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 591, Mon Apr 3 01:19:20 2000 UTC revision 824, Tue May 8 21:31:05 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: Lal George
17    Date: 2001/05/08  17:26:09 EDT
18    Tag: george-20010508-omit-frameptr
19    Description:
20    
21    Changes to implement the omit-frame-pointer optimization to support
22    raw C calls. For now, there is only support on the Intel x86, but
23    other architectures will follow as more experience is gained with this.
24    
25    
26    ----------------------------------------------------------------------
27    Name: Matthias Blume
28    Date: 2001/05/07 14:40:00 EDT
29    Tag: blume-20010507-proxies
30    Description:
31    
32    I made into "proxy libraries" all libraries that qualify for such a
33    change.  (A qualifying library is a library that has another library or
34    groups as its sole member and repeats that member's export list
35    verbatim.  A proxy library avoids this repetition by omitting its export
36    list, effectively inheriting the list that its (only) member exports.
37    See the CM manual for more explanation.)
38    The main effect is that explicit export lists for these libraries
39    do not have to be kepts in sync, making maintenance a bit easier.
40    
41    I also added copyright notices to many .cm-files.
42    
43    Last but not least, I made a new set of bootfiles.
44    
45    ----------------------------------------------------------------------
46    Name: Matthias Blume
47    Date: 2001/05/04 17:00:00 EDT
48    Tag: blume-20010504-cm-lsplit
49    Description:
50    
51    0. John merged pending changes to $/smlnj-lib.cm
52    
53    1. Allen's previous change accidentally backed out of one of Lal's
54       earlier changes.  I undid this mistake (re-introducing Lal's change).
55    
56    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
57       within the compiler where applicable.  There is some code simplification
58       because of that.
59    
60    3. The "split" phase (in FLINT) is now part of the default list of phases.
61       Compiler.Control.LambdaSplitting.* can be used to globally control the
62       lambda-splitting (cross-module-inlining) engine.  In addition to that,
63       it can now also be controlled on a per-source basis: CM has been taught
64       a new tool parameter applicable to ML source files.
65    
66       - To turn lambda-splitting off completely:
67            local open Compiler.Control.LambdaSplitting in
68                val _ = set Off
69            end
70       - To make "no lambda-splitting" the global default (but allow per-source
71         overriding); this is the initial setting:
72            local open Compiler.Control.LambdaSplitting in
73                val _ = set (Default NONE)
74            end
75       - To make "lambda-splitting with aggressiveness a" the global default
76         (and allow per-source overriding):
77            local open Compiler.Control.LambdaSplitting in
78                val _ = set (Default (SOME a))
79            end
80    
81       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
82         write (in the respective .cm-file):
83            a.sml (lambdasplitting:off)
84       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
85            a.sml (lambdasplitting:on)
86       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
87         <a> is a decimal non-negative integer):
88            a.sml (lambdasplitting:<a>)
89       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
90            a.sml (lambdasplitting:infinity)
91       - To use the global default for a.sml:
92            a.sml (lambdasplitting:default)
93         or simply
94            a.sml
95    
96    ----------------------------------------------------------------------
97    Name: Allen Leung
98    Date: 2001/05/04 01:57:00 EDT
99    Tag: leunga-20010504-sync
100    Description:
101    
102      MLRISC features.
103    
104      1. Fix to CMPXCHG instructions.
105      2. Changed RA interface to allow annotations in callbacks.
106      3. Added a new method to the stream interface to allow annotations updates.
107    
108    ----------------------------------------------------------------------
109    Name: Matthias Blume
110    Date: 2001/05/01 11:45:00 EDT
111    Tag: blume-20010501-pcedittmp
112    Description:
113    
114    Changed install.sh to use the current working directory instead of
115    /usr/tmp for a temporary file (pcedittmp).  The previous choice
116    of /usr/tmp caused trouble with MacOS X because of file premission
117    problems.
118    
119    ----------------------------------------------------------------------
120    Name: Matthias Blume
121    Date: 2001/04/20 11:10:00 EDT
122    Tag: blume-20010420-inMLflag
123    Description:
124    
125     - added vp_limitPtrMask to vproc-state.h
126       (for use by the raw-C-calls mechanism to implement proper interrupt
127        handling)
128     - made the ML compiler aware of various data-structure offsets so it
129       can generate code for accessing the vp_inML flag and vp_limitPtrMask
130     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
131       raw C-calls
132    
133    ----------------------------------------------------------------------
134    Name: Lal George
135    Date: 2001/04/20 09:15:28 EDT
136    Tag: george-20010420-macosX
137    Description:
138    
139     - Changes to port to Mac OS X; Darwin.
140    
141     - In the process I found that sqrt was broken on the PPC, because the
142       fsqrt instruction is not implemented.
143    
144    ----------------------------------------------------------------------
145    Name: Matthias Blume
146    Date: 2001/04/18 12:45:00 EDT
147    Tag: blume-20010418-ccalls
148    Description:
149    
150     - fixed two off-by-4 errors in the x86-specific c-calls implementation
151       (this bug prevented structure arguments containing pointers from being
152        passed correctly)
153     - changed the raw-C-call code in mlriscGen.sml in such a way that
154       structure arguments are represented as a pointer to the beginning
155       of the structure (instead of having a series of synthesized arguments,
156       one for each structure member)
157    
158     - made makeml script's verbosity level configurable via environment
159       variable (MAKEML_VERBOSITY)
160    
161     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
162       in rawmem-x86.sml; we are now using the real thing
163    
164    ----------------------------------------------------------------------
165    Name: Matthias Blume
166    Date: 2001/03/22 16:25:00 EST
167    Tag: blume-20010322-bootfiles
168    Description:
169    
170    Created a new set of bootfiles (for your automatic installation convenience).
171    
172    ----------------------------------------------------------------------
173    Name: Matthias Blume
174    Date: 2001/03/22 15:10:00 EST
175    Tag: blume-20010322-rawmem-parcm
176    Description:
177    
178    1. All "raw memory access" primitives for the new FFI are implemented now
179       (at least on the x86).
180    2. Some further cleanup of CM's parallel make mechanism.
