    2. Made pickle-lib.cm
661    3. Eliminated all uses of intset.sml (from comp-lib.cm)
662    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
663       equivalent constructs from smlnj-lib.cm (INtHashTable).
664    5. Point 4. also goes for those uses of intmap.* in MLRISC.
665       Duplicated intmap modules thrown out.
666    6. Hunted down all duplicated SCC code and replaced it with
667       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
668    7. Rewrote Feedback module.
669    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
670       should be thrown out and equivalent modules from smlnj-lib.cm
671       should be used (IntRedBlackSet, IntListSet, ...).
672    
673    Confirmed that compiler compiles to fixpoint.
674    
675    ----------------------------------------------------------------------
676    Name: Allen Leung
677    Date: 2000/11/10 18:00:00
678    Tag: leunga-20001110-new-x86-fp
679    
680    A new x86 floating point code generator has been added.
681    By default this is turned off.  To turn this on, do:
682    
683        CM.autoload "$smlnj/compiler.cm";
684        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
685    
686    Changes:
687    
688    1.  Changed FTAN to FPTAN so that the assembly output is correct.
689    2.  Changed the extension callback for FTANGENT to generate:
690    
691              fptan
692              fstp  %st(0)
693        instead of
694              fptan
695              fstpl ftempmem
696    
697    3.  Numerous assembly fixes for x86.
698    
699    5.  Cleaned up the machine code output module x86/x86MC.sml and added
700        support for a whole bunch of instructions and addressing modes:
701    
702          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
703          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
704          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
705          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
706          fxch %st(n)
707          fld %st(n)
708          fst %st(n)
709          fst mem
710          fstp %st(n)
711          fucom %st(n)
712          fucomp %st(n)
713    
714        All these are now generated when the fast fp mode is turned on.
715    
716    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
717    
718    ----------------------------------------------------------------------
719    Name: Matthias Blume
720    Date: 2000/11/09 11:20:00 JST
721    Tag: blume-20001109-scc
722    Description:
723    
724    Eliminated some code duplication:
725    
726    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
727       (Otherwise the functor is useless.)
728    2. Used GraphSCCFn where SCCUtilFun was used previously.
729    3. Got rid of SCCUtilFun (in comp-lib.cm).
730    
731    ----------------------------------------------------------------------
732    Name: Lal George
733    Date: 2000/11/06 09:02:21 EST 2000
734    Tag: Release_110_30
735    Description:
736    
737    - Version 110.30
738    ----------------------------------------------------------------------
739    Name: Matthias Blume
740    Date: 2000/11/04 14:45:00
741    Tag: blume-20001104-mlbuild
742    Description:
743    
744    - Made ml-build faster on startup.
745    - Documentation fixes.
746    
747    ----------------------------------------------------------------------
748    Name: Matthias Blume
749    Date: 2000/11/02 17:00:00 JST
750    Tag: blume-20001102-condcomp
751    Description:
752    
753    - Small tweaks to pickler -- new BOOTFILES!
754    - Version bumped to 110.29.2.
755    - Added conditional compilation facility to init.cmi (see comment there).
756    ----------------------------------------------------------------------
757    Name: Allen Leung
758    Date: 2000/10/23 19:31:00
759    Tag: leunga-20001023-demo-ra
760    
761    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
762    2. Test programs for the graph library updated
763    3. Some new MLRISC demo programs added
764    
765    ----------------------------------------------------------------------
766    Name: Matthias Blume
767    Date: 2000/08/31 22:15:00 JST
768    Tag: blume-20001017-errmsg
769    Description:
770    
771    More error message grief: Where there used to be no messages, there
772    now were some that had bogus error regions.  Fixed.
773    
774    ----------------------------------------------------------------------
775    Name: Matthias Blume
776    Date: 2000/08/31 17:30:00 JST
777    Tag: blume-20001017-v110p29p1
778    Description:
779    
780    I made a version 110.29.1 with new bootfiles.
781    
782    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
783              CM documentation changes and a small bugfix in CM's error reporting.
784    
785    ----------------------------------------------------------------------
786    Name: Lal George
787    Date: 2000/09/27 14:42:35 EDT
788    Tag: george-20000927-nodestatus
789    Description:
790    
791    Changed the type of the nodestatus, so that:
792    
793            SPILLED(~1)             is now SPILLED
794            SPILLED(m) where m>=0   is now MEMREG(m)
795            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
796    
797    ----------------------------------------------------------------------
798    Name: Matthias Blume
799    Date: 2000/09/07 14:45:00 JST
800    Tag: blume-20000907-cmerrmsg
801    Description:
802    
803    Small tweak to CM to avoid getting ML syntax error messages twice.
804    
805    ----------------------------------------------------------------------
806    Name: Matthias Blume
807    Date: 2000/08/31 18:00:00 JST
808    Tag: blume-20000831-cvsbootfiles
809    Description:
810    
811    New URL for boot files (because the 110.29 files on the BL server do
812    now work correctly with my updated install scripts for yacc and lex).
813    
814    ----------------------------------------------------------------------
815    Name: Matthias Blume
816    Date: 2000/08/08 12:33:00 JST
817    Tag: blume-20000808-manual
818    Description:
819    
820    Tiny update to CM manual.
821    
822    ----------------------------------------------------------------------
823    Name: Allen Leung
824    Date: 2000/08/7 19:31:00
825    Tag: leunga-20000807-a-whole-bunch-of-stuff
826    
827      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
828    and Mobdy.
829    
830    1.  x86
831    
832       a.  Fixes to peephole module by John and Dan.
833       b.  Assembly fix to SETcc by Allen.
834       c.  Fix to c-call by John.
835       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
836       e.  Instruction selection optimization to SETcc as suggested by John.
837    
838           For example,
839    
840            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
841    
842           should generate:
843    
844            MOVL a, x
845            SUBL b, x
846            SHRL 31, x
847    
848    2.  IR stuff
849    
850         A bunch of new DJ-graph related algorithms added.  These
851         speed up SSA construction.
852    
853    3.  SSA + Scheduling
854    
855         Added code for SSA and scheduling to the repository
856    
857    ----------------------------------------------------------------------
858    Name: Lal George
859    Date: 2000/07/27 11:53:14 EDT
860    
861    Tag: lal-20000727-linux-ppc
862    Description:
863    
864     Made changes to support Linux PPC.
865     p.s. I have confirmation that the 110.29 boot files work fine.
866    
867    ----------------------------------------------------------------------
868    Name: Matthias Blume
869    Date: 2000/07/27 17:40:00 JST
870    Tag: blume-20000727-scripts
871    Description:
872    
873    !!!! WARNING !!!!
874    You must recompile the runtime system!
875    !!!! WARNING !!!!
876    
877    This is basically another round of script-enhancements:
878    
879    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
880       and undefine CM preprocessor symbols.
881    
882    2. ml-build avoids generating a new heap image if it finds that the
883       existing one is still ok.  (The condition is that no ML file had to
884       be recompiled and all ML files are found to be older that the heap
885       file.)
886    
887       To make this work smoothly, I also hacked the runtime system as
888       well as SMLofNJ.SysInfo to get access to the heap image suffix
889       (.sparc-solaris, ...) that is currently being used.
890    
891       Moreover, the signature of CM.mk_standalone has changed.  See the
892       CM manual.
893    
894    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
895       CM manual for details.)
896    
897    4. More CM manual updates:
898        - all of the above has been documented.
899        - there is now a section describing the (CM-related) command line
900          arguments that are accepted by the "sml" command
901    
902    ----------------------------------------------------------------------
903    Name: Matthias Blume
904    Date: 2000/07/25 16:20:00 JST
905    Tag: blume-20000725-makedepend
906    Description:
907    
908    Added a script called ml-makedepend.  This can be used in makefiles
909    for Unix' make in a way very similar to the "makedepend" command for
910    C.
911    
912    The script internally uses function CM.sources.
913    
914    Synopsis:
915    
916        ml-makedepend [-f makefile] cmfile targetname
917    
918    The default for the makefile is "makefile" (or "Makefile" should
919    "makefile" not exist).
920    
921    ml-makedepend adds a cmfile/targetname-specific section to this
922    makefile (after removing the previous version of this section).  The
923    section contains a single dependency specification with targetname on
924    the LHS (targetname is an arbitrary name), and a list of files derived
925    from the cmfile on the RHS.  Some of the files on the RHS are
926    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
927    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
928    path names.  The makefile writer is responsible for making sure that
929    these variables have correct at the time "make" is invoked.
930    
931    ----------------------------------------------------------------------
932    Name: Matthias Blume
933    Date: 2000/07/22 23:30:00 JST
934    Tag: blume-20000722-urlupdate
935    Description:
936    
937    Changed BOOT and config/srcarchiveurl to point to BL server:
938    
939        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
940    
941    ----------------------------------------------------------------------
942    Name: Matthias Blume
943    Date: 2000/07/18 18:00:00 JST
944    Tag: blume-20000718-Version_110_29
945    Description:
946    
947    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
948    
949    2. Updated config/version to 110.29
950    
951    3. Updated config/srcarchiveurl
952    
953    3. New boot files!
