Home My Page Projects Code Snippets Project Openings SML/NJ
Summary Activity Forums Tracker Lists Tasks Docs Surveys News SCM Files

SCM Repository

[smlnj] Diff of /sml/trunk/HISTORY
ViewVC logotype

Diff of /sml/trunk/HISTORY

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

revision 607, Mon Apr 10 16:31:03 2000 UTC revision 832, Tue May 22 20:30:36 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 16:30:00 EDT
18    Tag: blume-20010522-minitut
19    Description:
20    
21    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
22       struct returns.
23    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
24       incomplete, preliminary documentation for NLFFI.
25    
26    ----------------------------------------------------------------------
27    Name: Matthias Blume
28    Date: 2001/05/14 11:30:00 EDT
29    Tag: blume-20010514-script
30    Description:
31    
32    Some bugs in install script fixed.
33    
34    In addition to that I also made a slight change to the NLFFI API:
35    Functors generated by ml-nlffigen now take the dynamic library as a
36    straight functor argument, not as a suspended one.  (The original
37    functor code used to force the suspension right away anyway, so there
38    was nothing gained by this complication of the interface.)
39    
40    ----------------------------------------------------------------------
41    Name: Matthias Blume
42    Date: 2001/05/11 14:35:00 EDT
43    Tag: blume-20010511-ml-nlffi
44    Description:
45    
46    I finally took the plunge and added my new FFI code to the main
47    repository.  For x86-linux it is now ready for prime-time.
48    
49    There are two new subdirectories of "src":
50    
51      - ml-nlffi-lib:
52           The utility library for programs using the FFI interface.
53           Here is the implementation of $/c.cm and its associated low-level
54           partners $/c-int.cm and $/memory.cm.
55      - ml-nlffigen:
56           A stand-alone program for generating ML glue code from C source
57           code.
58    
59    Building ml-nlffigen requires $/ckit-lib.cm.
60    
61    The config/install.sh script has been updates to do the Right Thing
62    (hopefully).
63    
64    Notice that the source tree for the C-Kit will not be put under "src"
65    but directly under the installation root directory.  (This is the
66    structure that currently exists on the CVS server when you check out
67    module "sml".)  Fortunately, config/install.sh knows about this oddity.
68    
69    Bugs: No documentation yet.
70    
71    ----------------------------------------------------------------------
72    Name: Matthias Blume
73    Date: 2001/05/09 16:35:00 EDT
74    Tag: blume-20010509-cpscontract
75    Description:
76    
77    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
78    wrapper/unwrapper elimination did not decrement usage counts and some
79    dead variables got overlooked by the dead-up logic.)
80    
81    ----------------------------------------------------------------------
82    Name: Lal George
83    Date: 2001/05/08  17:26:09 EDT
84    Tag: george-20010508-omit-frameptr
85    Description:
86    
87    Changes to implement the omit-frame-pointer optimization to support
88    raw C calls. For now, there is only support on the Intel x86, but
89    other architectures will follow as more experience is gained with this.
90    
91    
92    ----------------------------------------------------------------------
93    Name: Matthias Blume
94    Date: 2001/05/07 14:40:00 EDT
95    Tag: blume-20010507-proxies
96    Description:
97    
98    I made into "proxy libraries" all libraries that qualify for such a
99    change.  (A qualifying library is a library that has another library or
100    groups as its sole member and repeats that member's export list
101    verbatim.  A proxy library avoids this repetition by omitting its export
102    list, effectively inheriting the list that its (only) member exports.
103    See the CM manual for more explanation.)
104    The main effect is that explicit export lists for these libraries
105    do not have to be kepts in sync, making maintenance a bit easier.
106    
107    I also added copyright notices to many .cm-files.
108    
109    Last but not least, I made a new set of bootfiles.
110    
111    ----------------------------------------------------------------------
112    Name: Matthias Blume
113    Date: 2001/05/04 17:00:00 EDT
114    Tag: blume-20010504-cm-lsplit
115    Description:
116    
117    0. John merged pending changes to $/smlnj-lib.cm
118    
119    1. Allen's previous change accidentally backed out of one of Lal's
120       earlier changes.  I undid this mistake (re-introducing Lal's change).
121    
122    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
123       within the compiler where applicable.  There is some code simplification
124       because of that.
125    
126    3. The "split" phase (in FLINT) is now part of the default list of phases.
127       Compiler.Control.LambdaSplitting.* can be used to globally control the
128       lambda-splitting (cross-module-inlining) engine.  In addition to that,
129       it can now also be controlled on a per-source basis: CM has been taught
130       a new tool parameter applicable to ML source files.
131    
132       - To turn lambda-splitting off completely:
133            local open Compiler.Control.LambdaSplitting in
134                val _ = set Off
135            end
136       - To make "no lambda-splitting" the global default (but allow per-source
137         overriding); this is the initial setting:
138            local open Compiler.Control.LambdaSplitting in
139                val _ = set (Default NONE)
140            end
141       - To make "lambda-splitting with aggressiveness a" the global default
142         (and allow per-source overriding):
143            local open Compiler.Control.LambdaSplitting in
144                val _ = set (Default (SOME a))
145            end
146    
147       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
148         write (in the respective .cm-file):
149            a.sml (lambdasplitting:off)
150       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
151            a.sml (lambdasplitting:on)
152       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
153         <a> is a decimal non-negative integer):
154            a.sml (lambdasplitting:<a>)
155       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
156            a.sml (lambdasplitting:infinity)
157       - To use the global default for a.sml:
158            a.sml (lambdasplitting:default)
159         or simply
160            a.sml
161    
162    ----------------------------------------------------------------------
163    Name: Allen Leung
164    Date: 2001/05/04 01:57:00 EDT
165    Tag: leunga-20010504-sync
166    Description:
167    
168      MLRISC features.
169    
170      1. Fix to CMPXCHG instructions.
171      2. Changed RA interface to allow annotations in callbacks.
172      3. Added a new method to the stream interface to allow annotations updates.
173    
174    ----------------------------------------------------------------------
175    Name: Matthias Blume
176    Date: 2001/05/01 11:45:00 EDT
177    Tag: blume-20010501-pcedittmp
178    Description:
179    
180    Changed install.sh to use the current working directory instead of
181    /usr/tmp for a temporary file (pcedittmp).  The previous choice
182    of /usr/tmp caused trouble with MacOS X because of file premission
183    problems.
184    
185    ----------------------------------------------------------------------
186    Name: Matthias Blume
187    Date: 2001/04/20 11:10:00 EDT
188    Tag: blume-20010420-inMLflag
189    Description:
190    
191     - added vp_limitPtrMask to vproc-state.h
192       (for use by the raw-C-calls mechanism to implement proper interrupt
193        handling)
194     - made the ML compiler aware of various data-structure offsets so it
195       can generate code for accessing the vp_inML flag and vp_limitPtrMask
196     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
197       raw C-calls
198    
199    ----------------------------------------------------------------------
200    Name: Lal George
201    Date: 2001/04/20 09:15:28 EDT
202    Tag: george-20010420-macosX
203    Description:
204    
205     - Changes to port to Mac OS X; Darwin.
206    
207     - In the process I found that sqrt was broken on the PPC, because the
208       fsqrt instruction is not implemented.
209    
210    ----------------------------------------------------------------------
211    Name: Matthias Blume
212    Date: 2001/04/18 12:45:00 EDT
213    Tag: blume-20010418-ccalls
214    Description:
215    
216     - fixed two off-by-4 errors in the x86-specific c-calls implementation
217       (this bug prevented structure arguments containing pointers from being
218        passed correctly)
219     - changed the raw-C-call code in mlriscGen.sml in such a way that
220       structure arguments are represented as a pointer to the beginning
221       of the structure (instead of having a series of synthesized arguments,
222       one for each structure member)
223    
224     - made makeml script's verbosity level configurable via environment
225       variable (MAKEML_VERBOSITY)
226    
227     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
228       in rawmem-x86.sml; we are now using the real thing
229    
230    ----------------------------------------------------------------------
231    Name: Matthias Blume
232    Date: 2001/03/22 16:25:00 EST
233    Tag: blume-20010322-bootfiles
234    Description:
235    
236    Created a new set of bootfiles (for your automatic installation convenience).
237    
238    ----------------------------------------------------------------------
239    Name: Matthias Blume
240    Date: 2001/03/22 15:10:00 EST
241    Tag: blume-20010322-rawmem-parcm
242    Description:
243    
244    1. All "raw memory access" primitives for the new FFI are implemented now
245       (at least on the x86).
246    2. Some further cleanup of CM's parallel make mechanism.