181    
182    ----------------------------------------------------------------------
183    Name: Matthias Blume
184    Date: 2001/03/19 17:53:00 EST
185    Tag: blume-20010319-parallel
186    Description:
187    
188    Parallel make (using compile servers) now works again.
189    
190    To this end, CM.stabilize and CMB.make have been modified to work in
191    two passes when compile servers are attached:
192       1. Compile everything, do not perform stabilization; this pass
193          uses compile servers
194       2. Stabilize everything; this pass does not use compile servers
195    If there are no compile servers, the two passes are combined into one
196    (as before).  Splitting the passes increases the inherent parallelism
197    in the dependency graph because the entire graph including all
198    libraries is available at the same time.  This, in turn, improves
199    server utilization.  The downside is that the master process will
200    have to do some extra work after compilation is done (because for
201    technical reasons it must re-read all the binfiles during stabilization).
202    
203    ----------------------------------------------------------------------
204    Name: Matthias Blume
205    Date: 2001/03/16 12:22:00 EST
206    Tag: blume-20010316-bootfiles
207    Description:
208    
209    Created a new set of bootfiles (for your automatic installation convenience).
210    
211    ----------------------------------------------------------------------
212    Name: Matthias Blume
213    Date: 2001/03/16 11:00:00 EST
214    Tag: blume-20010316-MLTREE-fixup
215    Description:
216    
217    This is a minor fixup for an (untagged) earlier commit by Allen.
218    (A file was missing).
219    
220    ----------------------------------------------------------------------
221    Name: Allen Leung
222    Date: Mon Mar  5 18:54:57 EST 2001
223    Tag: leunga-20010305-cut-support
224    
225    1. New support for alternative control-flow in MLTREE.
226       Currently we support
227    
228          FLOW_TO(CALL ...., [k1,...,kn])
229    
230       This is needed for 'cuts to' in C-- and try/handle-like constructs
231       in Moby
232    
233       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
234    
235    2. Register Allocator
236    
237       Changes in interface [from Fermin, John]
238    
239    3. Alpha 8-bit SLL support [Fermin]
240    
241    4. All architectures
242    
243       A new module (ClusterExpandCopies) for expanding parallel copies.
244    
245    ----------------------------------------------------------------------
246    Name: Allen Leung
247    Date: 2001/02/27 23:07:00 EST
248    Tag: leunga-20010227-minor-stuff
249    
250    1. Alpha bug fix for CMOVNE
251    2. Handle mltree COND(..,FCMP ...,...)
252    3. Bug fix in simplifier
253    
254    ----------------------------------------------------------------------
255    Name: Matthias Blume
256    Date: 2001/01/30 17:50:00 EST
257    Tag: blume-20010130-sync
258    Description:
259    
260    This is just a minor update to sync my devel branch with the main brach.
261    The only visible change is the addition of some README files.
262    
263    ----------------------------------------------------------------------
264    Name: Matthias Blume
265    Date: 2001/01/12 23:30:00 JST
266    Tag: blume-20010112-bootfiles
267    Description:
268    
269    Made a new set of bootfiles that goes with the current state of the
270    repository.
271    
272    ----------------------------------------------------------------------
273    Name: Matthias Blume
274    Date: 2001/01/12 21:20:00 JST
275    Tag: blume-20010112-sync
276    Description:
277    
278    I am just flushing out some minor changes that had accumulated in
279    my private branch in order to sync with the main tree.  (This is
280    mainly because I had CVS trouble when trying to merge _into_ my
281    private branch.)
282    
283    Most people should be completely unaffected by this.
284    
285    ----------------------------------------------------------------------
286    Name: Allen Leung
287    Date: Thu Jan 11 21:03:00 EST 2001
288    Tag: leunga-20010111-labexp=mltree
289    Description:
290    
291    1.  Removed the type LabelExp and replace it by MLTree.
292    2.  Rewritten mltree-simplify with the pattern matcher tool.
293    3.  There were some bugs in alpha code generator which would break
294        64-bit code generation.
295    4.  Redo the tools to generate code with the
296    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
297        are now generated by perl scripts.
298    
299    ----------------------------------------------------------------------
300    Name: Matthias Blume
301    Date: 2001/01/10 21:55:00 JST
302    Tag: blume-20010110-rcc
303    Description:
304    
305    The RCC stuff now seems to work (but only on the x86).
306    This required hacking of the c-calls interface (and -implementation) in
307    MLRISC.
308    
309    Normal compiler users should be unaffected.
310    
311    ----------------------------------------------------------------------
312    Name: Matthias Blume
313    Date: 2001/01/09 01:20:00 JST
314    Tag: blume-20010109-rcc
315    Description:
316    
317    This is a fairly big patch, flushing out a large number of pending
318    changes that I made to my development copy over the last couple of days.
319    
320    Of practical relevance at this moment is a workaround for a pickling
321    bug that Allen ran into the other day.  The cause of the bug itself is
322    still unknown and it might be hard to fix it properly, but the
323    workaround has some merits of its own (namely somewhat reducing pickling
324    overhead for certain libraries).  Therefore, I think this solution should
325    be satisfactory at this time.
326    
327    The rest of the changes (i.e., the vast majority) has to do with my
328    ongoing efforts of providing direct support for C function calls from
329    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
330    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
331    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
332    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
333    handle RCC.
334    
335    The last part (the changes to mlriscGen.sml) are still known to be
336    wrong on the x86 and not implemented on all other architectures.  But
337    the infrastructure is in place. I had to change a few functor
338    signatures in the backend to be able to route the CCalls interface
339    from MLRISC there, and I had to specialize the mltree type (on the
340    x86) to include the necessary extensions. (The extensions themselves
341    were already there and redy to go in MLRISC/x86).
342    
343    Everything should be very happy as soon as someone helps me with
344    mlriscGen.sml...
345    
346    In any case, nothing of this should matter to anyone as long as the
347    new primop is not being used (which is going to be the case unless you
348    find it where I hid it :). The rest of the compiler is completely
349    unaffected.