954       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
955    
956    ----------------------------------------------------------------------
957    Name: Matthias Blume
958    Date: 2000/07/11 13:58:00 JST
959    Tag: blume-20000711-doctypo
960    Description:
961    
962    Fixed a few typos in CM manual.
963    
964    ----------------------------------------------------------------------
965    Name: Allen Leung
966    Date: 2000/06/15 00:38:00
967    Tag: leunga-20000704-sparc-x86
968    
969    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
970    2. fix to x86 RET bug [found by Dan Grossman]
971    3. sparc assembly bug fix for ticc instructions [found by Fermin]
972    
973       Affects c-- and moby only
974    
975    ----------------------------------------------------------------------
976    Name: Matthias Blume
977    Date: 2000/07/04 15:26:00
978    Tag: blume-20000704-trigger
979    Description:
980    
981    1. Improvements to CM manual.
982    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
983       of getting a back-trace.  The function, when called, raises an
984       internal exception which explicitly carries the full back-trace history,
985       so it is unaffected by any intervening handle-raise pairs ("trivial"
986       or not).  The interactive loop will print that history once it arrives
987       at top level.
988       Short of having all exceptions implicitly carry the full history, the
989       recommended way of using this facility is:
990         - compile your program with instrumentation "on"
991         - run it, when it raises an exception, look at the history
992         - if the history is "cut off" because of some handler, go and modify
993           your program so that it explicitly calls BTrace.trigger
994         - recompile (still instrumented), and rerun; look at the full history
995    
996    ----------------------------------------------------------------------
997    Name: Matthias Blume
998    Date: 2000/07/03 15:36:00 JST
999    Tag: blume-20000702-manual
1000    Description:
1001    
1002    Small corrections and updates to CM manual.
1003    
1004    ----------------------------------------------------------------------
1005    Name: Matthias Blume
1006    Date: 2000/06/29 16:04:00 JST
1007    Tag: blume-20000629-yacctool
1008    Description:
1009    
1010    Changes:
1011    
1012    1. Class "mlyacc" now takes separate arguments to pass options to
1013       generated .sml- and .sig-files independently.
1014    2. Corresponding CM manual updates.
1015    3. BTrace module now also reports call sites.  (However, for loop clusters
1016       it only shows from where the cluster was entered.)  There are associated
1017       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1018    
1019    ----------------------------------------------------------------------
1020    Name: Matthias Blume
1021    Date: 2000/06/27 16:51:00 JST
1022    Tag: blume-20000627-noweb
1023    Description:
1024    
1025    Changes:
1026    
1027     1. Implemented "subdir" and "witness" options for noweb tool.
1028        This caused some slight internal changes in CM's tool implementation.
1029     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
1030        some remaining issues from earlier path anchor changes.
1031     3. Updated CM manual accordingly.
1032    
1033     4. Changed implementation of back-tracing so that I now consider it
1034        ready for prime-time.
1035    
1036        In particular, you don't have to explicitly trigger the back-trace
1037        anymore.  Instead, if you are running BTrace-instrumented code and
1038        there is an uncaught exception (regardless of whether or not it was
1039        raised in instrumented code), the top-level evalloop will print
1040        the back-trace.
1041    
1042        Features:
1043    
1044          - Instrumented and uninstrumented code work together seemlessly.
1045            (Of course, uninstrumented code is never mentioned in actual
1046             back-traces.)
1047    
1048          - Asymptotic time- and space-complexity of instrumented code is
1049            equal to that of uninstrumented code.  (This means that
1050            tail-recursion is preserved by the instrumentation phase.)
1051    
1052          - Modules whose code has been instrumented in different sessions
1053            work together without problem.
1054    
1055          - There is no penalty whatsoever on uninstrumented code.
1056    
1057          - There is no penalty on "raise" expressions, even in
1058            instrumented code.
1059    
1060        A potential bug (or perhaps it is a feature, too):
1061    
1062          A back-trace reaches no further than the outermost instrumented
1063          non-trivial "raise".  Here, a "trivial" raise is one that is the
1064          sole RHS of a "handle" rule.  Thus, back-traces reach trough
1065    
1066               <exp> handle e => raise e
1067    
1068          and even
1069    
1070               <exp> handle Foo => raise Bar
1071    
1072          and, of course, through
1073    
1074               <exp> handle Foo => ...
1075    
1076         if the exception was not Foo.
1077    
1078         Back-traces always reach right through any un-instrumented code
1079         including any of its "handle" expressions, trivial or not.
1080    
1081       To try this out, do the following:
1082    
1083         - Erase all existing binfiles for your program.
1084           (You may keep binfiles for those modules where you think you
1085            definitely don't need back-tracing.)
1086         - Turn on back-trace instrumentation:
1087              SMLofNJ.Internals.BTrace.mode (SOME true);
1088         - Recompile your program.  (I.e., run "CM.make" or "use".)
1089         - You may now turn instrumentation off again (if you want):
1090              SMLofNJ.Internals.BTrace.mode (SOME false);
1091         - Run your program as usual.  If it raises an exception that
1092           reaches the interactive toplevel, then a back-trace will
1093           automatically be printed.  After that, the toplevel loop
1094           will print the exception history as usual.
1095    
1096    ----------------------------------------------------------------------
1097    Name: Matthias Blume
1098    Date: 2000/06/26 09:56:46 JST
1099    Tag: blume-20000626-setup
1100    Description:
1101    
1102    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1103          ML code before and after compiling a file (e.g., to set compiler
1104          flags)
1105    
1106    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1107              - associated changes to btrace.sml (BTrace instrumentation pass)
1108              - cleaner implementation of btimp.sml (BTrace tracing and report
1109                module)
1110    
1111    CM manual: * new path encoding documented
1112               * description of setup-parameter to "sml" added
1113    
1114    The biggest user-visible change to back-tracing is that it is no
1115    longer necessary to compile all traced modules within the same
1116    session.  (This was a real limitation.)
1117    
1118    ----------------------------------------------------------------------
1119    Name: Matthias Blume
1120    Date: 2000/06/24 12:40:00 JST
1121    Tag: blume-20000624-startup
1122    Description:
1123    
1124    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1125    bit too often -- to put it mildly. :)
1126    
1127    ----------------------------------------------------------------------
1128    Name: Matthias Blume
1129    Date: 2000/06/23 18:20:00 JST
1130    Tag: blume-20000623-btrace
1131    Description:
1132    
1133    This updates adds a backtrace facility to aid programmers in debugging
1134    their programs.  This involves the following changes:
1135    
1136    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1137       keeping track of the current call stack.  When programs are compiled
1138       in a special mode, the compiler will insert calls to these hooks
1139       into the user program.
1140       "Hook" means that it is possible for different implementations of
1141       back-tracing to register themselves (at different times).
1142    
1143    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1144       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1145       It can be turned on using this call:
1146         SMLofNJ.Internals.BTrace.mode (SOME true);
1147       Turning it off again:
1148         SMLofNJ.Internals.BTrace.mode (SOME false);
1149       Querying the current status:
1150         SMLofNJ.Internals.BTrace.mode NONE;
1151       Annotated programs are about twice as big as normal ones, and they
1152       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1153       where all hooks do nothing).  The slowdown with a plugin that is
1154       actually useful (such as the one supplied by default) is even greater,
1155       but in the case of the default plugin it is still only an constant
1156       factor (amortized).
1157    
1158    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1159       with a sub-structure BTrace for controlling back-tracing.  In particular,
1160       the above-mentioned function "mode" controls whether the annotation
1161       phase is invoked by the compiler.  Another important function is
1162       "trigger": when called it aborts the current execution and causes
1163       the top-level loop to print a full back-trace.
1164    
1165    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1166       for back-tracing.  It keeps track of the dynamic call stack and in
1167       addition to that it keeps a partial history at each "level" of that
1168       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1169       d and b (at separate times, dynamically), then the report will show:
1170    
1171       GOTO   d
1172             /c
1173       GOTO  \b
1174       CALL   a
1175    
1176       This shows that there was an initial non-tail call of a, then a
1177       tail-call to b or c, looping behavior in a cluster of functions that
1178       consist of b and c, and then a goto from that cluster (i.e., either from
1179       b or from c) to d.
1180    
1181       Note that (depending on the user program) the amount of information
1182       that the back-trace module has to keep track of at each level is bounded
1183       by a constant.  Thus, the whole implementation has the same asymptotical
1184       complexity as the original program (both in space and in time).
1185    
1186    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1187       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1188       which is raised by the "trigger" function mentioned above.
1189    
1190    Notes on usage:
1191    
1192    - Annotated code works well together with unannotated code:
1193    Unannotated calls simply do not show up at all in the backtrace.
1194    
1195    - It is not a good idea to let modules that were annotated during
1196    different sessions run at the same time.  This is because the compiler
1197    chooses small integers to identify individual functions, and there
1198    will be clashes if different modules were compiled in separate sessions.