247    
248    ----------------------------------------------------------------------
249    Name: Matthias Blume
250    Date: 2001/03/19 17:53:00 EST
251    Tag: blume-20010319-parallel
252    Description:
253    
254    Parallel make (using compile servers) now works again.
255    
256    To this end, CM.stabilize and CMB.make have been modified to work in
257    two passes when compile servers are attached:
258       1. Compile everything, do not perform stabilization; this pass
259          uses compile servers
260       2. Stabilize everything; this pass does not use compile servers
261    If there are no compile servers, the two passes are combined into one
262    (as before).  Splitting the passes increases the inherent parallelism
263    in the dependency graph because the entire graph including all
264    libraries is available at the same time.  This, in turn, improves
265    server utilization.  The downside is that the master process will
266    have to do some extra work after compilation is done (because for
267    technical reasons it must re-read all the binfiles during stabilization).
268    
269    ----------------------------------------------------------------------
270    Name: Matthias Blume
271    Date: 2001/03/16 12:22:00 EST
272    Tag: blume-20010316-bootfiles
273    Description:
274    
275    Created a new set of bootfiles (for your automatic installation convenience).
276    
277    ----------------------------------------------------------------------
278    Name: Matthias Blume
279    Date: 2001/03/16 11:00:00 EST
280    Tag: blume-20010316-MLTREE-fixup
281    Description:
282    
283    This is a minor fixup for an (untagged) earlier commit by Allen.
284    (A file was missing).
285    
286    ----------------------------------------------------------------------
287    Name: Allen Leung
288    Date: Mon Mar  5 18:54:57 EST 2001
289    Tag: leunga-20010305-cut-support
290    
291    1. New support for alternative control-flow in MLTREE.
292       Currently we support
293    
294          FLOW_TO(CALL ...., [k1,...,kn])
295    
296       This is needed for 'cuts to' in C-- and try/handle-like constructs
297       in Moby
298    
299       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
300    
301    2. Register Allocator
302    
303       Changes in interface [from Fermin, John]
304    
305    3. Alpha 8-bit SLL support [Fermin]
306    
307    4. All architectures
308    
309       A new module (ClusterExpandCopies) for expanding parallel copies.
310    
311    ----------------------------------------------------------------------
312    Name: Allen Leung
313    Date: 2001/02/27 23:07:00 EST
314    Tag: leunga-20010227-minor-stuff
315    
316    1. Alpha bug fix for CMOVNE
317    2. Handle mltree COND(..,FCMP ...,...)
318    3. Bug fix in simplifier
319    
320    ----------------------------------------------------------------------
321    Name: Matthias Blume
322    Date: 2001/01/30 17:50:00 EST
323    Tag: blume-20010130-sync
324    Description:
325    
326    This is just a minor update to sync my devel branch with the main brach.
327    The only visible change is the addition of some README files.
328    
329    ----------------------------------------------------------------------
330    Name: Matthias Blume
331    Date: 2001/01/12 23:30:00 JST
332    Tag: blume-20010112-bootfiles
333    Description:
334    
335    Made a new set of bootfiles that goes with the current state of the
336    repository.
337    
338    ----------------------------------------------------------------------
339    Name: Matthias Blume
340    Date: 2001/01/12 21:20:00 JST
341    Tag: blume-20010112-sync
342    Description:
343    
344    I am just flushing out some minor changes that had accumulated in
345    my private branch in order to sync with the main tree.  (This is
346    mainly because I had CVS trouble when trying to merge _into_ my
347    private branch.)
348    
349    Most people should be completely unaffected by this.
350    
351    ----------------------------------------------------------------------
352    Name: Allen Leung
353    Date: Thu Jan 11 21:03:00 EST 2001
354    Tag: leunga-20010111-labexp=mltree
355    Description:
356    
357    1.  Removed the type LabelExp and replace it by MLTree.
358    2.  Rewritten mltree-simplify with the pattern matcher tool.
359    3.  There were some bugs in alpha code generator which would break
360        64-bit code generation.
361    4.  Redo the tools to generate code with the
362    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
363        are now generated by perl scripts.
364    
365    ----------------------------------------------------------------------
366    Name: Matthias Blume
367    Date: 2001/01/10 21:55:00 JST
368    Tag: blume-20010110-rcc
369    Description:
370    
371    The RCC stuff now seems to work (but only on the x86).
372    This required hacking of the c-calls interface (and -implementation) in
373    MLRISC.
374    
375    Normal compiler users should be unaffected.
376    
377    ----------------------------------------------------------------------
378    Name: Matthias Blume
379    Date: 2001/01/09 01:20:00 JST
380    Tag: blume-20010109-rcc
381    Description:
382    
383    This is a fairly big patch, flushing out a large number of pending
384    changes that I made to my development copy over the last couple of days.
385    
386    Of practical relevance at this moment is a workaround for a pickling
387    bug that Allen ran into the other day.  The cause of the bug itself is
388    still unknown and it might be hard to fix it properly, but the
389    workaround has some merits of its own (namely somewhat reducing pickling
390    overhead for certain libraries).  Therefore, I think this solution should
391    be satisfactory at this time.
392    
393    The rest of the changes (i.e., the vast majority) has to do with my
394    ongoing efforts of providing direct support for C function calls from
395    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
396    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
397    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
398    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
399    handle RCC.
400    
401    The last part (the changes to mlriscGen.sml) are still known to be
402    wrong on the x86 and not implemented on all other architectures.  But
403    the infrastructure is in place. I had to change a few functor
404    signatures in the backend to be able to route the CCalls interface
405    from MLRISC there, and I had to specialize the mltree type (on the
406    x86) to include the necessary extensions. (The extensions themselves
407    were already there and redy to go in MLRISC/x86).
408    
409    Everything should be very happy as soon as someone helps me with
410    mlriscGen.sml...
411    
412    In any case, nothing of this should matter to anyone as long as the
413    new primop is not being used (which is going to be the case unless you
414    find it where I hid it :). The rest of the compiler is completely
415    unaffected.
416    
417    ----------------------------------------------------------------------
418    Name: Matthias Blume
419    Date: 2001/01/05 00:30:00 JST
420    Tag: blume-20010105-primops
421    Description:
422    
423    Added some experimental support for work that I am doing right now.
424    These changes mostly concern added primops, but there is also a new
425    experimental C library in the runtime system (but currently not enabled
426    anywhere except on Linux/X86).
427    
428    In the course of adding primops (and playing with them), I discovered that
429    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
430    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
431    To recover, I made type information in INL_PRIM mandatory and changed
432    prim.sml as well as built-in.sml accordingly.  The InLine structure now
433    has complete, correct type information (i.e., no bottom types).
434    
435    Since all these changes mean that we need new binfiles, I also bumped the
436    version number to 110.32.1.
437    
438    ----------------------------------------------------------------------
439    Name: Matthias Blume
440    Date: 2000/12/30 22:10:00 JST
441    Tag: blume-20001230-various
442    Description:
443    
444    Added proxy libraries for MLRISC and let MLRISC libraries refer
445    to each other using path anchors.  (See CM manual for explanation.)
446    
447    Updated CM documentation.
448    
449    Fixed some bugs in CM.
450    
451    Implemented "proxy" libraries (= syntactic sugar for CM).
452    
453    Added "-quiet" option to makeml and changed runtime system accordingly.
454    
455    Added cleanup handler for exportML to reset timers and compiler stats.
456    
457    ----------------------------------------------------------------------
458    Name: Lal George
459    Date: 2000/12/22 22:22:58 EST 2000
460    Tag: Release_110_32
461    Description:
462    
463            Infinite precision used throughout MLRISC.
464            see MLRISC/mltree/machine-int.sig
465    
466    ----------------------------------------------------------------------
467    Name: Matthias Blume
468    Date: 2000/12/22 23:16:00 JST
469    Tag: blume-20001222-warn
470    Description:
471    
472    Corrected wording and formatting of some CM warning message which I
473    broke in my previous patch.
474    
475    ----------------------------------------------------------------------
476    Name: Matthias Blume
477    Date: 2000/12/22 21:20:00 JST
478    Tag: blume-20001222-anchorenv
479    Description:
480    
481    Fixed CM's handling of anchor environments in connection with CMB.make.
482    
483    ----------------------------------------------------------------------
484    Name: Matthias Blume
485    Date: 2000/12/22 13:15:00 JST
486    Tag: blume-20001222-cleanup
487    Description:
488    
489    Removed src/cm/ffi which does not (and did not) belong here.
490    
491    ----------------------------------------------------------------------
492    Name: Matthias Blume
493    Date: 2000/12/21 23:55:00 JST
494    Tag: blume-20001221-exn
495    Description:
496    
497    Probably most important: CM no longer silently swallows all exceptions
498    in the compiler.
499    Plus: some other minor CM changes.  For example, CM now reports some
500    sizes for generated binfiles (code, data, envpickle, lambdapickle).
501    
502    ----------------------------------------------------------------------
503    Name: Matthias Blume
504    Date: 2000/12/15 00:01:05 JST
505    Tag: blume-20001215-dirtool
506    Description:
507    
508    - "dir" tool added.