350    
351    ----------------------------------------------------------------------
352    Name: Matthias Blume
353    Date: 2001/01/05 00:30:00 JST
354    Tag: blume-20010105-primops
355    Description:
356    
357    Added some experimental support for work that I am doing right now.
358    These changes mostly concern added primops, but there is also a new
359    experimental C library in the runtime system (but currently not enabled
360    anywhere except on Linux/X86).
361    
362    In the course of adding primops (and playing with them), I discovered that
363    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
364    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
365    To recover, I made type information in INL_PRIM mandatory and changed
366    prim.sml as well as built-in.sml accordingly.  The InLine structure now
367    has complete, correct type information (i.e., no bottom types).
368    
369    Since all these changes mean that we need new binfiles, I also bumped the
370    version number to 110.32.1.
371    
372    ----------------------------------------------------------------------
373    Name: Matthias Blume
374    Date: 2000/12/30 22:10:00 JST
375    Tag: blume-20001230-various
376    Description:
377    
378    Added proxy libraries for MLRISC and let MLRISC libraries refer
379    to each other using path anchors.  (See CM manual for explanation.)
380    
381    Updated CM documentation.
382    
383    Fixed some bugs in CM.
384    
385    Implemented "proxy" libraries (= syntactic sugar for CM).
386    
387    Added "-quiet" option to makeml and changed runtime system accordingly.
388    
389    Added cleanup handler for exportML to reset timers and compiler stats.
390    
391    ----------------------------------------------------------------------
392    Name: Lal George
393    Date: 2000/12/22 22:22:58 EST 2000
394    Tag: Release_110_32
395    Description:
396    
397            Infinite precision used throughout MLRISC.
398            see MLRISC/mltree/machine-int.sig
399    
400    ----------------------------------------------------------------------
401    Name: Matthias Blume
402    Date: 2000/12/22 23:16:00 JST
403    Tag: blume-20001222-warn
404    Description:
405    
406    Corrected wording and formatting of some CM warning message which I
407    broke in my previous patch.
408    
409    ----------------------------------------------------------------------
410    Name: Matthias Blume
411    Date: 2000/12/22 21:20:00 JST
412    Tag: blume-20001222-anchorenv
413    Description:
414    
415    Fixed CM's handling of anchor environments in connection with CMB.make.
416    
417    ----------------------------------------------------------------------
418    Name: Matthias Blume
419    Date: 2000/12/22 13:15:00 JST
420    Tag: blume-20001222-cleanup
421    Description:
422    
423    Removed src/cm/ffi which does not (and did not) belong here.
424    
425    ----------------------------------------------------------------------
426    Name: Matthias Blume
427    Date: 2000/12/21 23:55:00 JST
428    Tag: blume-20001221-exn
429    Description:
430    
431    Probably most important: CM no longer silently swallows all exceptions
432    in the compiler.
433    Plus: some other minor CM changes.  For example, CM now reports some
434    sizes for generated binfiles (code, data, envpickle, lambdapickle).
435    
436    ----------------------------------------------------------------------
437    Name: Matthias Blume
438    Date: 2000/12/15 00:01:05 JST
439    Tag: blume-20001215-dirtool
440    Description:
441    
442    - "dir" tool added.
443    - improvements and cleanup to Tools structure
444    - documentation updates
445    
446    ----------------------------------------------------------------------
447    Name: Allen Leung
448    Date: Thu Dec 14 03:45:24 EST 2000
449    Description:
450    Tag:  leunga-20001214-int-inf
451    Description:
452    
453       In IntInf, added these standard functions, which are missing from our
454    implementation:
455    
456        andb : int * int -> int
457        xorb : int * int -> int
458        orb  : int * int -> int
459        notb : int -> int
460         <<   : int * word -> int
461        ~>>  : int * word -> int
462    
463       Not tested, I hope they are correct.
464    
465    ----------------------------------------------------------------------
466    Name: Allen Leung
467    Date: Fri Dec  8 19:23:26 EST 2000
468    Description:
469    Tag:  leunga-20001208-nowhere
470    Description:
471    
472      Slight improvements to the 'nowhere' tool to handle OR-patterns,
473    to generate better error messages etc.  Plus a brief manual.
474    
475    ----------------------------------------------------------------------
476    Name: Lal George
477    Date: 2000/12/08 09:54:02 EST 2000
478    Tag: Release_110_31
479    Description:
480    
481    - Version 110.31
482    ----------------------------------------------------------------------
483    Name: Allen Leung
484    Date: Thu Dec  7 22:01:04 EST 2000
485    Tag:  leunga-20001207-cell-monster-hack
486    Description:
487    
488    Major MLRISC internal changes.  Affect all clients.
489    Summary:
490    
491    1.  Type CELLS.cell = int is now replaced by a datatype.
492        As a result, the old regmap is now gone.  Almost all interfaces
493        in MLRISC change as a consequence.
494    
495    2.  A new brand version of machine description tool (v3.0) that generates
496        modules expecting the new interface.  The old version is removed.
497    
498    3.  The RA interface has been further abstracted into two new functors.
499        RISC_RA and X86RA.  These functors have much simpler interfaces.
500        [See also directory MLRISC/demo.]
501    
502    4.  Some other new source->source code generation tools are available:
503    
504        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
505        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
506           I use this tool to generate the peephole optimizers---with the new
507           cell type changes, peephole rules are becoming difficult to write
508           without conditional pattern matching.
509    
510    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
511        cover the entire MLRISC source tree so many things broke.
512    
513    6.  CM files have been moved to the subdirectory MLRISC/cm.
514        They are moved because there are a lot of them and they clutter up the
515        root dir.
516    
517    7.  More detailed documentation to come...
518    
519        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
520        a makeml -rebuild first.  This is because of other other
521        changes that Matthias has made (see below).
522    
523    
524    ----------------------------------------------------------------------
525    Name: Matthias Blume
526    Date: 2000/11/30 23:12:00 JST
527    Tag: blume-20001130-filereorg
528    Description:
529    
530    Some manual updates and some file reorganizations in CM.
531    
532    ----------------------------------------------------------------------
533    Name: Matthias Blume
534    Date: 2000/11/24 17:45:00 JST
535    Tag: blume-20001124-link
536    Description:
537    
538    Drastically improved link traversal code for the case that the dynamic
539    value was already loaded at bootstrap time.  As a result, CM and CMB
540    now both load blazingly fast -- even on a very slow machine.  Also,
541    memory consumption has been further reduced by this.