1199    (Nothing will crash, and you will even be told about the clashes, but
1200    back-trace information will in general not be useful.)
1201    
1202    - Back-tracing can be confused by callcc and capture.
1203    
1204    - The only way of getting a back-trace right now is to explicitly
1205    invoke the "trigger" function from your user program.  Eventually, we
1206    should make every exception carry back-trace information (if
1207    available).  But since this creates more overhead at "raise"-time
1208    (similar to the current exnHistory overhead), I have not yet
1209    implemented this.  (The implementation will be rather easy.)  With
1210    exceptions carrying back-trace information, this facility will be even
1211    more useful because users don't need to modify their programs...
1212    
1213    - While it is possible to compile the compiler with back-trace
1214    annotations turned on (I did it to get some confidence in
1215    correctness), you must make absolutely sure that core.sml and
1216    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1217    be compiled with annotation because there is no core access yet, but
1218    if you compile btimp.sml with annotation, then the system will go into
1219    an infinite recursion and crash.)
1220    Since CM currently does not know about BTrace, the only way to turn
1221    annotations on and off for different modules of the compiler is to
1222    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1223    this is awkward and clumsy.
1224    
1225    Sample sessions:
1226    
1227    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1228    - SMLofNJ.Internals.BTrace.mode (SOME true);
1229    [autoloading]
1230    [autoloading done]
1231    val it = false : bool
1232    - structure X = struct
1233    -     fun main n = let
1234    -         fun a (x, 0) = d x
1235    -           | a (x, n) = b (x, n - 1)
1236    -         and b (x, n) = c (x, n)
1237    -         and c (x, n) = a (x, n)
1238    -         and d x = e (x, 3)
1239    -         and e (x, 0) = f x
1240    -           | e (x, n) = e (x, n - 1)
1241    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1242    -           | f n = n * g (n - 1)
1243    -         and g n = a (n, 3)
1244    -     in
1245    -         f n
1246    -     end
1247    - end;
1248    structure X : sig val main : int -> int end
1249    - X.main 3;
1250    *** BACK-TRACE ***
1251    GOTO   stdIn:4.2-13.20: X.main[2].f
1252    GOTO-( stdIn:4.2-13.20: X.main[2].e
1253    GOTO   stdIn:4.2-13.20: X.main[2].d
1254         / stdIn:4.2-13.20: X.main[2].a
1255         | stdIn:4.2-13.20: X.main[2].b
1256    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1257    CALL   stdIn:4.2-13.20: X.main[2].g
1258    GOTO   stdIn:4.2-13.20: X.main[2].f
1259    GOTO-( stdIn:4.2-13.20: X.main[2].e
1260    GOTO   stdIn:4.2-13.20: X.main[2].d
1261         / stdIn:4.2-13.20: X.main[2].a
1262         | stdIn:4.2-13.20: X.main[2].b
1263    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1264    CALL   stdIn:4.2-13.20: X.main[2].g
1265    GOTO   stdIn:4.2-13.20: X.main[2].f
1266    GOTO-( stdIn:4.2-13.20: X.main[2].e
1267    GOTO   stdIn:4.2-13.20: X.main[2].d
1268         / stdIn:4.2-13.20: X.main[2].a
1269         | stdIn:4.2-13.20: X.main[2].b
1270    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1271    CALL   stdIn:4.2-13.20: X.main[2].g
1272    GOTO   stdIn:4.2-13.20: X.main[2].f
1273    CALL   stdIn:2.15-17.4: X.main[2]
1274    -
1275    
1276    (Note that because of a FLINt bug the above code currently does not
1277    compile without BTrace turned on.)
1278    
1279    Here is another example, using my modified Tiger compiler:
1280    
1281    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1282    - SMLofNJ.Internals.BTrace.mode (SOME true);
1283    [autoloading]
1284    [autoloading done]
1285    val it = false : bool
1286    - CM.make "sources.cm";
1287    [autoloading]
1288    ...
1289    [autoloading done]
1290    [scanning sources.cm]
1291    [parsing (sources.cm):parse.sml]
1292    [creating directory CM/SKEL ...]
1293    [parsing (sources.cm):tiger.lex.sml]
1294    ...
1295    [wrote CM/sparc-unix/semant.sml]
1296    [compiling (sources.cm):main.sml]
1297    [wrote CM/sparc-unix/main.sml]
1298    [New bindings added.]
1299    val it = true : bool
1300    - Main.compile ("../testcases/merge.tig", "foo.out");
1301    *** BACK-TRACE ***
1302    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
1303    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1304    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1305    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1306    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1307    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1308    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
1309         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
1310    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
1311    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1312    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
1313    CALL   main.sml:1.18-118.4: Main.compile[2]
1314    -
1315    
1316    ----------------------------------------------------------------------
1317    Name: Matthias Blumen
1318    Date: 2000/06/21 18:00:00 JST
1319    Tag: blume-20000621-manual
1320    Description:
1321    
1322    CM manual update: Path environments documented.
1323    
1324    ----------------------------------------------------------------------
1325    Name: Matthias Blume
1326    Date: 2000/06/19 13:40:00
1327    Tag: blume-20000619-manual
1328    Description:
1329    
1330    CM manual and system/README update.  This only covers the fact that
1331    there are no more implicit anchors.  (Path environments and the "bind"
1332    option to "cm" have yet to be documented.)
1333    
1334    ----------------------------------------------------------------------
1335    Name: Matthias Blume
1336    Date: 2000/06/19 11:05:00 JST
1337    Tag: blume-20000619-chdir-bugfix
1338    Description:
1339    
1340    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1341    
1342    ----------------------------------------------------------------------
1343    Name: Matthias Blume
1344    Date: 2000/06/18 22:00:10 JST
1345    Tag: blume-20000618-implicit-anchors-really-gone
1346    Description:
1347    
1348    I updates the previous HISTORY entry where I forgot to mention that
1349    implicit anchors are no longer with us.
1350    
1351    The current update also gets rid of the (now useless) controller
1352    CM.Control.implicit_anchors.
1353    
1354    ----------------------------------------------------------------------
1355    Name: Matthias Blume
1356    Date: 2000/06/16 17:30:00 JST
1357    Tag: blume-20000616-anchorenv
1358    Description:
1359    
1360    This patch implements the long anticipated (just kidding :) "anchor
1361    environment" mechanism.  In the course of doing this, I also
1362    re-implemented CM's internal "SrcPath" module from scratch.  The new
1363    one should be more robust in certain boundary cases.  In any case, it
1364    is a lot cleaner than its predecessor (IMHO).
1365    
1366    This time, although there is yet another boot file format change, I
1367    kept the unpickler backward-compatible.  As a result, no new bootfiles
1368    are necessary and bootstrapping is straightforward.  (You cannot read
1369    new bootfiles into an old system, but the other way around is no
1370    problem.)
1371    
1372    Visible changes:
1373    
1374    ** 0. Implicit path anchors (without the leading $-symbol) are no
1375    longer recognized at all. This means that such path names are not
1376    illegal either.  For example, the name basis.cm simply refers to a
1377    local file called "basis.cm" (i.e, the name is an ordinary path
1378    relative to .cm-files directory).  Or, to put it differently, only
1379    names that start with $ are anchored paths.
1380    
1381    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1382    vanished.
1383    
1384    John (Reppy) had critizised this as soon as I originally proposed and
1385    implemented it, but at that time I did not really deeply believe
1386    him. :) Now I came full-circle because I need the $<singlearc> syntax
1387    in another place where it cannot be seen as an abbreviation for
1388    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1389    seems to mean (i.e., it "expands" into the corresponding anchor
1390    value).
1391    
1392    However, when paths are used as members in CM description files, it
1393    continues to be true that there must be at least another arc after the
1394    anchor.  This is now enforced separately during semantic analysis
1395    (i.e., from a lexical/syntactical point of view, the notation is ok.)
1396    
1397    ** 2. The "cm" class now accepts an option "bind".  The option's value
1398    is a sub-option list of precisely two items -- one labeled "anchor"
1399    and the other one labeled "value".  As you might expect, "anchor" is
1400    used to specify an anchor name to be bound, and "value" specifies what
1401    the anchor is being bound to.
1402    
1403    The value must be a directory name and can be given in either standard
1404    syntax (including the possibility that it is itself an anchored path)
1405    or native syntax.
1406    
1407    Examples:
1408    
1409       foo.cm (bind:(anchor:bar value:$mystuff/bar))
1410       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1411    
1412    and so on.
1413    
1414    The meaning of this is that the .cm-file will be processed with an
1415    augmented anchor environment where the given anchor(s) is/are bound to
1416    the given values(s).
1417    
1418    The rationale for having this feature is this: Suppose you are trying
1419    to use two different (already stable) libraries a.cm and b.cm (that
1420    you perhaps didn't write yourself).  Further, suppose each of these
1421    two libraries internally uses its own auxiliary library $aux/lib.cm.