509    - improvements and cleanup to Tools structure
510    - documentation updates
511    
512    ----------------------------------------------------------------------
513    Name: Allen Leung
514    Date: Thu Dec 14 03:45:24 EST 2000
515    Description:
516    Tag:  leunga-20001214-int-inf
517    Description:
518    
519       In IntInf, added these standard functions, which are missing from our
520    implementation:
521    
522        andb : int * int -> int
523        xorb : int * int -> int
524        orb  : int * int -> int
525        notb : int -> int
526         <<   : int * word -> int
527        ~>>  : int * word -> int
528    
529       Not tested, I hope they are correct.
530    
531    ----------------------------------------------------------------------
532    Name: Allen Leung
533    Date: Fri Dec  8 19:23:26 EST 2000
534    Description:
535    Tag:  leunga-20001208-nowhere
536    Description:
537    
538      Slight improvements to the 'nowhere' tool to handle OR-patterns,
539    to generate better error messages etc.  Plus a brief manual.
540    
541    ----------------------------------------------------------------------
542    Name: Lal George
543    Date: 2000/12/08 09:54:02 EST 2000
544    Tag: Release_110_31
545    Description:
546    
547    - Version 110.31
548    ----------------------------------------------------------------------
549    Name: Allen Leung
550    Date: Thu Dec  7 22:01:04 EST 2000
551    Tag:  leunga-20001207-cell-monster-hack
552    Description:
553    
554    Major MLRISC internal changes.  Affect all clients.
555    Summary:
556    
557    1.  Type CELLS.cell = int is now replaced by a datatype.
558        As a result, the old regmap is now gone.  Almost all interfaces
559        in MLRISC change as a consequence.
560    
561    2.  A new brand version of machine description tool (v3.0) that generates
562        modules expecting the new interface.  The old version is removed.
563    
564    3.  The RA interface has been further abstracted into two new functors.
565        RISC_RA and X86RA.  These functors have much simpler interfaces.
566        [See also directory MLRISC/demo.]
567    
568    4.  Some other new source->source code generation tools are available:
569    
570        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
571        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
572           I use this tool to generate the peephole optimizers---with the new
573           cell type changes, peephole rules are becoming difficult to write
574           without conditional pattern matching.
575    
576    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
577        cover the entire MLRISC source tree so many things broke.
578    
579    6.  CM files have been moved to the subdirectory MLRISC/cm.
580        They are moved because there are a lot of them and they clutter up the
581        root dir.
582    
583    7.  More detailed documentation to come...
584    
585        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
586        a makeml -rebuild first.  This is because of other other
587        changes that Matthias has made (see below).
588    
589    
590    ----------------------------------------------------------------------
591    Name: Matthias Blume
592    Date: 2000/11/30 23:12:00 JST
593    Tag: blume-20001130-filereorg
594    Description:
595    
596    Some manual updates and some file reorganizations in CM.
597    
598    ----------------------------------------------------------------------
599    Name: Matthias Blume
600    Date: 2000/11/24 17:45:00 JST
601    Tag: blume-20001124-link
602    Description:
603    
604    Drastically improved link traversal code for the case that the dynamic
605    value was already loaded at bootstrap time.  As a result, CM and CMB
606    now both load blazingly fast -- even on a very slow machine.  Also,
607    memory consumption has been further reduced by this.
608    
609    Warning: The format of the PIDMAP file has changed.  THerefore, to
610    bootstrap you have to do this:
611    
612    1. Run CMB.make
613    2. Make a symbolic link for the boot directory:
614         ln -s sml.boot.ARCH-OS xxx
615    3. "Rebuild" the boot directory:
616         ./makeml -boot xxx -rebuild sml ; rm xxx
617    4. Boot normally:
618          ./makeml
619    
620    ----------------------------------------------------------------------
621    Name: Matthias Blume
622    Date: 2000/11/21 21:20:00 JST
623    Tag: blume-20001121-tools
624    Description:
625    
626    Continued hacking on autoloading problem -- with success this time.
627    Also changed tool-plugin mechanism.  See new CM manual.
628    
629    ----------------------------------------------------------------------
630    Name: Matthias Blume
631    Date: 2000/11/19 14:30:00 JST
632    Tag:  blume-20001119-autoload
633    Description:
634    
635    Some hacking to make autoloading faster.  Success for CMB, no success
636    so far for CM.  There is a reduced structure CM' that autoloads faster.
637    (This is a temporary, non-documented hack to be eliminated again when
638    the general problem is solved.)
639    
640    ----------------------------------------------------------------------
641    Name: Matthias Blume
642    Date: 2000/11/17 14:10:00 JST
643    Tag: blume-20001117-pickle-lib
644    Description:
645    
646    1. Eliminated comp-lib.cm
647    2. Made pickle-lib.cm
648    3. Eliminated all uses of intset.sml (from comp-lib.cm)
649    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
650       equivalent constructs from smlnj-lib.cm (INtHashTable).
651    5. Point 4. also goes for those uses of intmap.* in MLRISC.
652       Duplicated intmap modules thrown out.
653    6. Hunted down all duplicated SCC code and replaced it with
654       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
655    7. Rewrote Feedback module.
656    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
657       should be thrown out and equivalent modules from smlnj-lib.cm
658       should be used (IntRedBlackSet, IntListSet, ...).
659    
660    Confirmed that compiler compiles to fixpoint.
661    
662    ----------------------------------------------------------------------
663    Name: Allen Leung
664    Date: 2000/11/10 18:00:00
665    Tag: leunga-20001110-new-x86-fp
666    
667    A new x86 floating point code generator has been added.
668    By default this is turned off.  To turn this on, do:
669    
670        CM.autoload "$smlnj/compiler.cm";
671        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
672    
673    Changes:
674    
675    1.  Changed FTAN to FPTAN so that the assembly output is correct.
676    2.  Changed the extension callback for FTANGENT to generate:
677    
678              fptan
679              fstp  %st(0)
680        instead of
681              fptan
682              fstpl ftempmem
683    
684    3.  Numerous assembly fixes for x86.
685    
686    5.  Cleaned up the machine code output module x86/x86MC.sml and added
687        support for a whole bunch of instructions and addressing modes:
688    
689          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
690          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
691          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
692          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
693          fxch %st(n)
694          fld %st(n)
695          fst %st(n)
696          fst mem
697          fstp %st(n)
698          fucom %st(n)
699          fucomp %st(n)
700    
701        All these are now generated when the fast fp mode is turned on.
702    
703    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
704    
705    ----------------------------------------------------------------------
706    Name: Matthias Blume
707    Date: 2000/11/09 11:20:00 JST
708    Tag: blume-20001109-scc
709    Description:
710    
711    Eliminated some code duplication:
712    
713    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
714       (Otherwise the functor is useless.)
715    2. Used GraphSCCFn where SCCUtilFun was used previously.
716    3. Got rid of SCCUtilFun (in comp-lib.cm).
717    
718    ----------------------------------------------------------------------
719    Name: Lal George
720    Date: 2000/11/06 09:02:21 EST 2000
721    Tag: Release_110_30
722    Description:
723    
724    - Version 110.30
725    ----------------------------------------------------------------------
726    Name: Matthias Blume
727    Date: 2000/11/04 14:45:00
728    Tag: blume-20001104-mlbuild
729    Description:
730    
731    - Made ml-build faster on startup.
732    - Documentation fixes.
733    
734    ----------------------------------------------------------------------
735    Name: Matthias Blume
736    Date: 2000/11/02 17:00:00 JST
737    Tag: blume-20001102-condcomp
738    Description:
739    
740    - Small tweaks to pickler -- new BOOTFILES!
741    - Version bumped to 110.29.2.
742    - Added conditional compilation facility to init.cmi (see comment there).
743    ----------------------------------------------------------------------
744    Name: Allen Leung
745    Date: 2000/10/23 19:31:00
746    Tag: leunga-20001023-demo-ra
747    
748    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
749    2. Test programs for the graph library updated
750    3. Some new MLRISC demo programs added
751    
752    ----------------------------------------------------------------------
753    Name: Matthias Blume
754    Date: 2000/08/31 22:15:00 JST
755    Tag: blume-20001017-errmsg
756    Description:
757    
758    More error message grief: Where there used to be no messages, there
759    now were some that had bogus error regions.  Fixed.