542    
543    Warning: The format of the PIDMAP file has changed.  THerefore, to
544    bootstrap you have to do this:
545    
546    1. Run CMB.make
547    2. Make a symbolic link for the boot directory:
548         ln -s sml.boot.ARCH-OS xxx
549    3. "Rebuild" the boot directory:
550         ./makeml -boot xxx -rebuild sml ; rm xxx
551    4. Boot normally:
552          ./makeml
553    
554    ----------------------------------------------------------------------
555    Name: Matthias Blume
556    Date: 2000/11/21 21:20:00 JST
557    Tag: blume-20001121-tools
558    Description:
559    
560    Continued hacking on autoloading problem -- with success this time.
561    Also changed tool-plugin mechanism.  See new CM manual.
562    
563    ----------------------------------------------------------------------
564    Name: Matthias Blume
565    Date: 2000/11/19 14:30:00 JST
566    Tag:  blume-20001119-autoload
567    Description:
568    
569    Some hacking to make autoloading faster.  Success for CMB, no success
570    so far for CM.  There is a reduced structure CM' that autoloads faster.
571    (This is a temporary, non-documented hack to be eliminated again when
572    the general problem is solved.)
573    
574    ----------------------------------------------------------------------
575    Name: Matthias Blume
576    Date: 2000/11/17 14:10:00 JST
577    Tag: blume-20001117-pickle-lib
578    Description:
579    
580    1. Eliminated comp-lib.cm
581    2. Made pickle-lib.cm
582    3. Eliminated all uses of intset.sml (from comp-lib.cm)
583    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
584       equivalent constructs from smlnj-lib.cm (INtHashTable).
585    5. Point 4. also goes for those uses of intmap.* in MLRISC.
586       Duplicated intmap modules thrown out.
587    6. Hunted down all duplicated SCC code and replaced it with
588       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
589    7. Rewrote Feedback module.
590    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
591       should be thrown out and equivalent modules from smlnj-lib.cm
592       should be used (IntRedBlackSet, IntListSet, ...).
593    
594    Confirmed that compiler compiles to fixpoint.
595    
596    ----------------------------------------------------------------------
597    Name: Allen Leung
598    Date: 2000/11/10 18:00:00
599    Tag: leunga-20001110-new-x86-fp
600    
601    A new x86 floating point code generator has been added.
602    By default this is turned off.  To turn this on, do:
603    
604        CM.autoload "$smlnj/compiler.cm";
605        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
606    
607    Changes:
608    
609    1.  Changed FTAN to FPTAN so that the assembly output is correct.
610    2.  Changed the extension callback for FTANGENT to generate:
611    
612              fptan
613              fstp  %st(0)
614        instead of
615              fptan
616              fstpl ftempmem
617    
618    3.  Numerous assembly fixes for x86.
619    
620    5.  Cleaned up the machine code output module x86/x86MC.sml and added
621        support for a whole bunch of instructions and addressing modes:
622    
623          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
624          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
625          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
626          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
627          fxch %st(n)
628          fld %st(n)
629          fst %st(n)
630          fst mem
631          fstp %st(n)
632          fucom %st(n)
633          fucomp %st(n)
634    
635        All these are now generated when the fast fp mode is turned on.
636    
637    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
638    
639    ----------------------------------------------------------------------
640    Name: Matthias Blume
641    Date: 2000/11/09 11:20:00 JST
642    Tag: blume-20001109-scc
643    Description:
644    
645    Eliminated some code duplication:
646    
647    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
648       (Otherwise the functor is useless.)
649    2. Used GraphSCCFn where SCCUtilFun was used previously.
650    3. Got rid of SCCUtilFun (in comp-lib.cm).
651    
652    ----------------------------------------------------------------------
653    Name: Lal George
654    Date: 2000/11/06 09:02:21 EST 2000
655    Tag: Release_110_30
656    Description:
657    
658    - Version 110.30
659    ----------------------------------------------------------------------
660    Name: Matthias Blume
661    Date: 2000/11/04 14:45:00
662    Tag: blume-20001104-mlbuild
663    Description:
664    
665    - Made ml-build faster on startup.
666    - Documentation fixes.
667    
668    ----------------------------------------------------------------------
669    Name: Matthias Blume
670    Date: 2000/11/02 17:00:00 JST
671    Tag: blume-20001102-condcomp
672    Description:
673    
674    - Small tweaks to pickler -- new BOOTFILES!
675    - Version bumped to 110.29.2.
676    - Added conditional compilation facility to init.cmi (see comment there).
677    ----------------------------------------------------------------------
678    Name: Allen Leung
679    Date: 2000/10/23 19:31:00
680    Tag: leunga-20001023-demo-ra
681    
682    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
683    2. Test programs for the graph library updated
684    3. Some new MLRISC demo programs added
685    
686    ----------------------------------------------------------------------
687    Name: Matthias Blume
688    Date: 2000/08/31 22:15:00 JST
689    Tag: blume-20001017-errmsg
690    Description:
691    
692    More error message grief: Where there used to be no messages, there
693    now were some that had bogus error regions.  Fixed.
694    
695    ----------------------------------------------------------------------
696    Name: Matthias Blume
697    Date: 2000/08/31 17:30:00 JST
698    Tag: blume-20001017-v110p29p1
699    Description:
700    
701    I made a version 110.29.1 with new bootfiles.
702    
703    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
704              CM documentation changes and a small bugfix in CM's error reporting.
705    
706    ----------------------------------------------------------------------
707    Name: Lal George
708    Date: 2000/09/27 14:42:35 EDT
709    Tag: george-20000927-nodestatus
710    Description:
711    
712    Changed the type of the nodestatus, so that:
713    
714            SPILLED(~1)             is now SPILLED
715            SPILLED(m) where m>=0   is now MEMREG(m)
716            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
717    
718    ----------------------------------------------------------------------
719    Name: Matthias Blume
720    Date: 2000/09/07 14:45:00 JST
721    Tag: blume-20000907-cmerrmsg
722    Description:
723    
724    Small tweak to CM to avoid getting ML syntax error messages twice.
725    
726    ----------------------------------------------------------------------
727    Name: Matthias Blume
728    Date: 2000/08/31 18:00:00 JST
729    Tag: blume-20000831-cvsbootfiles
730    Description:
731    
732    New URL for boot files (because the 110.29 files on the BL server do
733    now work correctly with my updated install scripts for yacc and lex).