1422    Normally you would now have a problem because the anchor "lib" can not
1423    be bound to more than one value globally.  Therefore, the project that
1424    uses both a.cm and b.cm must locally redirect the anchor to some other
1425    place:
1426    
1427       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1428       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1429    
1430    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1431    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
1432    
1433    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1434    you when you do so at the time of CM.stabilize).  Therefore, you can
1435    also use an anchored path as the value:
1436    
1437      a.cm (bind:(anchor:lib value:$a-lib))
1438      b.cm (bind:(anchor:lib value:$b-lib))
1439    
1440    Now you can globally configure (using the usual CM.Anchor.anchor or
1441    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1442    itself is always locally bound, setting it globally is no longer
1443    meaningful or necessary (but it does not hurt either).  In fact, "lib"
1444    can still be used as a global anchor for separate purposes.  As a
1445    matter of fact, one can locally define "lib" in terms of a global
1446    "lib":
1447    
1448      a.cm (bind:(anchor:lib value:$lib/a))
1449      b.cm (bind:(anchor:lib value:$lib/b))
1450    
1451    ** 3: The encoding of path names has changed.  This affects the way
1452    path names are shown in CM's progress report and also the internal
1453    protocol encoding used for parallel make.
1454    
1455    The encoding now uses one or more ':'-separated segments.  Each
1456    segments corresponds to a file that has been specified relative to the
1457    file given by its preceding segment.  The first segment is either
1458    relative to the CWD, absolute, or anchored.  Each segment itself is
1459    basically a Unix pathname; all segments but the first are relative.
1460    
1461    Example:
1462    
1463       $foo/bar/baz.cm:a/b/c.sml
1464    
1465    This path denotes the file bar/a/b/c.sml relative to the directory
1466    denoted by anchor "foo".  Notice that the encoding also includes
1467    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1468    paths are resolved relative to the .cm-files directory, so baz.cm must
1469    be ignored to get the "real" pathname.
1470    
1471    To make this fact more obvious, CM puts the names of such "virtual
1472    arcs" into parentheses when they appear in progress reports. (No
1473    parentheses will appear in the internal protocol encoding.)  Thus,
1474    what you really see is:
1475    
1476      $foo/bar/(baz.cm):a/b/c.sml
1477    
1478    I find this notation to be much more informative than before.
1479    
1480    Another new feature of the encoding is that special characters
1481    including parentheses, colons, (back)slashes, and white space are
1482    written as \ddd (where ddd is the decimal encoding of the character).
1483    
1484    *** The CM manual still needs to be updated.
1485    
1486    ----------------------------------------------------------------------
1487    Name: Allen Leung
1488    Date: 2000/06/15 00:38:00
1489    Tag: leunga-20000615-x86-peephole
1490    
1491    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1492    
1493    ----------------------------------------------------------------------
1494    Name: Matthias Blume
1495    Date: 2000/06/12 11:40:00
1496    Tag: blume-20000612-parmakefix
1497    Description:
1498    
1499    More cleanup after changing the file naming scheme: This time I
1500    repaired the parallel make mechanism for CMB.make which I broke earlier.
1501    
1502    ----------------------------------------------------------------------
1503    Name: Allen Leung
1504    Date: 2000/06/09 01:25:00
1505    Tag: leunga-20000609-various
1506    
1507    None of these things should affect normal SML/NJ operations
1508    
1509    1. Peephole improvements provided by Fermin (c--)
1510    2. New annotation DEFUSE for adding extra dependence (moby)
1511    3. New X86 LOCK instructions (moby)
1512    4. New machine description language for reservation tables (scheduling)
1513    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1514       trees etc.)
1515    6. I've changed the CM files so that they can work with versions
1516       110.0.6, 110.25 and 110.28
1517    
1518    ----------------------------------------------------------------------
1519    Name: Matthias Blume
1520    Date: 2000/06/09 12:40:00
1521    Tag: blume-20000609-log
1522    Description:
1523    
1524    - Removed all(?) remaining RCS Log entries from sources.
1525    
1526    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1527      anchored paths).
1528    
1529    ----------------------------------------------------------------------
1530    Name: Matthias Blume
1531    Date: 2000/06/07 17:00:00 JST
1532    Tag: blume-20000607-no-implicit-anchors
1533    Description:
1534    
1535    1. This update changes the default setting for
1536    CM.Control.implicit_anchors from true to false.  This means that
1537    implicit anchors are no longer permitted by default.  I also tried to
1538    make sure that nothing else still relies on implicit anchors.
1539    (This is the next step on the schedule towards a CM that does not even
1540    have the notion of implicit anchors anymore.)
1541    
1542    2. More CM manual updates.
1543    
1544    3. I managed to track down and fix the pickling bug I mentioned last
1545    time.  Because of the previously existing workaround, this entails no
1546    immediate practical changes.
1547    
1548    ----------------------------------------------------------------------
1549    Name: Matthias Blume
1550    Date: 2000/06/06 11:15:00 JST
1551    Tag: blume-20000606-lazierpickle
1552    Description:
1553    
1554    !!!! NEW BOOT FILES !!!!
1555    
1556    * The main purpose of this update is to make library pickles lazier in
1557    order to reduce the initial space penalty for autoloading a library.
1558    As a result, it is now possible to have $smlnj/compiler.cm
1559    pre-registered.  This should take care of the many complaints or
1560    inquiries about missing structure Compiler.  This required changes to
1561    CM's internal data structures and small tweaks to some algorithms.
1562    
1563    As a neat additional effect, it is no longer necessary (for the sake
1564    of lean heap image files) to distinguish between a "minimal" CM and a
1565    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1566    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1567    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1568    
1569    To make the life of compiler-hackers easier, "makeml" now also
1570    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1571    words, after you bootstrap a new sml for the first time, you will not
1572    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1573    around you will still have to do it, though.)
1574    
1575    * A second change consists of major updates to the CM manual.  There
1576    are now several appendices with summary information and also a full
1577    specification of the CM description file syntax.
1578    
1579    * In directory src/system I added the script "allcross".  This script
1580    invokes sml and cross-compiles the compiler for all supported
1581    architectures.  (Useful when providing a new set of boot files.)
1582    
1583    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1584    added a small tweak to pickle-util.sml to work around this problem,
1585    but it is not a proper fix yet.  I will investigate further.  (The
1586    effect of the bug was an inflation of library pickle size.)
1587    
1588    * Version number increased to 110.28.1 (to avoid compatibility problems).
1589    
1590    ----------------------------------------------------------------------
1591    Name: Allen Leung
1592    Date: 2000/05/25 17:28 EDT
1593    Tag: leunga-20000525-ra
1594    Description:
1595    
1596      Fixed a bug in freezing phase of the register allocator.
1597    
1598    ----------------------------------------------------------------------
1599    Name: Allen Leung
1600    Date: 2000/05/15 22:53 EDT
1601    Tag: leunga-20000515-alpha-x86-ra
1602    Description:
1603    
1604      1. Alpha
1605    
1606          Slight cleanup.  Removed the instruction SGNXL
1607    
1608      2. X86
1609    
1610          Added the following instructions to the instruction set:
1611    
1612            ROLx, RORx,
1613            BTx, BTSx, BTLx, BTRx,
1614            XCHGx, and variants with the LOCK prefix
1615    
1616      3. Register Allocation
1617    
1618          The module ra-rewrite-with-renaming has been improved.
1619    
1620      These have no effect on SML/NJ.
1621    
1622    ----------------------------------------------------------------------
1623    Name: Matthias Blume
1624    Date: 2000/05/15 16:20:00 JST
1625    Tag: blume-20000515-lightrebuild
1626    Description:
1627    
1628    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1629       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1630       defined.  In effect, the command will not build any cross-compiler
1631       backends and therefore finish more quickly.
1632    
1633       The "fixpt" script also takes a "-light" switch to be able to use
1634       this new facility while compiling for a fixpoint.
1635    
1636    2. I replaced all mentions of anchored paths in group owner specifications
1637       with simple relative paths (usually starting with "..").
1638       The rationale is that a library's internal workings should not be
1639       compromised by the lack of some anchor.  (An anchor is necessary
1640       for someone who wants to refer to the library by an anchored path,
1641       but it should not be necessary to build the same library in the first
1642       place.)
1643    
1644    3. I changed the way CM's tool mechanism determines the shell command
1645       string used for things like ml-yacc etc. so that it does not break
1646       when CM.Control.implicit_anchors is turned off.
1647    
1648    ----------------------------------------------------------------------
1649    Name: Matthias Blume
1650    Date: 2000/05/12 18:20:00 JST
1651    Tag: blume-20000512-ml-build
1652    Description:
1653    
1654    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1655    getting installed properly (by config/install.sh).
1656    
1657    ----------------------------------------------------------------------
1658    Name: Matthias Blume
1659    Date: 2000/05/12 17:30:00 JST
1660    Tag: blume-20000512-anchors
1661    Description:
1662    
1663    !!! NEW BOOT FILES !!!