760    
761    ----------------------------------------------------------------------
762    Name: Matthias Blume
763    Date: 2000/08/31 17:30:00 JST
764    Tag: blume-20001017-v110p29p1
765    Description:
766    
767    I made a version 110.29.1 with new bootfiles.
768    
769    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
770              CM documentation changes and a small bugfix in CM's error reporting.
771    
772    ----------------------------------------------------------------------
773    Name: Lal George
774    Date: 2000/09/27 14:42:35 EDT
775    Tag: george-20000927-nodestatus
776    Description:
777    
778    Changed the type of the nodestatus, so that:
779    
780            SPILLED(~1)             is now SPILLED
781            SPILLED(m) where m>=0   is now MEMREG(m)
782            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
783    
784    ----------------------------------------------------------------------
785    Name: Matthias Blume
786    Date: 2000/09/07 14:45:00 JST
787    Tag: blume-20000907-cmerrmsg
788    Description:
789    
790    Small tweak to CM to avoid getting ML syntax error messages twice.
791    
792    ----------------------------------------------------------------------
793    Name: Matthias Blume
794    Date: 2000/08/31 18:00:00 JST
795    Tag: blume-20000831-cvsbootfiles
796    Description:
797    
798    New URL for boot files (because the 110.29 files on the BL server do
799    now work correctly with my updated install scripts for yacc and lex).
800    
801    ----------------------------------------------------------------------
802    Name: Matthias Blume
803    Date: 2000/08/08 12:33:00 JST
804    Tag: blume-20000808-manual
805    Description:
806    
807    Tiny update to CM manual.
808    
809    ----------------------------------------------------------------------
810    Name: Allen Leung
811    Date: 2000/08/7 19:31:00
812    Tag: leunga-20000807-a-whole-bunch-of-stuff
813    
814      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
815    and Mobdy.
816    
817    1.  x86
818    
819       a.  Fixes to peephole module by John and Dan.
820       b.  Assembly fix to SETcc by Allen.
821       c.  Fix to c-call by John.
822       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
823       e.  Instruction selection optimization to SETcc as suggested by John.
824    
825           For example,
826    
827            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
828    
829           should generate:
830    
831            MOVL a, x
832            SUBL b, x
833            SHRL 31, x
834    
835    2.  IR stuff
836    
837         A bunch of new DJ-graph related algorithms added.  These
838         speed up SSA construction.
839    
840    3.  SSA + Scheduling
841    
842         Added code for SSA and scheduling to the repository
843    
844    ----------------------------------------------------------------------
845    Name: Lal George
846    Date: 2000/07/27 11:53:14 EDT
847    
848    Tag: lal-20000727-linux-ppc
849    Description:
850    
851     Made changes to support Linux PPC.
852     p.s. I have confirmation that the 110.29 boot files work fine.
853    
854    ----------------------------------------------------------------------
855    Name: Matthias Blume
856    Date: 2000/07/27 17:40:00 JST
857    Tag: blume-20000727-scripts
858    Description:
859    
860    !!!! WARNING !!!!
861    You must recompile the runtime system!
862    !!!! WARNING !!!!
863    
864    This is basically another round of script-enhancements:
865    
866    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
867       and undefine CM preprocessor symbols.
868    
869    2. ml-build avoids generating a new heap image if it finds that the
870       existing one is still ok.  (The condition is that no ML file had to
871       be recompiled and all ML files are found to be older that the heap
872       file.)
873    
874       To make this work smoothly, I also hacked the runtime system as
875       well as SMLofNJ.SysInfo to get access to the heap image suffix
876       (.sparc-solaris, ...) that is currently being used.
877    
878       Moreover, the signature of CM.mk_standalone has changed.  See the
879       CM manual.
880    
881    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
882       CM manual for details.)
883    
884    4. More CM manual updates:
885        - all of the above has been documented.
886        - there is now a section describing the (CM-related) command line
887          arguments that are accepted by the "sml" command
888    
889    ----------------------------------------------------------------------
890    Name: Matthias Blume
891    Date: 2000/07/25 16:20:00 JST
892    Tag: blume-20000725-makedepend
893    Description:
894    
895    Added a script called ml-makedepend.  This can be used in makefiles
896    for Unix' make in a way very similar to the "makedepend" command for
897    C.
898    
899    The script internally uses function CM.sources.
900    
901    Synopsis:
902    
903        ml-makedepend [-f makefile] cmfile targetname
904    
905    The default for the makefile is "makefile" (or "Makefile" should
906    "makefile" not exist).
907    
908    ml-makedepend adds a cmfile/targetname-specific section to this
909    makefile (after removing the previous version of this section).  The
910    section contains a single dependency specification with targetname on
911    the LHS (targetname is an arbitrary name), and a list of files derived
912    from the cmfile on the RHS.  Some of the files on the RHS are
913    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
914    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
915    path names.  The makefile writer is responsible for making sure that
916    these variables have correct at the time "make" is invoked.
917    
918    ----------------------------------------------------------------------
919    Name: Matthias Blume
920    Date: 2000/07/22 23:30:00 JST
921    Tag: blume-20000722-urlupdate
922    Description:
923    
924    Changed BOOT and config/srcarchiveurl to point to BL server:
925    
926        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
927    
928    ----------------------------------------------------------------------
929    Name: Matthias Blume
930    Date: 2000/07/18 18:00:00 JST
931    Tag: blume-20000718-Version_110_29
932    Description:
933    
934    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
935    
936    2. Updated config/version to 110.29
937    
938    3. Updated config/srcarchiveurl
939    
940    3. New boot files!
941       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
942    
943    ----------------------------------------------------------------------
944    Name: Matthias Blume
945    Date: 2000/07/11 13:58:00 JST
946    Tag: blume-20000711-doctypo
947    Description:
948    
949    Fixed a few typos in CM manual.
950    
951    ----------------------------------------------------------------------
952    Name: Allen Leung
953    Date: 2000/06/15 00:38:00
954    Tag: leunga-20000704-sparc-x86
955    
956    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
957    2. fix to x86 RET bug [found by Dan Grossman]
958    3. sparc assembly bug fix for ticc instructions [found by Fermin]
959    
960       Affects c-- and moby only
961    
962    ----------------------------------------------------------------------
963    Name: Matthias Blume
964    Date: 2000/07/04 15:26:00
965    Tag: blume-20000704-trigger
966    Description:
967    
968    1. Improvements to CM manual.
969    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
970       of getting a back-trace.  The function, when called, raises an
971       internal exception which explicitly carries the full back-trace history,
972       so it is unaffected by any intervening handle-raise pairs ("trivial"
973       or not).  The interactive loop will print that history once it arrives
974       at top level.
975       Short of having all exceptions implicitly carry the full history, the
976       recommended way of using this facility is:
977         - compile your program with instrumentation "on"
978         - run it, when it raises an exception, look at the history
979         - if the history is "cut off" because of some handler, go and modify
980           your program so that it explicitly calls BTrace.trigger
981         - recompile (still instrumented), and rerun; look at the full history
982    
983    ----------------------------------------------------------------------
984    Name: Matthias Blume
985    Date: 2000/07/03 15:36:00 JST
986    Tag: blume-20000702-manual
987    Description:
988    
989    Small corrections and updates to CM manual.
990    
991    ----------------------------------------------------------------------
992    Name: Matthias Blume
993    Date: 2000/06/29 16:04:00 JST
994    Tag: blume-20000629-yacctool
995    Description:
996    
997    Changes:
998    
999    1. Class "mlyacc" now takes separate arguments to pass options to
1000       generated .sml- and .sig-files independently.
1001    2. Corresponding CM manual updates.
1002    3. BTrace module now also reports call sites.  (However, for loop clusters
1003       it only shows from where the cluster was entered.)  There are associated
1004       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1005    
1006    ----------------------------------------------------------------------
1007    Name: Matthias Blume
1008    Date: 2000/06/27 16:51:00 JST
1009    Tag: blume-20000627-noweb
1010    Description:
1011    
1012    Changes:
1013    
1014     1. Implemented "subdir" and "witness" options for noweb tool.
1015        This caused some slight internal changes in CM's tool implementation.
1016     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
1017        some remaining issues from earlier path anchor changes.
1018     3. Updated CM manual accordingly.
1019    
1020     4. Changed implementation of back-tracing so that I now consider it
1021        ready for prime-time.
1022    
1023        In particular, you don't have to explicitly trigger the back-trace
1024        anymore.  Instead, if you are running BTrace-instrumented code and
1025        there is an uncaught exception (regardless of whether or not it was
1026        raised in instrumented code), the top-level evalloop will print
1027        the back-trace.
1028    
1029        Features:
1030    
1031          - Instrumented and uninstrumented code work together seemlessly.
1032            (Of course, uninstrumented code is never mentioned in actual
1033             back-traces.)
1034    
1035          - Asymptotic time- and space-complexity of instrumented code is
1036            equal to that of uninstrumented code.  (This means that
1037            tail-recursion is preserved by the instrumentation phase.)
1038    
1039          - Modules whose code has been instrumented in different sessions
1040            work together without problem.
1041    
1042          - There is no penalty whatsoever on uninstrumented code.
1043    
1044          - There is no penalty on "raise" expressions, even in
1045            instrumented code.