734    
735    ----------------------------------------------------------------------
736    Name: Matthias Blume
737    Date: 2000/08/08 12:33:00 JST
738    Tag: blume-20000808-manual
739    Description:
740    
741    Tiny update to CM manual.
742    
743    ----------------------------------------------------------------------
744    Name: Allen Leung
745    Date: 2000/08/7 19:31:00
746    Tag: leunga-20000807-a-whole-bunch-of-stuff
747    
748      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
749    and Mobdy.
750    
751    1.  x86
752    
753       a.  Fixes to peephole module by John and Dan.
754       b.  Assembly fix to SETcc by Allen.
755       c.  Fix to c-call by John.
756       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
757       e.  Instruction selection optimization to SETcc as suggested by John.
758    
759           For example,
760    
761            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
762    
763           should generate:
764    
765            MOVL a, x
766            SUBL b, x
767            SHRL 31, x
768    
769    2.  IR stuff
770    
771         A bunch of new DJ-graph related algorithms added.  These
772         speed up SSA construction.
773    
774    3.  SSA + Scheduling
775    
776         Added code for SSA and scheduling to the repository
777    
778    ----------------------------------------------------------------------
779    Name: Lal George
780    Date: 2000/07/27 11:53:14 EDT
781    
782    Tag: lal-20000727-linux-ppc
783    Description:
784    
785     Made changes to support Linux PPC.
786     p.s. I have confirmation that the 110.29 boot files work fine.
787    
788    ----------------------------------------------------------------------
789    Name: Matthias Blume
790    Date: 2000/07/27 17:40:00 JST
791    Tag: blume-20000727-scripts
792    Description:
793    
794    !!!! WARNING !!!!
795    You must recompile the runtime system!
796    !!!! WARNING !!!!
797    
798    This is basically another round of script-enhancements:
799    
800    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
801       and undefine CM preprocessor symbols.
802    
803    2. ml-build avoids generating a new heap image if it finds that the
804       existing one is still ok.  (The condition is that no ML file had to
805       be recompiled and all ML files are found to be older that the heap
806       file.)
807    
808       To make this work smoothly, I also hacked the runtime system as
809       well as SMLofNJ.SysInfo to get access to the heap image suffix
810       (.sparc-solaris, ...) that is currently being used.
811    
812       Moreover, the signature of CM.mk_standalone has changed.  See the
813       CM manual.
814    
815    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
816       CM manual for details.)
817    
818    4. More CM manual updates:
819        - all of the above has been documented.
820        - there is now a section describing the (CM-related) command line
821          arguments that are accepted by the "sml" command
822    
823    ----------------------------------------------------------------------
824    Name: Matthias Blume
825    Date: 2000/07/25 16:20:00 JST
826    Tag: blume-20000725-makedepend
827    Description:
828    
829    Added a script called ml-makedepend.  This can be used in makefiles
830    for Unix' make in a way very similar to the "makedepend" command for
831    C.
832    
833    The script internally uses function CM.sources.
834    
835    Synopsis:
836    
837        ml-makedepend [-f makefile] cmfile targetname
838    
839    The default for the makefile is "makefile" (or "Makefile" should
840    "makefile" not exist).
841    
842    ml-makedepend adds a cmfile/targetname-specific section to this
843    makefile (after removing the previous version of this section).  The
844    section contains a single dependency specification with targetname on
845    the LHS (targetname is an arbitrary name), and a list of files derived
846    from the cmfile on the RHS.  Some of the files on the RHS are
847    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
848    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
849    path names.  The makefile writer is responsible for making sure that
850    these variables have correct at the time "make" is invoked.
851    
852    ----------------------------------------------------------------------
853    Name: Matthias Blume
854    Date: 2000/07/22 23:30:00 JST
855    Tag: blume-20000722-urlupdate
856    Description:
857    
858    Changed BOOT and config/srcarchiveurl to point to BL server:
859    
860        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
861    
862    ----------------------------------------------------------------------
863    Name: Matthias Blume
864    Date: 2000/07/18 18:00:00 JST
865    Tag: blume-20000718-Version_110_29
866    Description:
867    
868    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
869    
870    2. Updated config/version to 110.29
871    
872    3. Updated config/srcarchiveurl
873    
874    3. New boot files!
875       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
876    
877    ----------------------------------------------------------------------
878    Name: Matthias Blume
879    Date: 2000/07/11 13:58:00 JST
880    Tag: blume-20000711-doctypo
881    Description:
882    
883    Fixed a few typos in CM manual.
884    
885    ----------------------------------------------------------------------
886    Name: Allen Leung
887    Date: 2000/06/15 00:38:00
888    Tag: leunga-20000704-sparc-x86
889    
890    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
891    2. fix to x86 RET bug [found by Dan Grossman]
892    3. sparc assembly bug fix for ticc instructions [found by Fermin]
893    
894       Affects c-- and moby only
895    
896    ----------------------------------------------------------------------
897    Name: Matthias Blume
898    Date: 2000/07/04 15:26:00
899    Tag: blume-20000704-trigger
900    Description:
901    
902    1. Improvements to CM manual.
903    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
904       of getting a back-trace.  The function, when called, raises an
905       internal exception which explicitly carries the full back-trace history,
906       so it is unaffected by any intervening handle-raise pairs ("trivial"
907       or not).  The interactive loop will print that history once it arrives
908       at top level.