1664    
1665    This change is in preparation of fading out support for "implicitly
1666    anchored path names".  I went through all sources and used the
1667    explicit (and relatively new) $-notation.  See system/README and the
1668    CM manual for more info on this.
1669    
1670    I also modified the anchoring scheme for some things such as "smlnj",
1671    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1672    anchors are more expressive: anchor name and first arc do not have to
1673    coincide.  This entails the following user-visible change:
1674    
1675    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1676    particular, when you fire up sml with a command-line argument, say,
1677    e.g.:
1678    
1679       sml '$smlnj/cmb.cm'
1680    
1681    At the ML toplevel prompt:
1682    
1683       CM.autoload "$smlnj/cmb.cm";
1684    
1685    There is also a new controller in CM.Control that can be used to turn
1686    off all remaining support for implicit anchors by saying:
1687    
1688        CM.autoload "$smlnj/
1689        #set CM.Control.implicit_anchors false;
1690    
1691    This causes CM to reject implicitly anchored paths.  This is (for the
1692    time being) less permissive than the "final" version where there will
1693    be no more such implicit anchors and relative paths will be just that:
1694    relative.
1695    
1696    The next step (version after next version?) will be to make the
1697    default for CM.Control.implicit_anchors false.  After the dust has
1698    settled, I can then produce the "final" version of this...
1699    
1700    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1701    
1702    ----------------------------------------------------------------------
1703    Name: Matthias Blume
1704    Date: 2000/05/11 16:30:00 JST
1705    Tag: blume-20000511-sources
1706    Description:
1707    
1708    The main change is that I added function CM.sources as a generalized
1709    version of the earlier CM.makedepend.  This entails the following
1710    additional changes:
1711    
1712      - CM.makedepend has been dropped.
1713    
1714      - CM manual has been updated.
1715    
1716      - TOOLS signature and API have been changed.
1717    
1718    ----------------------------------------------------------------------
1719    Name: Allen Leung
1720    Date: 2000/05/10 21:17 EDT
1721    Tag: leunga-20000510-moby-c--ssa
1722    Description:
1723    
1724      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1725    None of these affect SML/NJ.
1726    
1727    1. Register Allocation
1728    
1729        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1730           This module tries to remove local (i.e. basic block level) redundancies
1731           during spilling.
1732    
1733        b. A new framework for performing region based register allocation.
1734           Not yet entirely functional.
1735    
1736    2. X86
1737    
1738       a. DefUse for POP was missing the stack pointer [found by Lal]
1739       b. Reload for CALL was incorrect in X86Spill [found by John]
1740       c. Various fixes in X86Spill so that it can be used correctly for
1741          the new spilling module.
1742    
1743    3. SSA/IR
1744    
1745       a. New module ir/dj-dataflow.sml implements elimination based
1746          data flow analysis.
1747    
1748    4. MLRiscGen
1749    
1750       a. Fix for gc type annotation
1751    
1752    5. MDGen
1753    
1754       Various fixes for machine description -> ml code translation.  For ssa
1755       only.
1756    
1757    ----------------------------------------------------------------------
1758    Name: Allen Leung
1759    Date: 2000/05/08 22:17 EDT
1760    Tag: leunga-20000508-labexp
1761    Description:
1762    
1763      Fermin has found a few assembly problems with constant expressions
1764      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1765      which choke on dumb assemblers.  This is his fix.
1766    
1767    ----------------------------------------------------------------------
1768    Name: Dave MacQueen
1769    Date: 2000/04/09 14:00 EDT
1770    Tag: dbm-20000502-Version_110_28
1771    Description:
1772    
1773    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1774    
1775    2. Updated config/version to 110.28
1776    
1777    3. Updated config/srcarchiveurl
1778    
1779    3. New boot files!
1780       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1781    
1782    ----------------------------------------------------------------------
1783    Name: Matthias Blume
1784    Date: 2000/05/01 19:05:00 JST
1785    Tag: blume-20000501-noweb
1786    Description:
1787    
1788    A new noweb tool has been added.  The existing system is entirely
1789    unaffected by this, but some CM users have asked for renewed noweb
1790    support.  Everything is documented in the CM manual.
1791    
1792    New (plugin) libraries:
1793    
1794       noweb-tool.cm
1795       nw-ext.cm
1796    
1797    ----------------------------------------------------------------------
1798    Name: Dave MacQueen
1799    Date: 2000/04/30 12:40PM EDT
1800    Tag: dbm-20000430-bug_fixes
1801    Description:
1802    
1803    1. Fix for bug 1498
1804       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1805       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1806         added toRealArray function
1807       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1808         added check for tag Obj.RealArray to array printing case in ppObj
1809    
1810    2. Fix for bug 1510
1811       smlnj/src/compiler/Semant/types/typesutil.sml
1812         fixed definition of dummyargs (used by equalTycon) so that
1813         dummy args are distinct types
1814    
1815    ----------------------------------------------------------------------
1816    Name: Matthias Blume
1817    Date: 2000/04/30 01:00:00 JST
1818    Tag: blume-20000430-versions
1819    Description:
1820    
1821    1. CM version numbering added.  This is an implementation of Lal's
1822       proposal for adding version numbers and version checking to .cm
1823       files.  Lal said that his proposal was just that -- a proposal.
1824       For the time being I went ahead and implemented it so that people
1825       can comment on it.  Everything is completely backward-compatible
1826       (except for the stable library format, i.e., new bootfiles!).
1827    
1828       As usual, see the CM manual for details.
1829    
1830    2. An alternative syntax for anchored paths has been implemented.
1831       Dave has recently voiced the same concerns that I had when I did
1832       this, so there should be some support.  My take is that eventually
1833       I will let support for the current syntax (where anchors are
1834       "implicit") fade out in favor of the new, explicit syntax.
1835       In order to be backward-compatible, both old and new syntax are
1836       currently supported.
1837    
1838       Again, see the CM manual for details.
1839    
1840    3. Parallel make is trying to be slightly smarter:  When the master
1841       process finds a "bottleneck", i.e., when there is only one
1842       compilation unit that can be compiled and everybody else is
1843       waiting on it, then it will simply compile it directly instead
1844       of clumsily telling one of the slaves to do it.
1845    
1846    4. Support for "unsharing" added.  This is necessary in order to be
1847       able to have two different versions of the same library running
1848       at the same time (e.g., for trying out a new MLRISC while still
1849       having the old MLRISC linked into the current compiler, etc.)
1850       See the CM manual.
1851    
1852    5. Simple "makedepend" functionality added for generating Makefile
1853       dependency information.  (This is rather crude at the moment.
1854       Expect some changes here in the future.)
1855    
1856    6. ".fun" added as a recognized suffix for ML files. Also documented
1857       explicitly in the manual that the fallback behavior (unknown suffix
1858       -> ML file) is not an official feature!
1859    
1860    7. Small changes to the pickler for stable libraries.
1861    
1862    8. Several internal changes to CM (for cleanup/improvement).
1863    
1864    
1865    !!!! NEW BINFILES !!!!
1866    
1867    ----------------------------------------------------------------------
1868    Name: Matthias Blume
1869    Date: 2000/04/28 17:30:00 JST
1870    Tag: blume-20000428-pathconfig
1871    Description:
1872    
1873    1. I changed config/install.sh to remove duplicate entries from the
1874       lib/pathconfig file at the end.  Moreover, the final version of
1875       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1876       in src/system/installml.
1877    
1878    2. The config/install.sh script now consistently uses relative
1879       pathnames in lib/pathconfig whenever the anchor is in the lib
1880       directory.  (So far this was true for the libraries that come
1881       pre-compiled and bundled as part of the bootfiles but not for
1882       libraries that are compiled by the script itself.)
1883    
1884    ----------------------------------------------------------------------
1885    Name: Matthias Blume
1886    Date: 2000/04/26 13:10:00 JST
1887    Tag: blume-20000426-fun_suffix
1888    Description:
1889    
1890    Added ".fun" as a recognized file name suffix (for ML code).
1891    
1892    ----------------------------------------------------------------------
1893    Name: Allen Leung
1894    Date: 2000/04/25 17:00:00 EST
1895    Tag: leunga-20000425-alpha-ra
1896    Description:
1897    
1898    1. Alpha
1899    
1900        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1901    in C--.
1902    
1903    2. RA
1904    
1905       Added a flag "ra-dump-size" to print out the size of the flowgraph
1906       and the interference graph.
1907    
1908    ----------------------------------------------------------------------
1909    Name: Dave MacQueen
1910    Date: 2000/04/25/
1911    Tag: dbm-20000425-mlyacc_doc_examples
1912    Description:
1913      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1914      Updated all three examples in src/ml-yacc/examples to run
1915      under 110.* using CM.make.
1916    
1917    ----------------------------------------------------------------------
1918    Name: Allen Leung
1919    Date: 2000/04/20 23:04:00 EST
1920    Tag: leunga-20000420-ssa-c---stuff
1921    Description:
1922    
1923      This update synchronizes my repository with Yale's.  Most of these
1924    changes, however, do not affect SML/NJ at all (the RA is an exception).