1046    
1047        A potential bug (or perhaps it is a feature, too):
1048    
1049          A back-trace reaches no further than the outermost instrumented
1050          non-trivial "raise".  Here, a "trivial" raise is one that is the
1051          sole RHS of a "handle" rule.  Thus, back-traces reach trough
1052    
1053               <exp> handle e => raise e
1054    
1055          and even
1056    
1057               <exp> handle Foo => raise Bar
1058    
1059          and, of course, through
1060    
1061               <exp> handle Foo => ...
1062    
1063         if the exception was not Foo.
1064    
1065         Back-traces always reach right through any un-instrumented code
1066         including any of its "handle" expressions, trivial or not.
1067    
1068       To try this out, do the following:
1069    
1070         - Erase all existing binfiles for your program.
1071           (You may keep binfiles for those modules where you think you
1072            definitely don't need back-tracing.)
1073         - Turn on back-trace instrumentation:
1074              SMLofNJ.Internals.BTrace.mode (SOME true);
1075         - Recompile your program.  (I.e., run "CM.make" or "use".)
1076         - You may now turn instrumentation off again (if you want):
1077              SMLofNJ.Internals.BTrace.mode (SOME false);
1078         - Run your program as usual.  If it raises an exception that
1079           reaches the interactive toplevel, then a back-trace will
1080           automatically be printed.  After that, the toplevel loop
1081           will print the exception history as usual.
1082    
1083    ----------------------------------------------------------------------
1084    Name: Matthias Blume
1085    Date: 2000/06/26 09:56:46 JST
1086    Tag: blume-20000626-setup
1087    Description:
1088    
1089    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1090          ML code before and after compiling a file (e.g., to set compiler
1091          flags)
1092    
1093    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1094              - associated changes to btrace.sml (BTrace instrumentation pass)
1095              - cleaner implementation of btimp.sml (BTrace tracing and report
1096                module)
1097    
1098    CM manual: * new path encoding documented
1099               * description of setup-parameter to "sml" added
1100    
1101    The biggest user-visible change to back-tracing is that it is no
1102    longer necessary to compile all traced modules within the same
1103    session.  (This was a real limitation.)
1104    
1105    ----------------------------------------------------------------------
1106    Name: Matthias Blume
1107    Date: 2000/06/24 12:40:00 JST
1108    Tag: blume-20000624-startup
1109    Description:
1110    
1111    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1112    bit too often -- to put it mildly. :)
1113    
1114    ----------------------------------------------------------------------
1115    Name: Matthias Blume
1116    Date: 2000/06/23 18:20:00 JST
1117    Tag: blume-20000623-btrace
1118    Description:
1119    
1120    This updates adds a backtrace facility to aid programmers in debugging
1121    their programs.  This involves the following changes:
1122    
1123    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1124       keeping track of the current call stack.  When programs are compiled
1125       in a special mode, the compiler will insert calls to these hooks
1126       into the user program.
1127       "Hook" means that it is possible for different implementations of
1128       back-tracing to register themselves (at different times).
1129    
1130    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1131       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1132       It can be turned on using this call:
1133         SMLofNJ.Internals.BTrace.mode (SOME true);
1134       Turning it off again:
1135         SMLofNJ.Internals.BTrace.mode (SOME false);
1136       Querying the current status:
1137         SMLofNJ.Internals.BTrace.mode NONE;
1138       Annotated programs are about twice as big as normal ones, and they
1139       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1140       where all hooks do nothing).  The slowdown with a plugin that is
1141       actually useful (such as the one supplied by default) is even greater,
1142       but in the case of the default plugin it is still only an constant
1143       factor (amortized).
1144    
1145    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1146       with a sub-structure BTrace for controlling back-tracing.  In particular,
1147       the above-mentioned function "mode" controls whether the annotation
1148       phase is invoked by the compiler.  Another important function is
1149       "trigger": when called it aborts the current execution and causes
1150       the top-level loop to print a full back-trace.
1151    
1152    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1153       for back-tracing.  It keeps track of the dynamic call stack and in
1154       addition to that it keeps a partial history at each "level" of that
1155       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1156       d and b (at separate times, dynamically), then the report will show:
1157    
1158       GOTO   d
1159             /c
1160       GOTO  \b
1161       CALL   a
1162    
1163       This shows that there was an initial non-tail call of a, then a
1164       tail-call to b or c, looping behavior in a cluster of functions that
1165       consist of b and c, and then a goto from that cluster (i.e., either from
1166       b or from c) to d.
1167    
1168       Note that (depending on the user program) the amount of information
1169       that the back-trace module has to keep track of at each level is bounded
1170       by a constant.  Thus, the whole implementation has the same asymptotical
1171       complexity as the original program (both in space and in time).
1172    
1173    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1174       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1175       which is raised by the "trigger" function mentioned above.
1176    
1177    Notes on usage:
1178    
1179    - Annotated code works well together with unannotated code:
1180    Unannotated calls simply do not show up at all in the backtrace.
1181    
1182    - It is not a good idea to let modules that were annotated during
1183    different sessions run at the same time.  This is because the compiler
1184    chooses small integers to identify individual functions, and there
1185    will be clashes if different modules were compiled in separate sessions.
1186    (Nothing will crash, and you will even be told about the clashes, but
1187    back-trace information will in general not be useful.)
1188    
1189    - Back-tracing can be confused by callcc and capture.
1190    
1191    - The only way of getting a back-trace right now is to explicitly
1192    invoke the "trigger" function from your user program.  Eventually, we
1193    should make every exception carry back-trace information (if
1194    available).  But since this creates more overhead at "raise"-time
1195    (similar to the current exnHistory overhead), I have not yet
1196    implemented this.  (The implementation will be rather easy.)  With
1197    exceptions carrying back-trace information, this facility will be even
1198    more useful because users don't need to modify their programs...
1199    
1200    - While it is possible to compile the compiler with back-trace
1201    annotations turned on (I did it to get some confidence in
1202    correctness), you must make absolutely sure that core.sml and
1203    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1204    be compiled with annotation because there is no core access yet, but
1205    if you compile btimp.sml with annotation, then the system will go into
1206    an infinite recursion and crash.)
1207    Since CM currently does not know about BTrace, the only way to turn
1208    annotations on and off for different modules of the compiler is to
1209    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1210    this is awkward and clumsy.
1211    
1212    Sample sessions:
1213    
1214    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1215    - SMLofNJ.Internals.BTrace.mode (SOME true);
1216    [autoloading]
1217    [autoloading done]
1218    val it = false : bool
1219    - structure X = struct
1220    -     fun main n = let
1221    -         fun a (x, 0) = d x
1222    -           | a (x, n) = b (x, n - 1)
1223    -         and b (x, n) = c (x, n)
1224    -         and c (x, n) = a (x, n)
1225    -         and d x = e (x, 3)
1226    -         and e (x, 0) = f x
1227    -           | e (x, n) = e (x, n - 1)
1228    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1229    -           | f n = n * g (n - 1)
1230    -         and g n = a (n, 3)
1231    -     in
1232    -         f n
1233    -     end
1234    - end;
1235    structure X : sig val main : int -> int end
1236    - X.main 3;
1237    *** BACK-TRACE ***
1238    GOTO   stdIn:4.2-13.20: X.main[2].f
1239    GOTO-( stdIn:4.2-13.20: X.main[2].e
1240    GOTO   stdIn:4.2-13.20: X.main[2].d
1241         / stdIn:4.2-13.20: X.main[2].a
1242         | stdIn:4.2-13.20: X.main[2].b
1243    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1244    CALL   stdIn:4.2-13.20: X.main[2].g
1245    GOTO   stdIn:4.2-13.20: X.main[2].f
1246    GOTO-( stdIn:4.2-13.20: X.main[2].e
1247    GOTO   stdIn:4.2-13.20: X.main[2].d
1248         / stdIn:4.2-13.20: X.main[2].a
1249         | stdIn:4.2-13.20: X.main[2].b
1250    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1251    CALL   stdIn:4.2-13.20: X.main[2].g
1252    GOTO   stdIn:4.2-13.20: X.main[2].f
1253    GOTO-( stdIn:4.2-13.20: X.main[2].e
1254    GOTO   stdIn:4.2-13.20: X.main[2].d
1255         / stdIn:4.2-13.20: X.main[2].a
1256         | stdIn:4.2-13.20: X.main[2].b
1257    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1258    CALL   stdIn:4.2-13.20: X.main[2].g
1259    GOTO   stdIn:4.2-13.20: X.main[2].f
1260    CALL   stdIn:2.15-17.4: X.main[2]
1261    -
1262    
1263    (Note that because of a FLINt bug the above code currently does not
1264    compile without BTrace turned on.)
1265    
1266    Here is another example, using my modified Tiger compiler:
1267    
1268    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1269    - SMLofNJ.Internals.BTrace.mode (SOME true);
1270    [autoloading]
1271    [autoloading done]
1272    val it = false : bool
1273    - CM.make "sources.cm";
1274    [autoloading]
1275    ...