909       Short of having all exceptions implicitly carry the full history, the
910       recommended way of using this facility is:
911         - compile your program with instrumentation "on"
912         - run it, when it raises an exception, look at the history
913         - if the history is "cut off" because of some handler, go and modify
914           your program so that it explicitly calls BTrace.trigger
915         - recompile (still instrumented), and rerun; look at the full history
916    
917    ----------------------------------------------------------------------
918    Name: Matthias Blume
919    Date: 2000/07/03 15:36:00 JST
920    Tag: blume-20000702-manual
921    Description:
922    
923    Small corrections and updates to CM manual.
924    
925    ----------------------------------------------------------------------
926    Name: Matthias Blume
927    Date: 2000/06/29 16:04:00 JST
928    Tag: blume-20000629-yacctool
929    Description:
930    
931    Changes:
932    
933    1. Class "mlyacc" now takes separate arguments to pass options to
934       generated .sml- and .sig-files independently.
935    2. Corresponding CM manual updates.
936    3. BTrace module now also reports call sites.  (However, for loop clusters
937       it only shows from where the cluster was entered.)  There are associated
938       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
939    
940    ----------------------------------------------------------------------
941    Name: Matthias Blume
942    Date: 2000/06/27 16:51:00 JST
943    Tag: blume-20000627-noweb
944    Description:
945    
946    Changes:
947    
948     1. Implemented "subdir" and "witness" options for noweb tool.
949        This caused some slight internal changes in CM's tool implementation.
950     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
951        some remaining issues from earlier path anchor changes.
952     3. Updated CM manual accordingly.
953    
954     4. Changed implementation of back-tracing so that I now consider it
955        ready for prime-time.
956    
957        In particular, you don't have to explicitly trigger the back-trace
958        anymore.  Instead, if you are running BTrace-instrumented code and
959        there is an uncaught exception (regardless of whether or not it was
960        raised in instrumented code), the top-level evalloop will print
961        the back-trace.
962    
963        Features:
964    
965          - Instrumented and uninstrumented code work together seemlessly.
966            (Of course, uninstrumented code is never mentioned in actual
967             back-traces.)
968    
969          - Asymptotic time- and space-complexity of instrumented code is
970            equal to that of uninstrumented code.  (This means that
971            tail-recursion is preserved by the instrumentation phase.)
972    
973          - Modules whose code has been instrumented in different sessions
974            work together without problem.
975    
976          - There is no penalty whatsoever on uninstrumented code.
977    
978          - There is no penalty on "raise" expressions, even in
979            instrumented code.
980    
981        A potential bug (or perhaps it is a feature, too):
982    
983          A back-trace reaches no further than the outermost instrumented
984          non-trivial "raise".  Here, a "trivial" raise is one that is the
985          sole RHS of a "handle" rule.  Thus, back-traces reach trough
986    
987               <exp> handle e => raise e
988    
989          and even
990    
991               <exp> handle Foo => raise Bar
992    
993          and, of course, through
994    
995               <exp> handle Foo => ...
996    
997         if the exception was not Foo.
998    
999         Back-traces always reach right through any un-instrumented code
1000         including any of its "handle" expressions, trivial or not.
1001    
1002       To try this out, do the following:
1003    
1004         - Erase all existing binfiles for your program.
1005           (You may keep binfiles for those modules where you think you
1006            definitely don't need back-tracing.)
1007         - Turn on back-trace instrumentation:
1008              SMLofNJ.Internals.BTrace.mode (SOME true);
1009         - Recompile your program.  (I.e., run "CM.make" or "use".)
1010         - You may now turn instrumentation off again (if you want):
1011              SMLofNJ.Internals.BTrace.mode (SOME false);
1012         - Run your program as usual.  If it raises an exception that
1013           reaches the interactive toplevel, then a back-trace will
1014           automatically be printed.  After that, the toplevel loop
1015           will print the exception history as usual.
1016    
1017    ----------------------------------------------------------------------
1018    Name: Matthias Blume
1019    Date: 2000/06/26 09:56:46 JST
1020    Tag: blume-20000626-setup
1021    Description:
1022    
1023    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1024          ML code before and after compiling a file (e.g., to set compiler
1025          flags)
1026    
1027    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1028              - associated changes to btrace.sml (BTrace instrumentation pass)
1029              - cleaner implementation of btimp.sml (BTrace tracing and report
1030                module)
1031    
1032    CM manual: * new path encoding documented
1033               * description of setup-parameter to "sml" added
1034    
1035    The biggest user-visible change to back-tracing is that it is no
1036    longer necessary to compile all traced modules within the same
1037    session.  (This was a real limitation.)
1038    
1039    ----------------------------------------------------------------------
1040    Name: Matthias Blume
1041    Date: 2000/06/24 12:40:00 JST
1042    Tag: blume-20000624-startup
1043    Description:
1044    
1045    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1046    bit too often -- to put it mildly. :)
1047    
1048    ----------------------------------------------------------------------
1049    Name: Matthias Blume
1050    Date: 2000/06/23 18:20:00 JST
1051    Tag: blume-20000623-btrace
1052    Description:
1053    
1054    This updates adds a backtrace facility to aid programmers in debugging
1055    their programs.  This involves the following changes:
1056    
1057    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1058       keeping track of the current call stack.  When programs are compiled
1059       in a special mode, the compiler will insert calls to these hooks
1060       into the user program.
1061       "Hook" means that it is possible for different implementations of
1062       back-tracing to register themselves (at different times).
1063    
1064    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1065       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1066       It can be turned on using this call:
1067         SMLofNJ.Internals.BTrace.mode (SOME true);
1068       Turning it off again:
1069         SMLofNJ.Internals.BTrace.mode (SOME false);
1070       Querying the current status:
1071         SMLofNJ.Internals.BTrace.mode NONE;
1072       Annotated programs are about twice as big as normal ones, and they
1073       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1074       where all hooks do nothing).  The slowdown with a plugin that is
1075       actually useful (such as the one supplied by default) is even greater,
1076       but in the case of the default plugin it is still only an constant
1077       factor (amortized).
1078    
1079    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1080       with a sub-structure BTrace for controlling back-tracing.  In particular,
1081       the above-mentioned function "mode" controls whether the annotation
1082       phase is invoked by the compiler.  Another important function is
1083       "trigger": when called it aborts the current execution and causes
1084       the top-level loop to print a full back-trace.