1925    
1926    1. Register Allocator
1927    
1928       a. An improvement in the interference graph construction:
1929          Given a copy
1930    
1931                s <- t
1932    
1933          no interference edge between s and t is added for this definition of s.
1934    
1935       b. I've added two new spill heuristic modules that Fermin and I developed
1936          (in the new library RA.cm). These are unused in SML/NJ but maybe
1937          useful for others (Moby?)
1938    
1939    2. X86
1940    
1941       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1942    
1943    3. Alpha
1944    
1945       a. Added the BSR instruction and code generation that goes with it [C--]
1946       b. Other fixes too numerous to recount provided by Fermin [C--]
1947    
1948    4. Regmaps
1949    
1950       a. The regmaps are not initialized with the identity physical bindings
1951          at creation time.  This is unneeded.
1952    
1953    5. MLRISC Optimizations
1954    
1955       a. The DJ-Graph module can now compute the iterated dominance frontiers
1956          intersects with liveness incrementally in linear time! Woohoo!
1957          This is now used in my new SSA construction algorithm.
1958    
1959       b. THe branch reorganization module is now smarter about linear chains of
1960          basic blocks.
1961    
1962    
1963    ----------------------------------------------------------------------
1964    Name: Matthias Blume
1965    Date: 2000/04/12 13:52:00 JST
1966    Tag: blume_main_v110p27_1
1967    Description:
1968    
1969    Changed install.sh script to handle archive files without version number
1970    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1971    name of the boot file archive.
1972    
1973    ----------------------------------------------------------------------
1974    Name: Dave MacQueen
1975    Date: 2000/04/09 14:00 EDT
1976    Tag: dbm-20000410-Version_110_27
1977    Description:
1978    
1979    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1980    
1981    2. Updated src/config/version to 110.27
1982    
1983    3. New boot files!
1984    
1985    ----------------------------------------------------------------------
1986    Name: Allen Leung
1987    Date: 2000/04/09 19:09:00 EST
1988    Tag: leunga-20000409-misc
1989    Description:
1990    
1991    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1992    
1993    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1994    
1995    ----------------------------------------------------------------------
1996    Name: Stefan
1997    Date: 2000/04/07 10:00:00 EDT
1998    Tag: monnier-20000406-branch-handling
1999    Description:
2000    
2001    Improved handling of branches (mostly those generated from
2002    polymorphic equality), removed switchoff and changed the
2003    default optimization settings (more cpsopt and less flintopt).
2004    
2005    ----------------------------------------------------------------------
2006    Name: Allen Leung
2007    Date: 2000/04/06 01:30:00 EST
2008    Tag: leunga-20000406-peephole-x86-SSA-2
2009    Description:
2010    
2011       Forgot a few files.
2012    
2013    ----------------------------------------------------------------------
2014    Name: Allen Leung
2015    Date: 2000/04/06 00:36:00 EST
2016    Tag: leunga-20000406-peephole-x86-SSA
2017    Description:
2018    
2019    1.  New Peephole code
2020    
2021    2.  Minor improvement to X86 instruction selection
2022    
2023    3.  Various fixes to SSA and machine description -> code translator
2024    
2025    ----------------------------------------------------------------------
2026    Name: Matthias Blume
2027    Date: 2000/04/05 12:30:00 JST
2028    Tag: blume_main_v110p26p2_3
2029    Description:
2030    
2031    This update just merges three minor cosmetic updates to CM's sources
2032    to get ready for the 110.27 code freeze on Friday.  No functionality
2033    has changed.
2034    
2035    ----------------------------------------------------------------------
2036    Name: Allen Leung
2037    Date: 2000/04/04 19:39:00 EST
2038    Tag: leunga-20000404-x86-asm
2039    Description:
2040    
2041    1.  Fixed a problem in X86 assembly.
2042    
2043        Things like
2044    
2045           jmp %eax
2046           jmp (%eax)
2047    
2048        should be output as
2049    
2050           jmp *%eax
2051           jmp *(%eax)
2052    
2053    2.  Assembly output
2054    
2055          Added a new flag
2056    
2057              "asm-indent-copies" (default to false)
2058    
2059          When this flag is on, parallel copies will be indented an extra level.
2060    
2061    ----------------------------------------------------------------------
2062    Name: Allen Leung
2063    Date: 2000/04/04 03:18:00 EST
2064    Tag: leunga-20000404-C--Moby
2065    Description:
2066    
2067        All of these fixes are related to C--, Moby, and my own optimization
2068        stuff; so they shouldn't affect SML/NJ.
2069    
2070    1.  X86
2071    
2072        Various fixes related floating point, and extensions.
2073    
2074    2.  Alpha
2075    
2076        Some extra patterns related to loads with signed/zero extension
2077        provided by Fermin.
2078    
2079    3.  Assembly
2080    
2081        When generating assembly, resolve the value of client defined constants,
2082        instead of generating symbolic values.  This is controlled by the
2083        new flag "asm-resolve-constants", which is default to true.
2084    
2085    4.  Machine Descriptions
2086    
2087        a. The precedence parser was slightly broken when parsing infixr symbols.
2088        b. The type generalizing code had the bound variables reversed, resulting
2089           in a problem during arity raising.
2090        c. Various fixes in machine descriptions.
2091    
2092    ----------------------------------------------------------------------
2093    Name: Matthias Blume
2094    Date: 2000/04/03 16:05:00 JST
2095    Tag: blume_main_v110p26p2_2
2096    Description:
2097    
2098    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
2099    now done via the ordinary static environment that is context to each
2100    compilation unit.
2101    
2102    To this end, I arranged that instead of "structure Core" as "structure
2103    _Core" is bound in the pervasive environment.  Core access is done via
2104    _Core (which can never be accidentally rebound because _Core is not a
2105    legal surface-syntax symbol).
2106    
2107    The current solution is much cleaner because the core environment is
2108    now simply part of the pervasive environment which is part of every
2109    compilation unit's context anyway.  In particular, this eliminates all
2110    special-case handling that was necessary until now in order to deal
2111    with dynamic and symbolic parts of the core environment.
2112    
2113    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2114    compilation manager's bootstrap compiler (actually: in the "init group"
2115    handling).  See the comments in src/system/smlnj/init/init.cmi for
2116    more details.
2117    
2118    I also tried to track down all mentions of "Core" (as string argument
2119    to Symbol.strSymbol) in the compiler and replaced them with a
2120    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
2121    name appears in one place only.
2122    
2123    Binfile and bootfile format have not changed, but the switchover from
2124    the old "init.cmi" to the new one is a bit tricky, so I supplied new
2125    bootfiles anyway.
2126    
2127    ----------------------------------------------------------------------
2128    Name: Allen Leung
2129    Date: 2000/04/02 21:17:00 EST
2130    Tag: leunga-20000402-mltree
2131    Description:
2132    
2133       1. Renamed the constructor CALL in MLTREE by popular demand.
2134       2. Added a bunch of files from my repository.  These are currently
2135          used by other non-SMLNJ backends.
2136    
2137    ----------------------------------------------------------------------
2138    Name: Allen Leung
2139    Date: 2000/03/31 21:15:00 EST
2140    Tag: leunga-20000331-aliasing
2141    Description:
2142    
2143    This update contains a rewritten (and hopefully more correct) module
2144    for extracting aliasing information from CPS.
2145    
2146       To turn on this feature:
2147    
2148            Compiler.Control.CG.memDisambiguate := true
2149    
2150       To pretty print the region information with assembly
2151    
2152           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
2153    
2154       To control how many levels of aliasing information are printed, use:
2155    
2156           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
2157    
2158       The default of n is 3.
2159    
2160    ----------------------------------------------------------------------
2161    Name: David MacQueen
2162    Date: 2000/03/31 11:15:00 EST
2163    Tag: dbm-20000331-runtime_fix
2164    Description:
2165    
2166    This update contains:
2167    
2168    1. runtime/c-lib/c-libraries.c
2169       includes added in revision 1.2 caused compilation errors on hppa-hpux
2170    
2171    2. fix for bug 1556
2172       system/Basis/Implementation/NJ/internal-signals.sml
2173    
2174    ----------------------------------------------------------------------
2175    Name: Matthias Blume
2176    Date: 2000/03/31 18:00:00 JST
2177    Tag: blume_main_v110p26p2_1
2178    Description:
2179    
2180    This update contains:
2181    
2182    1. A small change to CM's handling of stable libraries:
2183       CM now maintains one "global" modmap that is used for all stable
2184       libraries.  The use of such a global modmap maximizes sharing and
2185       minimizes the need for re-traversing parts of environments during
2186       modmap construction.  (However, this has minor impact since modmap
2187       construction seems to account for just one percent or less of total
2188       compile time.)
2189    
2190    2. I added a "genmap" phase to the statistics.  This is where I got the
2191       "one percent" number (see above).
2192    
2193    3. CM's new tool parameter mechanism just became _even_ better. :)
2194       - The parser understands named parameters and recursive options.
2195       - The "make" and "shell" tools use these new features.
2196         (This makes it a lot easier to cascade these tools.)