1276    [autoloading done]
1277    [scanning sources.cm]
1278    [parsing (sources.cm):parse.sml]
1279    [creating directory CM/SKEL ...]
1280    [parsing (sources.cm):tiger.lex.sml]
1281    ...
1282    [wrote CM/sparc-unix/semant.sml]
1283    [compiling (sources.cm):main.sml]
1284    [wrote CM/sparc-unix/main.sml]
1285    [New bindings added.]
1286    val it = true : bool
1287    - Main.compile ("../testcases/merge.tig", "foo.out");
1288    *** BACK-TRACE ***
1289    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
1290    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1291    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1292    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1293    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1294    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1295    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
1296         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
1297    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
1298    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1299    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
1300    CALL   main.sml:1.18-118.4: Main.compile[2]
1301    -
1302    
1303    ----------------------------------------------------------------------
1304    Name: Matthias Blumen
1305    Date: 2000/06/21 18:00:00 JST
1306    Tag: blume-20000621-manual
1307    Description:
1308    
1309    CM manual update: Path environments documented.
1310    
1311    ----------------------------------------------------------------------
1312    Name: Matthias Blume
1313    Date: 2000/06/19 13:40:00
1314    Tag: blume-20000619-manual
1315    Description:
1316    
1317    CM manual and system/README update.  This only covers the fact that
1318    there are no more implicit anchors.  (Path environments and the "bind"
1319    option to "cm" have yet to be documented.)
1320    
1321    ----------------------------------------------------------------------
1322    Name: Matthias Blume
1323    Date: 2000/06/19 11:05:00 JST
1324    Tag: blume-20000619-chdir-bugfix
1325    Description:
1326    
1327    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1328    
1329    ----------------------------------------------------------------------
1330    Name: Matthias Blume
1331    Date: 2000/06/18 22:00:10 JST
1332    Tag: blume-20000618-implicit-anchors-really-gone
1333    Description:
1334    
1335    I updates the previous HISTORY entry where I forgot to mention that
1336    implicit anchors are no longer with us.
1337    
1338    The current update also gets rid of the (now useless) controller
1339    CM.Control.implicit_anchors.
1340    
1341    ----------------------------------------------------------------------
1342    Name: Matthias Blume
1343    Date: 2000/06/16 17:30:00 JST
1344    Tag: blume-20000616-anchorenv
1345    Description:
1346    
1347    This patch implements the long anticipated (just kidding :) "anchor
1348    environment" mechanism.  In the course of doing this, I also
1349    re-implemented CM's internal "SrcPath" module from scratch.  The new
1350    one should be more robust in certain boundary cases.  In any case, it
1351    is a lot cleaner than its predecessor (IMHO).
1352    
1353    This time, although there is yet another boot file format change, I
1354    kept the unpickler backward-compatible.  As a result, no new bootfiles
1355    are necessary and bootstrapping is straightforward.  (You cannot read
1356    new bootfiles into an old system, but the other way around is no
1357    problem.)
1358    
1359    Visible changes:
1360    
1361    ** 0. Implicit path anchors (without the leading $-symbol) are no
1362    longer recognized at all. This means that such path names are not
1363    illegal either.  For example, the name basis.cm simply refers to a
1364    local file called "basis.cm" (i.e, the name is an ordinary path
1365    relative to .cm-files directory).  Or, to put it differently, only
1366    names that start with $ are anchored paths.
1367    
1368    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1369    vanished.
1370    
1371    John (Reppy) had critizised this as soon as I originally proposed and
1372    implemented it, but at that time I did not really deeply believe
1373    him. :) Now I came full-circle because I need the $<singlearc> syntax
1374    in another place where it cannot be seen as an abbreviation for
1375    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1376    seems to mean (i.e., it "expands" into the corresponding anchor
1377    value).
1378    
1379    However, when paths are used as members in CM description files, it
1380    continues to be true that there must be at least another arc after the
1381    anchor.  This is now enforced separately during semantic analysis
1382    (i.e., from a lexical/syntactical point of view, the notation is ok.)
1383    
1384    ** 2. The "cm" class now accepts an option "bind".  The option's value
1385    is a sub-option list of precisely two items -- one labeled "anchor"
1386    and the other one labeled "value".  As you might expect, "anchor" is
1387    used to specify an anchor name to be bound, and "value" specifies what
1388    the anchor is being bound to.
1389    
1390    The value must be a directory name and can be given in either standard
1391    syntax (including the possibility that it is itself an anchored path)
1392    or native syntax.
1393    
1394    Examples:
1395    
1396       foo.cm (bind:(anchor:bar value:$mystuff/bar))
1397       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1398    
1399    and so on.
1400    
1401    The meaning of this is that the .cm-file will be processed with an
1402    augmented anchor environment where the given anchor(s) is/are bound to
1403    the given values(s).
1404    
1405    The rationale for having this feature is this: Suppose you are trying
1406    to use two different (already stable) libraries a.cm and b.cm (that
1407    you perhaps didn't write yourself).  Further, suppose each of these
1408    two libraries internally uses its own auxiliary library $aux/lib.cm.
1409    Normally you would now have a problem because the anchor "lib" can not
1410    be bound to more than one value globally.  Therefore, the project that
1411    uses both a.cm and b.cm must locally redirect the anchor to some other
1412    place:
1413    
1414       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1415       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1416    
1417    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1418    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
1419    
1420    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1421    you when you do so at the time of CM.stabilize).  Therefore, you can
1422    also use an anchored path as the value:
1423    
1424      a.cm (bind:(anchor:lib value:$a-lib))
1425      b.cm (bind:(anchor:lib value:$b-lib))
1426    
1427    Now you can globally configure (using the usual CM.Anchor.anchor or
1428    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1429    itself is always locally bound, setting it globally is no longer
1430    meaningful or necessary (but it does not hurt either).  In fact, "lib"
1431    can still be used as a global anchor for separate purposes.  As a
1432    matter of fact, one can locally define "lib" in terms of a global
1433    "lib":
1434    
1435      a.cm (bind:(anchor:lib value:$lib/a))
1436      b.cm (bind:(anchor:lib value:$lib/b))
1437    
1438    ** 3: The encoding of path names has changed.  This affects the way
1439    path names are shown in CM's progress report and also the internal
1440    protocol encoding used for parallel make.
1441    
1442    The encoding now uses one or more ':'-separated segments.  Each
1443    segments corresponds to a file that has been specified relative to the
1444    file given by its preceding segment.  The first segment is either
1445    relative to the CWD, absolute, or anchored.  Each segment itself is
1446    basically a Unix pathname; all segments but the first are relative.
1447    
1448    Example:
1449    
1450       $foo/bar/baz.cm:a/b/c.sml
1451    
1452    This path denotes the file bar/a/b/c.sml relative to the directory
1453    denoted by anchor "foo".  Notice that the encoding also includes
1454    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1455    paths are resolved relative to the .cm-files directory, so baz.cm must
1456    be ignored to get the "real" pathname.
1457    
1458    To make this fact more obvious, CM puts the names of such "virtual
1459    arcs" into parentheses when they appear in progress reports. (No
1460    parentheses will appear in the internal protocol encoding.)  Thus,
1461    what you really see is:
1462    
1463      $foo/bar/(baz.cm):a/b/c.sml
1464    
1465    I find this notation to be much more informative than before.
1466    
1467    Another new feature of the encoding is that special characters
1468    including parentheses, colons, (back)slashes, and white space are
1469    written as \ddd (where ddd is the decimal encoding of the character).
1470    
1471    *** The CM manual still needs to be updated.
1472    
1473    ----------------------------------------------------------------------
1474    Name: Allen Leung
1475    Date: 2000/06/15 00:38:00
1476    Tag: leunga-20000615-x86-peephole
1477    
1478    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1479    
1480    ----------------------------------------------------------------------
1481    Name: Matthias Blume
1482    Date: 2000/06/12 11:40:00
1483    Tag: blume-20000612-parmakefix
1484    Description:
1485    
1486    More cleanup after changing the file naming scheme: This time I
1487    repaired the parallel make mechanism for CMB.make which I broke earlier.
1488    
1489    ----------------------------------------------------------------------
1490    Name: Allen Leung
1491    Date: 2000/06/09 01:25:00
1492    Tag: leunga-20000609-various
1493    
1494    None of these things should affect normal SML/NJ operations
1495    
1496    1. Peephole improvements provided by Fermin (c--)
1497    2. New annotation DEFUSE for adding extra dependence (moby)
1498    3. New X86 LOCK instructions (moby)
1499    4. New machine description language for reservation tables (scheduling)
1500    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1501       trees etc.)