1085    
1086    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1087       for back-tracing.  It keeps track of the dynamic call stack and in
1088       addition to that it keeps a partial history at each "level" of that
1089       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1090       d and b (at separate times, dynamically), then the report will show:
1091    
1092       GOTO   d
1093             /c
1094       GOTO  \b
1095       CALL   a
1096    
1097       This shows that there was an initial non-tail call of a, then a
1098       tail-call to b or c, looping behavior in a cluster of functions that
1099       consist of b and c, and then a goto from that cluster (i.e., either from
1100       b or from c) to d.
1101    
1102       Note that (depending on the user program) the amount of information
1103       that the back-trace module has to keep track of at each level is bounded
1104       by a constant.  Thus, the whole implementation has the same asymptotical
1105       complexity as the original program (both in space and in time).
1106    
1107    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1108       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1109       which is raised by the "trigger" function mentioned above.
1110    
1111    Notes on usage:
1112    
1113    - Annotated code works well together with unannotated code:
1114    Unannotated calls simply do not show up at all in the backtrace.
1115    
1116    - It is not a good idea to let modules that were annotated during
1117    different sessions run at the same time.  This is because the compiler
1118    chooses small integers to identify individual functions, and there
1119    will be clashes if different modules were compiled in separate sessions.
1120    (Nothing will crash, and you will even be told about the clashes, but
1121    back-trace information will in general not be useful.)
1122    
1123    - Back-tracing can be confused by callcc and capture.
1124    
1125    - The only way of getting a back-trace right now is to explicitly
1126    invoke the "trigger" function from your user program.  Eventually, we
1127    should make every exception carry back-trace information (if
1128    available).  But since this creates more overhead at "raise"-time
1129    (similar to the current exnHistory overhead), I have not yet
1130    implemented this.  (The implementation will be rather easy.)  With
1131    exceptions carrying back-trace information, this facility will be even
1132    more useful because users don't need to modify their programs...
1133    
1134    - While it is possible to compile the compiler with back-trace
1135    annotations turned on (I did it to get some confidence in
1136    correctness), you must make absolutely sure that core.sml and
1137    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1138    be compiled with annotation because there is no core access yet, but
1139    if you compile btimp.sml with annotation, then the system will go into
1140    an infinite recursion and crash.)
1141    Since CM currently does not know about BTrace, the only way to turn
1142    annotations on and off for different modules of the compiler is to
1143    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1144    this is awkward and clumsy.
1145    
1146    Sample sessions:
1147    
1148    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1149    - SMLofNJ.Internals.BTrace.mode (SOME true);
1150    [autoloading]
1151    [autoloading done]
1152    val it = false : bool
1153    - structure X = struct
1154    -     fun main n = let
1155    -         fun a (x, 0) = d x
1156    -           | a (x, n) = b (x, n - 1)
1157    -         and b (x, n) = c (x, n)
1158    -         and c (x, n) = a (x, n)
1159    -         and d x = e (x, 3)
1160    -         and e (x, 0) = f x
1161    -           | e (x, n) = e (x, n - 1)
1162    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1163    -           | f n = n * g (n - 1)
1164    -         and g n = a (n, 3)
1165    -     in
1166    -         f n
1167    -     end
1168    - end;
1169    structure X : sig val main : int -> int end
1170    - X.main 3;
1171    *** BACK-TRACE ***
1172    GOTO   stdIn:4.2-13.20: X.main[2].f
1173    GOTO-( stdIn:4.2-13.20: X.main[2].e
1174    GOTO   stdIn:4.2-13.20: X.main[2].d
1175         / stdIn:4.2-13.20: X.main[2].a
1176         | stdIn:4.2-13.20: X.main[2].b
1177    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1178    CALL   stdIn:4.2-13.20: X.main[2].g
1179    GOTO   stdIn:4.2-13.20: X.main[2].f
1180    GOTO-( stdIn:4.2-13.20: X.main[2].e
1181    GOTO   stdIn:4.2-13.20: X.main[2].d
1182         / stdIn:4.2-13.20: X.main[2].a
1183         | stdIn:4.2-13.20: X.main[2].b
1184    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1185    CALL   stdIn:4.2-13.20: X.main[2].g
1186    GOTO   stdIn:4.2-13.20: X.main[2].f
1187    GOTO-( stdIn:4.2-13.20: X.main[2].e
1188    GOTO   stdIn:4.2-13.20: X.main[2].d
1189         / stdIn:4.2-13.20: X.main[2].a
1190         | stdIn:4.2-13.20: X.main[2].b
1191    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1192    CALL   stdIn:4.2-13.20: X.main[2].g
1193    GOTO   stdIn:4.2-13.20: X.main[2].f
1194    CALL   stdIn:2.15-17.4: X.main[2]
1195    -
1196    
1197    (Note that because of a FLINt bug the above code currently does not
1198    compile without BTrace turned on.)
1199    
1200    Here is another example, using my modified Tiger compiler:
1201    
1202    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1203    - SMLofNJ.Internals.BTrace.mode (SOME true);
1204    [autoloading]
1205    [autoloading done]
1206    val it = false : bool
1207    - CM.make "sources.cm";
1208    [autoloading]
1209    ...
1210    [autoloading done]
1211    [scanning sources.cm]
1212    [parsing (sources.cm):parse.sml]
1213    [creating directory CM/SKEL ...]
1214    [parsing (sources.cm):tiger.lex.sml]
1215    ...
1216    [wrote CM/sparc-unix/semant.sml]
1217    [compiling (sources.cm):main.sml]
1218    [wrote CM/sparc-unix/main.sml]
1219    [New bindings added.]
1220    val it = true : bool
1221    - Main.compile ("../testcases/merge.tig", "foo.out");
1222    *** BACK-TRACE ***
1223    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
1224    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1225    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1226    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1227    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1228    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1229    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
1230         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
1231    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
1232    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1233    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
1234    CALL   main.sml:1.18-118.4: Main.compile[2]
1235    -
1236    
1237    ----------------------------------------------------------------------
1238    Name: Matthias Blumen
1239    Date: 2000/06/21 18:00:00 JST
1240    Tag: blume-20000621-manual
1241    Description:
1242    
1243    CM manual update: Path environments documented.
1244    
1245    ----------------------------------------------------------------------
1246    Name: Matthias Blume
1247    Date: 2000/06/19 13:40:00
1248    Tag: blume-20000619-manual
1249    Description:
1250    
1251    CM manual and system/README update.  This only covers the fact that
1252    there are no more implicit anchors.  (Path environments and the "bind"
1253    option to "cm" have yet to be documented.)