2197       - There is a small syntax change: named parameters use a
2198    
2199           <name> : ( <option> ... )            or
2200           <name> : <string>
2201    
2202         syntax.  Previously, named parameters were implemented in an
2203         ad-hoc fashion by each tool individually (by parsing strings)
2204         and had the form
2205    
2206           <name>=<string>
2207    
2208       See the CM manual for a full description of these issues.
2209    
2210    ----------------------------------------------------------------------
2211    Name: Matthias Blume
2212    Date: 2000/03/30 18:00:00 JST
2213    Tag: blume_main_v110p26p2_0
2214    Description:
2215    
2216    !!!!! WARNING !!!!!!
2217    !!  New binfiles  !!
2218    !!!!!!!!!!!!!!!!!!!!
2219    
2220    This update contains:
2221    
2222    1. Moderate changes to CM:
2223    
2224       - Changes to CM's tools mechanism.  In particular, it is now possible
2225       to have tools that accept additional "command line" parameters
2226       (specified in the .cm file at each instance where the tool's class is
2227       used).
2228    
2229       This was done to accommodate the new "make" and "shell" tools which
2230       facilitate fairly seamless hookup to portions of code managed using
2231       Makefiles or Shell scripts.
2232    
2233       There are no classes "shared" or "private" anymore.  Instead, the
2234       sharing annotation is now a parameter to the "sml" class.
2235    
2236       There is a bit of generic machinery for implementing one's own
2237       tools that accept command-line parameters.  However, I am not yet fully
2238       satisfied with that part, so expect changes here in the future.
2239    
2240       All existing tools are described in the CM manual.
2241    
2242       - Slightly better error handling.  (CM now suppresses many followup
2243       error messages that tended to be more annoying than helpful.)
2244    
2245    2. Major changes to the compiler's static environment data structures.
2246    
2247       - no CMStaticEnv anymore.
2248            - no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
2249              but it is called Environment), no conversions between different
2250              kinds of static environments
2251    
2252       - There is still a notion of a "modmap", but such modmaps are generated
2253         on demand at the time when they are needed.  This sounds slow, but I
2254         sped up the code that generates modmaps enough for this not to lead to
2255         a slowdown of the compiler (at least I didn't detect any).
2256    
2257       - To facilitate rapid modmap generation, static environments now
2258         contain an (optional) "modtree" structure.  Modtree annotations are
2259         constructed by the unpickler during unpickling.  (This means that
2260         the elaborator does not have to worry about modtrees at all.)
2261         Modtrees have the advantage that they are compositional in the same
2262         way as the environment data structure itself is compositional.
2263         As a result, modtrees never hang on to parts of an environment that
2264         has already been rendered "stale" by filtering or rebinding.
2265    
2266       - I went through many, many trials and errors before arriving at the
2267         current solution.  (The initial idea of "linkpaths" did not work.)
2268         But the result of all this is that I have touched a lot of files that
2269         depend on the "modules" and "types" data structures (most of the
2270         elaborator). There were a lot of changes during my "linkpath" trials
2271         that could have been reverted to their original state but weren't.
2272         Please, don't be too harsh on me for messing with this code a bit more
2273         than what was strictly necessary...  (I _did_ resist the tempation
2274         of doing any "global reformatting" to avoid an untimely death at
2275         Dave's hands. :)
2276    
2277       - One positive aspect of the previous point:  At least I made sure that
2278         all files that I touched now compile without warnings (other than
2279         "polyEqual").
2280    
2281       - compiler now tends to run "leaner" (i.e., ties up less memory in
2282         redundant modmaps)
2283    
2284    ----------------------------------------------------------------------
2285    Name: Allen Leung
2286    Date: 2000/03/29 18:00:00
2287    Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
2288    Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
2289    Description:
2290    
2291       This update contains *MAJOR* changes to the way code is generated from CPS
2292    in the module mlriscGen, and in various backend modules.
2293    
2294    CHANGES
2295    =======
2296    
2297    1. MLRiscGen: forward propagation fix.
2298    
2299       There was a bug in forward propagation introduced at about the same time
2300       as the MLRISC x86 backend, which prohibits coalescing to be
2301       performed effectively in loops.
2302    
2303       Effect: speed up of loops in RISC architectures.
2304               By itself, this actually slowed down certain benchmarks on the x86.
2305    
2306    2. MLRiscGen:  forward propagating addresses from consing.
2307    
2308       I've changed the way consing code is generated.  Basically I separated
2309       out the initialization part:
2310    
2311            store tag,   offset(allocptr)
2312            store elem1, offset+4(allocptr)
2313            store elem2, offset+8(allocptr)
2314            ...
2315            store elemn, offset+4n(allocptr)
2316    
2317       and the address computation part:
2318    
2319            celladdr <- offset+4+alloctpr
2320    
2321       and move the address computation part
2322    
2323       Effect:  register pressure is generally lower as a result.  This
2324                makes compilation of certain expressions much faster, such as
2325                long lists with non-trivial elements.
2326    
2327                 [(0,0), (0,0), .... (0,0)]
2328    
2329    3. MLRiscGen: base pointer elimination.
2330    
2331        As part of the linkage mechanism, we generate the sequence:
2332    
2333         L:  ...  <- start of the code fragment
2334    
2335         L1:
2336             base pointer <- linkreg - L1 + L
2337    
2338         The base pointer was then used for computing relocatable addresses
2339       in the code fragment.  Frequently (such as in lots of continuations)
2340       this is not needed.  We now eliminate this sequence whenever possible.
2341    
2342         For compile time efficiency, I'm using a very stupid local heuristic.
2343       But in general, this should be done as a control flow analysis.
2344    
2345       Effect:  Smaller code size.  Speed up of most programs.
2346    
2347    4. Hppa back end
2348    
2349         Long jumps in span dependence resolution used to depend on the existence
2350      of the base pointer.
2351    
2352         A jump to a long label L was expanded into the following sequence:
2353    
2354          LDIL %hi(L-8192), %r29
2355          LDO  %lo(L-8192)(%r29), %r29
2356          ADD  %r29, baseptr, %r29
2357          BV,n %r0(%r29)
2358    
2359         In the presence of change (3) above, this will not work.  I've changed
2360       it so that the following sequence of instructions are generated, which
2361       doesn't mention the base pointer at all:
2362    
2363             BL,n  L', %r29           /* branch and link, L' + 4 -> %r29 */
2364        L':  ADDIL L-(L'+4), %r29     /* Compute address of L */
2365             BV,n  %r0(%r29)          /* Jump */
2366    
2367    5. Alpha back end
2368    
2369          New alpha instructions LDB/LDW have been added, as per Fermin's
2370       suggestions.   This is unrelated to all other changes.
2371    
2372    6. X86 back end
2373    
2374         I've changed andl to testl in the floating point test sequence
2375         whenever appropriate.  The Intel optimization guide states that
2376         testl is preferable to andl.
2377    
2378    7. RA (x86 only)
2379    
2380         I've improved the spill propagation algorithm, using an approximation
2381       of maximal weighted independent sets.   This seems to be necessary to
2382       alleviate the negative effect in light of the slow down in (1).
2383    
2384         I'll write down the algorithm one of these days.
2385    
2386    8. MLRiscGen: frequencies
2387    
2388         I've added an annotation that states that all call gc blocks have zero
2389       execution frequencies.  This improves register allocation on the x86.
2390    
2391    BENCHMARKS
2392    ==========
2393    
2394       I've only perform the comparison on 110.25.
2395    
2396       The platforms are:
2397    
2398        HPPA  A four processor HP machine (E9000) with 5G of memory.
2399        X86   A 300Hhz Pentium II with 128M of memory, and
2400        SPARC An Ultra sparc 2 with 512M of memory.
2401    
2402       I used the following parameters for the SML benchmarks:
2403    
2404                 @SMLalloc
2405         HPPA    256k
2406         SPARC   512k
2407         X86     256k
2408    
2409    COMPILATION TIME
2410    ----------------
2411       Here are the numbers comparing the compilation times of the compilers.
2412       I've only compared 110.25 compiling the new sources versus
2413       a fixpoint version of the new compiler compiling the same.
2414    
2415                     110.25                                  New
2416               Total  Time in RA  Spill+Reload   Total  Time In RA Spill+Reload
2417         HPPA   627s    116s        2684+3584     599s    95s       1003+1879
2418         SPARC  892s    173s        2891+3870     708s    116s      1004+1880
2419         X86    999s    315s       94006+130691   987s    296s    108877+141957
2420    
2421                   110.25         New
2422                Code Size      Code Size
2423         HPPA   8596736         8561421
2424         SPARC  8974299         8785143
2425         X86    9029180         8716783
2426    
2427       So in summary, things are at least as good as before.   Dramatic
2428       reduction in compilation is obtained on the Sparc; I can't explain it,
2429       but it is reproducible.  Perhaps someone should try to reproduce this
2430       on their own machines.