1502    6. I've changed the CM files so that they can work with versions
1503       110.0.6, 110.25 and 110.28
1504    
1505    ----------------------------------------------------------------------
1506    Name: Matthias Blume
1507    Date: 2000/06/09 12:40:00
1508    Tag: blume-20000609-log
1509    Description:
1510    
1511    - Removed all(?) remaining RCS Log entries from sources.
1512    
1513    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1514      anchored paths).
1515    
1516    ----------------------------------------------------------------------
1517    Name: Matthias Blume
1518    Date: 2000/06/07 17:00:00 JST
1519    Tag: blume-20000607-no-implicit-anchors
1520    Description:
1521    
1522    1. This update changes the default setting for
1523    CM.Control.implicit_anchors from true to false.  This means that
1524    implicit anchors are no longer permitted by default.  I also tried to
1525    make sure that nothing else still relies on implicit anchors.
1526    (This is the next step on the schedule towards a CM that does not even
1527    have the notion of implicit anchors anymore.)
1528    
1529    2. More CM manual updates.
1530    
1531    3. I managed to track down and fix the pickling bug I mentioned last
1532    time.  Because of the previously existing workaround, this entails no
1533    immediate practical changes.
1534    
1535    ----------------------------------------------------------------------
1536    Name: Matthias Blume
1537    Date: 2000/06/06 11:15:00 JST
1538    Tag: blume-20000606-lazierpickle
1539    Description:
1540    
1541    !!!! NEW BOOT FILES !!!!
1542    
1543    * The main purpose of this update is to make library pickles lazier in
1544    order to reduce the initial space penalty for autoloading a library.
1545    As a result, it is now possible to have $smlnj/compiler.cm
1546    pre-registered.  This should take care of the many complaints or
1547    inquiries about missing structure Compiler.  This required changes to
1548    CM's internal data structures and small tweaks to some algorithms.
1549    
1550    As a neat additional effect, it is no longer necessary (for the sake
1551    of lean heap image files) to distinguish between a "minimal" CM and a
1552    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1553    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1554    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1555    
1556    To make the life of compiler-hackers easier, "makeml" now also
1557    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1558    words, after you bootstrap a new sml for the first time, you will not
1559    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1560    around you will still have to do it, though.)
1561    
1562    * A second change consists of major updates to the CM manual.  There
1563    are now several appendices with summary information and also a full
1564    specification of the CM description file syntax.
1565    
1566    * In directory src/system I added the script "allcross".  This script
1567    invokes sml and cross-compiles the compiler for all supported
1568    architectures.  (Useful when providing a new set of boot files.)
1569    
1570    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1571    added a small tweak to pickle-util.sml to work around this problem,
1572    but it is not a proper fix yet.  I will investigate further.  (The
1573    effect of the bug was an inflation of library pickle size.)
1574    
1575    * Version number increased to 110.28.1 (to avoid compatibility problems).
1576    
1577    ----------------------------------------------------------------------
1578    Name: Allen Leung
1579    Date: 2000/05/25 17:28 EDT
1580    Tag: leunga-20000525-ra
1581    Description:
1582    
1583      Fixed a bug in freezing phase of the register allocator.
1584    
1585    ----------------------------------------------------------------------
1586    Name: Allen Leung
1587    Date: 2000/05/15 22:53 EDT
1588    Tag: leunga-20000515-alpha-x86-ra
1589    Description:
1590    
1591      1. Alpha
1592    
1593          Slight cleanup.  Removed the instruction SGNXL
1594    
1595      2. X86
1596    
1597          Added the following instructions to the instruction set:
1598    
1599            ROLx, RORx,
1600            BTx, BTSx, BTLx, BTRx,
1601            XCHGx, and variants with the LOCK prefix
1602    
1603      3. Register Allocation
1604    
1605          The module ra-rewrite-with-renaming has been improved.
1606    
1607      These have no effect on SML/NJ.
1608    
1609    ----------------------------------------------------------------------
1610    Name: Matthias Blume
1611    Date: 2000/05/15 16:20:00 JST
1612    Tag: blume-20000515-lightrebuild
1613    Description:
1614    
1615    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1616       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1617       defined.  In effect, the command will not build any cross-compiler
1618       backends and therefore finish more quickly.
1619    
1620       The "fixpt" script also takes a "-light" switch to be able to use
1621       this new facility while compiling for a fixpoint.
1622    
1623    2. I replaced all mentions of anchored paths in group owner specifications
1624       with simple relative paths (usually starting with "..").
1625       The rationale is that a library's internal workings should not be
1626       compromised by the lack of some anchor.  (An anchor is necessary
1627       for someone who wants to refer to the library by an anchored path,
1628       but it should not be necessary to build the same library in the first
1629       place.)
1630    
1631    3. I changed the way CM's tool mechanism determines the shell command
1632       string used for things like ml-yacc etc. so that it does not break
1633       when CM.Control.implicit_anchors is turned off.
1634    
1635    ----------------------------------------------------------------------
1636    Name: Matthias Blume
1637    Date: 2000/05/12 18:20:00 JST
1638    Tag: blume-20000512-ml-build
1639    Description:
1640    
1641    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1642    getting installed properly (by config/install.sh).
1643    
1644    ----------------------------------------------------------------------
1645    Name: Matthias Blume
1646    Date: 2000/05/12 17:30:00 JST
1647    Tag: blume-20000512-anchors
1648    Description:
1649    
1650    !!! NEW BOOT FILES !!!
1651    
1652    This change is in preparation of fading out support for "implicitly
1653    anchored path names".  I went through all sources and used the
1654    explicit (and relatively new) $-notation.  See system/README and the
1655    CM manual for more info on this.
1656    
1657    I also modified the anchoring scheme for some things such as "smlnj",
1658    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1659    anchors are more expressive: anchor name and first arc do not have to
1660    coincide.  This entails the following user-visible change:
1661    
1662    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1663    particular, when you fire up sml with a command-line argument, say,
1664    e.g.:
1665    
1666       sml '$smlnj/cmb.cm'
1667    
1668    At the ML toplevel prompt:
1669    
1670       CM.autoload "$smlnj/cmb.cm";
1671    
1672    There is also a new controller in CM.Control that can be used to turn
1673    off all remaining support for implicit anchors by saying:
1674    
1675        CM.autoload "$smlnj/
1676        #set CM.Control.implicit_anchors false;
1677    
1678    This causes CM to reject implicitly anchored paths.  This is (for the
1679    time being) less permissive than the "final" version where there will
1680    be no more such implicit anchors and relative paths will be just that:
1681    relative.
1682    
1683    The next step (version after next version?) will be to make the
1684    default for CM.Control.implicit_anchors false.  After the dust has
1685    settled, I can then produce the "final" version of this...
1686    
1687    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1688    
1689    ----------------------------------------------------------------------
1690    Name: Matthias Blume
1691    Date: 2000/05/11 16:30:00 JST
1692    Tag: blume-20000511-sources
1693    Description:
1694    
1695    The main change is that I added function CM.sources as a generalized
1696    version of the earlier CM.makedepend.  This entails the following
1697    additional changes:
1698    
1699      - CM.makedepend has been dropped.
1700    
1701      - CM manual has been updated.
1702    
1703      - TOOLS signature and API have been changed.
1704    
1705    ----------------------------------------------------------------------
1706    Name: Allen Leung
1707    Date: 2000/05/10 21:17 EDT
1708    Tag: leunga-20000510-moby-c--ssa
1709    Description:
1710    
1711      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1712    None of these affect SML/NJ.
1713    
1714    1. Register Allocation
1715    
1716        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1717           This module tries to remove local (i.e. basic block level) redundancies
1718           during spilling.
1719    
1720        b. A new framework for performing region based register allocation.
1721           Not yet entirely functional.
1722    
1723    2. X86
1724    
1725       a. DefUse for POP was missing the stack pointer [found by Lal]
1726       b. Reload for CALL was incorrect in X86Spill [found by John]
1727       c. Various fixes in X86Spill so that it can be used correctly for
1728          the new spilling module.
1729    
1730    3. SSA/IR
1731    
1732       a. New module ir/dj-dataflow.sml implements elimination based
1733          data flow analysis.
1734    
1735    4. MLRiscGen
1736    
1737       a. Fix for gc type annotation
1738    
1739    5. MDGen
1740    
1741       Various fixes for machine description -> ml code translation.  For ssa
1742       only.
1743    
1744    ----------------------------------------------------------------------
1745    Name: Allen Leung
1746    Date: 2000/05/08 22:17 EDT
1747    Tag: leunga-20000508-labexp
1748    Description:
1749    
1750      Fermin has found a few assembly problems with constant expressions
1751      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1752      which choke on dumb assemblers.  This is his fix.
1753    
1754    ----------------------------------------------------------------------
1755    Name: Dave MacQueen
1756    Date: 2000/04/09 14:00 EDT
1757    Tag: dbm-20000502-Version_110_28
1758    Description:
1759    
1760    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1761    
1762    2. Updated config/version to 110.28
1763    
1764    3. Updated config/srcarchiveurl
1765    
1766    3. New boot files!