2431    
2432    SML BENCHMARKS
2433    --------------
2434    
2435        On the average, all benchmarks perform at least as well as before.
2436    
2437          HPPA         Compilation Time     Spill+Reload      Run Time
2438                     110.25  New            110.25    New   110.25  New
2439    
2440          barnesHut  3.158  3.015  4.75%    1+1       0+0   2.980  2.922   2.00%
2441              boyer  6.152  5.708  7.77%    0+0       0+0   0.218  0.213   2.34%
2442       count-graphs  1.168  1.120  4.32%    0+0       0+0  22.705 23.073  -1.60%
2443                fft  0.877  0.792 10.74%    1+3       1+3   0.602  0.587   2.56%
2444        knuthBendix  3.180  2.857 11.32%    0+0       0+0   0.675  0.662   2.02%
2445             lexgen  6.190  5.290 17.01%    0+0       0+0   0.913  0.788  15.86%
2446               life  0.803  0.703 14.22%   25+25      0+0   0.153  0.140   9.52%
2447              logic  2.048  2.007  2.08%    6+6       1+1   4.133  4.008   3.12%
2448         mandelbrot  0.077  0.080 -4.17%    0+0       0+0   0.765  0.712   7.49%
2449             mlyacc 22.932 20.937  9.53%  154+181    32+57  0.468  0.430   8.91%
2450            nucleic  5.183  5.060  2.44%    2+2       0+0   0.125  0.120   4.17%
2451      ratio-regions  3.357  3.142  6.84%    0+0       0+0  116.225 113.173 2.70%
2452                ray  1.283  1.290 -0.52%    0+0       0+0   2.887  2.855   1.11%
2453             simple  6.307  6.032  4.56%   28+30      5+7   3.705  3.658   1.28%
2454                tsp  0.888  0.862  3.09%    0+0       0+0   7.040  6.893   2.13%
2455               vliw 24.378 23.455  3.94%  106+127    25+45  2.758  2.707   1.91%
2456      --------------------------------------------------------------------------
2457       Average                     6.12%                                   4.09%
2458    
2459          SPARC        Compilation Time     Spill+Reload      Run Time
2460                     110.25  New            110.25    New   110.25  New
2461    
2462          barnesHut  3.778  3.592  5.20%    2+2       0+0   3.648  3.453    5.65%
2463              boyer  6.632  6.110  8.54%    0+0       0+0   0.258  0.242    6.90%
2464       count-graphs  1.435  1.325  8.30%    0+0       0+0  33.672 34.737   -3.07%
2465                fft  0.980  0.940  4.26%    3+9       2+6   0.838  0.827    1.41%
2466        knuthBendix  3.590  3.138 14.39%    0+0       0+0   0.962  0.967   -0.52%
2467             lexgen  6.593  6.072  8.59%    1+1       0+0   1.077  1.078   -0.15%
2468               life  0.972  0.868 11.90%   26+26      0+0   0.143  0.140    2.38%
2469              logic  2.525  2.387  5.80%    7+7       1+1   5.625  5.158    9.05%
2470         mandelbrot  0.090  0.093 -3.57%    0+0       0+0   0.855  0.728   17.39%
2471             mlyacc 26.732 23.827 12.19%  162+189    32+57  0.550  0.560   -1.79%
2472            nucleic  6.233  6.197  0.59%    3+3       0+0   0.163  0.173   -5.77%
2473      ratio-regions  3.780  3.507  7.79%    0+0       0+0 133.993 131.035   2.26%
2474                ray  1.595  1.550  2.90%    1+1       0+0   3.440  3.418    0.63%
2475             simple  6.972  6.487  7.48%   29+32      5+7   3.523  3.525   -0.05%
2476                tsp  1.115  1.063  4.86%    0+0       0+0   7.393  7.265    1.77%
2477               vliw 27.765 24.818 11.87%  110+135    25+45  2.265  2.135    6.09%
2478      ----------------------------------------------------------------------------
2479       Average                     6.94%                                    2.64%
2480    
2481          X86          Compilation Time     Spill+Reload      Run Time
2482                     110.25  New            110.25    New   110.25  New
2483    
2484          barnesHut  5.530  5.420  2.03%  593+893   597+915   3.532  3.440   2.66%
2485              boyer  8.768  7.747 13.19%  493+199   301+289   0.327  0.297  10.11%
2486       count-graphs  2.040  2.010  1.49%  298+394   315+457  26.578 28.660  -7.26%
2487                fft  1.327  1.302  1.92%  112+209   115+210   1.055  0.962   9.71%
2488        knuthBendix  5.218  5.475 -4.69%  451+598   510+650   0.928  0.932  -0.36%
2489             lexgen  9.970  9.623  3.60% 1014+841  1157+885   0.947  0.928   1.97%
2490               life  1.183  1.183  0.00%  162+182   145+148   0.127  0.103  22.58%
2491              logic  3.285  3.512 -6.45%  514+684   591+836   5.682  5.577   1.88%
2492         mandelbrot  0.147  0.143  2.33%   38+41     33+54    0.703  0.690   1.93%
2493             mlyacc 35.457 32.763  8.22% 3496+4564 3611+4860  0.552  0.550   0.30%
2494            nucleic  7.100  6.888  3.07%  239+168   201+158   0.175  0.173   0.96%
2495      ratio-regions  6.388  6.843 -6.65% 1182+257   981+300  120.142 120.345 -0.17%
2496                ray  2.332  2.338 -0.29%  346+398   402+494   3.593  3.540   1.51%
2497             simple  9.912  9.903  0.08% 1475+941  1579+1168  3.057  3.178  -3.83%
2498                tsp  1.623  1.532  5.98%  266+200   250+211   8.045  7.878   2.12%
2499               vliw 33.947 35.470 -4.29% 2629+2774 2877+3171  2.072  1.890   9.61%
2500      ----------------------------------------------------------------------------
2501       Average                     1.22%                                     3.36%
2502    
2503    ----------------------------------------------------------------------
2504    Name: Allen Leung
2505    Date: 2000/03/23 16:25:00
2506    Tag: leunga-20000323-fix_x86_alpha
2507    Description:
2508    
2509    1. X86 fixes/changes
2510    
2511       a.  The old code generated for SETcc was completely wrong.
2512           The Intel optimization guide is VERY misleading.
2513    
2514    2. ALPHA fixes/changes
2515    
2516       a.  Added the instructions LDBU, LDWU, STB, STW as per Fermin's suggestion.
2517       b.  Added a new mode byteWordLoadStores to the functor parameter to Alpha()
2518       c.  Added reassociation code for address computation.
2519    
2520    ----------------------------------------------------------------------
2521    Name: Allen Leung
2522    Date: 2000/03/22 01:23:00
2523    Tag: leunga-20000322-fix_x86_hppa_ra
2524    Description:
2525    
2526    1. X86 fixes/changes
2527    
2528       a.  x86Rewrite bug with MUL3 (found by Lal)
2529       b.  Added the instructions FSTS, FSTL
2530    
2531    2. PA-RISC fixes/changes
2532    
2533       a.  B label should not be a delay slot candidate!  Why did this work?
2534       b.  ADDT(32, REG(32, r), LI n) now generates one instruction instead of two,
2535           as it should be.
2536       c.  The assembly syntax for fstds and fstdd was wrong.
2537       d.  Added the composite instruction COMICLR/LDO, which is the immediate
2538           operand variant of COMCLR/LDO.
2539    
2540    3. Generic MLRISC
2541    
2542       a.  shuffle.sml rewritten to be slightly more efficient
2543       b.  DIV bug in mltree-simplify fixed (found by Fermin)
2544    
2545    4. Register Allocator
2546    
2547       a.  I now release the interference graph earlier during spilling.
2548           May improve memory usage.
2549    
2550    ----------------------------------------------------------------------
2551    Name: Matthias Blume
2552    Date: 2000/03/14 14:15:32
2553    Tag: blume_main_v110p26p1_2
2554    Description:
2555    
2556    1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2557    additional argument called "template" which is an optional string that
2558    specifies the layout of the tool command line.  See the CM manual for
2559    explanation.
2560    
2561    2. A special-purpose tool can be "registered" by simply dropping the
2562    corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2563    directory where the .cm file lives that uses this tool.  (The
2564    behavior/misfeature until now was to look for the tool description
2565    files in the current working directory.)  As before, tool description
2566    files could also be anchored -- in which case they can live anywhere
2567    they like.  Following the recent e-mail discussion, this change should
2568    make it easier to have special-purpose tools that are shipped together
2569    with the sources of the program that uses them.
2570    
2571  ----------------------------------------------------------------------  ----------------------------------------------------------------------
2572  Name: Matthias Blume  Name: Matthias Blume
2573  Date: 2000/03/10 07:48:34  Date: 2000/03/10 07:48:34
# Line 45  Line 2602 
2602  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
2603  <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,
2604  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
2605  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2606  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2607  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2608    
# Line 95  Line 2652 
2652    it from that remote directory.    it from that remote directory.
2653    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2654    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2655    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2656    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2657    
2658    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.577  
changed lines
  Added in v.833

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