1767       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1768    
1769    ----------------------------------------------------------------------
1770    Name: Matthias Blume
1771    Date: 2000/05/01 19:05:00 JST
1772    Tag: blume-20000501-noweb
1773    Description:
1774    
1775    A new noweb tool has been added.  The existing system is entirely
1776    unaffected by this, but some CM users have asked for renewed noweb
1777    support.  Everything is documented in the CM manual.
1778    
1779    New (plugin) libraries:
1780    
1781       noweb-tool.cm
1782       nw-ext.cm
1783    
1784    ----------------------------------------------------------------------
1785    Name: Dave MacQueen
1786    Date: 2000/04/30 12:40PM EDT
1787    Tag: dbm-20000430-bug_fixes
1788    Description:
1789    
1790    1. Fix for bug 1498
1791       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1792       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1793         added toRealArray function
1794       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1795         added check for tag Obj.RealArray to array printing case in ppObj
1796    
1797    2. Fix for bug 1510
1798       smlnj/src/compiler/Semant/types/typesutil.sml
1799         fixed definition of dummyargs (used by equalTycon) so that
1800         dummy args are distinct types
1801    
1802    ----------------------------------------------------------------------
1803    Name: Matthias Blume
1804    Date: 2000/04/30 01:00:00 JST
1805    Tag: blume-20000430-versions
1806    Description:
1807    
1808    1. CM version numbering added.  This is an implementation of Lal's
1809       proposal for adding version numbers and version checking to .cm
1810       files.  Lal said that his proposal was just that -- a proposal.
1811       For the time being I went ahead and implemented it so that people
1812       can comment on it.  Everything is completely backward-compatible
1813       (except for the stable library format, i.e., new bootfiles!).
1814    
1815       As usual, see the CM manual for details.
1816    
1817    2. An alternative syntax for anchored paths has been implemented.
1818       Dave has recently voiced the same concerns that I had when I did
1819       this, so there should be some support.  My take is that eventually
1820       I will let support for the current syntax (where anchors are
1821       "implicit") fade out in favor of the new, explicit syntax.
1822       In order to be backward-compatible, both old and new syntax are
1823       currently supported.
1824    
1825       Again, see the CM manual for details.
1826    
1827    3. Parallel make is trying to be slightly smarter:  When the master
1828       process finds a "bottleneck", i.e., when there is only one
1829       compilation unit that can be compiled and everybody else is
1830       waiting on it, then it will simply compile it directly instead
1831       of clumsily telling one of the slaves to do it.
1832    
1833    4. Support for "unsharing" added.  This is necessary in order to be
1834       able to have two different versions of the same library running
1835       at the same time (e.g., for trying out a new MLRISC while still
1836       having the old MLRISC linked into the current compiler, etc.)
1837       See the CM manual.
1838    
1839    5. Simple "makedepend" functionality added for generating Makefile
1840       dependency information.  (This is rather crude at the moment.
1841       Expect some changes here in the future.)
1842    
1843    6. ".fun" added as a recognized suffix for ML files. Also documented
1844       explicitly in the manual that the fallback behavior (unknown suffix
1845       -> ML file) is not an official feature!
1846    
1847    7. Small changes to the pickler for stable libraries.
1848    
1849    8. Several internal changes to CM (for cleanup/improvement).
1850    
1851    
1852    !!!! NEW BINFILES !!!!
1853    
1854    ----------------------------------------------------------------------
1855    Name: Matthias Blume
1856    Date: 2000/04/28 17:30:00 JST
1857    Tag: blume-20000428-pathconfig
1858    Description:
1859    
1860    1. I changed config/install.sh to remove duplicate entries from the
1861       lib/pathconfig file at the end.  Moreover, the final version of
1862       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1863       in src/system/installml.
1864    
1865    2. The config/install.sh script now consistently uses relative
1866       pathnames in lib/pathconfig whenever the anchor is in the lib
1867       directory.  (So far this was true for the libraries that come
1868       pre-compiled and bundled as part of the bootfiles but not for
1869       libraries that are compiled by the script itself.)
1870    
1871    ----------------------------------------------------------------------
1872    Name: Matthias Blume
1873    Date: 2000/04/26 13:10:00 JST
1874    Tag: blume-20000426-fun_suffix
1875    Description:
1876    
1877    Added ".fun" as a recognized file name suffix (for ML code).
1878    
1879    ----------------------------------------------------------------------
1880    Name: Allen Leung
1881    Date: 2000/04/25 17:00:00 EST
1882    Tag: leunga-20000425-alpha-ra
1883    Description:
1884    
1885    1. Alpha
1886    
1887        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1888    in C--.
1889    
1890    2. RA
1891    
1892       Added a flag "ra-dump-size" to print out the size of the flowgraph
1893       and the interference graph.
1894    
1895    ----------------------------------------------------------------------
1896    Name: Dave MacQueen
1897    Date: 2000/04/25/
1898    Tag: dbm-20000425-mlyacc_doc_examples
1899    Description:
1900      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1901      Updated all three examples in src/ml-yacc/examples to run
1902      under 110.* using CM.make.
1903    
1904    ----------------------------------------------------------------------
1905    Name: Allen Leung
1906    Date: 2000/04/20 23:04:00 EST
1907    Tag: leunga-20000420-ssa-c---stuff
1908    Description:
1909    
1910      This update synchronizes my repository with Yale's.  Most of these
1911    changes, however, do not affect SML/NJ at all (the RA is an exception).
1912    
1913    1. Register Allocator
1914    
1915       a. An improvement in the interference graph construction:
1916          Given a copy
1917    
1918                s <- t
1919    
1920          no interference edge between s and t is added for this definition of s.
1921    
1922       b. I've added two new spill heuristic modules that Fermin and I developed
1923          (in the new library RA.cm). These are unused in SML/NJ but maybe
1924          useful for others (Moby?)
1925    
1926    2. X86
1927    
1928       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1929    
1930    3. Alpha
1931    
1932       a. Added the BSR instruction and code generation that goes with it [C--]
1933       b. Other fixes too numerous to recount provided by Fermin [C--]
1934    
1935    4. Regmaps
1936    
1937       a. The regmaps are not initialized with the identity physical bindings
1938          at creation time.  This is unneeded.
1939    
1940    5. MLRISC Optimizations
1941    
1942       a. The DJ-Graph module can now compute the iterated dominance frontiers
1943          intersects with liveness incrementally in linear time! Woohoo!
1944          This is now used in my new SSA construction algorithm.
1945    
1946       b. THe branch reorganization module is now smarter about linear chains of
1947          basic blocks.
1948    
1949    
1950    ----------------------------------------------------------------------
1951    Name: Matthias Blume
1952    Date: 2000/04/12 13:52:00 JST
1953    Tag: blume_main_v110p27_1
1954    Description:
1955    
1956    Changed install.sh script to handle archive files without version number
1957    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1958    name of the boot file archive.
1959    
1960  ----------------------------------------------------------------------  ----------------------------------------------------------------------
1961  Name: Dave MacQueen  Name: Dave MacQueen
1962  Date: 2000/04/09 14:00 EDT  Date: 2000/04/09 14:00 EDT
# Line 119  Line 2065 
2065    
2066  3.  Assembly  3.  Assembly
2067    
2068      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
2069      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
2070      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
2071    
# Line 142  Line 2088 
2088    
2089  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
2090  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
2091  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
2092  legal surface-syntax symbol).  legal surface-syntax symbol).
2093    
2094  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 152  Line 2098 
2098  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
2099    
2100  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2101  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
2102  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
2103  more details.  more details.
2104    
# Line 267  Line 2213 
2213     (specified in the .cm file at each instance where the tool's class is     (specified in the .cm file at each instance where the tool's class is
2214     used).     used).
2215    
2216     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
2217     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
2218     Makefiles or Shell scripts.     Makefiles or Shell scripts.
2219    
2220     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 280  Line 2226 
2226    
2227     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
2228    
2229     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
2230     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
2231    
2232  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 414  Line 2360 
2360    
2361       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
2362       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
2363       testl is perferable to andl.       testl is preferable to andl.
2364    
2365  7. RA (x86 only)  7. RA (x86 only)
2366    
# Line 596  Line 2542 
2542    
2543  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2544  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2545  specifiel the layout of the tool command line.  See the CM manual for  specifies the layout of the tool command line.  See the CM manual for
2546  explanation.  explanation.
2547    
2548  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2549  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2550  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2551  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 643  Line 2589 
2589  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
2590  <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,
2591  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
2592  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2593  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2594  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2595    
# Line 693  Line 2639 
2639    it from that remote directory.    it from that remote directory.
2640    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2641    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2642    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2643    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2644    
2645    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.607  
changed lines
  Added in v.832

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