1254    
1255    ----------------------------------------------------------------------
1256    Name: Matthias Blume
1257    Date: 2000/06/19 11:05:00 JST
1258    Tag: blume-20000619-chdir-bugfix
1259    Description:
1260    
1261    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1262    
1263    ----------------------------------------------------------------------
1264    Name: Matthias Blume
1265    Date: 2000/06/18 22:00:10 JST
1266    Tag: blume-20000618-implicit-anchors-really-gone
1267    Description:
1268    
1269    I updates the previous HISTORY entry where I forgot to mention that
1270    implicit anchors are no longer with us.
1271    
1272    The current update also gets rid of the (now useless) controller
1273    CM.Control.implicit_anchors.
1274    
1275    ----------------------------------------------------------------------
1276    Name: Matthias Blume
1277    Date: 2000/06/16 17:30:00 JST
1278    Tag: blume-20000616-anchorenv
1279    Description:
1280    
1281    This patch implements the long anticipated (just kidding :) "anchor
1282    environment" mechanism.  In the course of doing this, I also
1283    re-implemented CM's internal "SrcPath" module from scratch.  The new
1284    one should be more robust in certain boundary cases.  In any case, it
1285    is a lot cleaner than its predecessor (IMHO).
1286    
1287    This time, although there is yet another boot file format change, I
1288    kept the unpickler backward-compatible.  As a result, no new bootfiles
1289    are necessary and bootstrapping is straightforward.  (You cannot read
1290    new bootfiles into an old system, but the other way around is no
1291    problem.)
1292    
1293    Visible changes:
1294    
1295    ** 0. Implicit path anchors (without the leading $-symbol) are no
1296    longer recognized at all. This means that such path names are not
1297    illegal either.  For example, the name basis.cm simply refers to a
1298    local file called "basis.cm" (i.e, the name is an ordinary path
1299    relative to .cm-files directory).  Or, to put it differently, only
1300    names that start with $ are anchored paths.
1301    
1302    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1303    vanished.
1304    
1305    John (Reppy) had critizised this as soon as I originally proposed and
1306    implemented it, but at that time I did not really deeply believe
1307    him. :) Now I came full-circle because I need the $<singlearc> syntax
1308    in another place where it cannot be seen as an abbreviation for
1309    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1310    seems to mean (i.e., it "expands" into the corresponding anchor
1311    value).
1312    
1313    However, when paths are used as members in CM description files, it
1314    continues to be true that there must be at least another arc after the
1315    anchor.  This is now enforced separately during semantic analysis
1316    (i.e., from a lexical/syntactical point of view, the notation is ok.)
1317    
1318    ** 2. The "cm" class now accepts an option "bind".  The option's value
1319    is a sub-option list of precisely two items -- one labeled "anchor"
1320    and the other one labeled "value".  As you might expect, "anchor" is
1321    used to specify an anchor name to be bound, and "value" specifies what
1322    the anchor is being bound to.
1323    
1324    The value must be a directory name and can be given in either standard
1325    syntax (including the possibility that it is itself an anchored path)
1326    or native syntax.
1327    
1328    Examples:
1329    
1330       foo.cm (bind:(anchor:bar value:$mystuff/bar))
1331       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1332    
1333    and so on.
1334    
1335    The meaning of this is that the .cm-file will be processed with an
1336    augmented anchor environment where the given anchor(s) is/are bound to
1337    the given values(s).
1338    
1339    The rationale for having this feature is this: Suppose you are trying
1340    to use two different (already stable) libraries a.cm and b.cm (that
1341    you perhaps didn't write yourself).  Further, suppose each of these
1342    two libraries internally uses its own auxiliary library $aux/lib.cm.
1343    Normally you would now have a problem because the anchor "lib" can not
1344    be bound to more than one value globally.  Therefore, the project that
1345    uses both a.cm and b.cm must locally redirect the anchor to some other
1346    place:
1347    
1348       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1349       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1350    
1351    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1352    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
1353    
1354    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1355    you when you do so at the time of CM.stabilize).  Therefore, you can
1356    also use an anchored path as the value:
1357    
1358      a.cm (bind:(anchor:lib value:$a-lib))
1359      b.cm (bind:(anchor:lib value:$b-lib))
1360    
1361    Now you can globally configure (using the usual CM.Anchor.anchor or
1362    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1363    itself is always locally bound, setting it globally is no longer
1364    meaningful or necessary (but it does not hurt either).  In fact, "lib"
1365    can still be used as a global anchor for separate purposes.  As a
1366    matter of fact, one can locally define "lib" in terms of a global
1367    "lib":
1368    
1369      a.cm (bind:(anchor:lib value:$lib/a))
1370      b.cm (bind:(anchor:lib value:$lib/b))
1371    
1372    ** 3: The encoding of path names has changed.  This affects the way
1373    path names are shown in CM's progress report and also the internal
1374    protocol encoding used for parallel make.
1375    
1376    The encoding now uses one or more ':'-separated segments.  Each
1377    segments corresponds to a file that has been specified relative to the
1378    file given by its preceding segment.  The first segment is either
1379    relative to the CWD, absolute, or anchored.  Each segment itself is
1380    basically a Unix pathname; all segments but the first are relative.
1381    
1382    Example:
1383    
1384       $foo/bar/baz.cm:a/b/c.sml
1385    
1386    This path denotes the file bar/a/b/c.sml relative to the directory
1387    denoted by anchor "foo".  Notice that the encoding also includes
1388    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1389    paths are resolved relative to the .cm-files directory, so baz.cm must
1390    be ignored to get the "real" pathname.
1391    
1392    To make this fact more obvious, CM puts the names of such "virtual
1393    arcs" into parentheses when they appear in progress reports. (No
1394    parentheses will appear in the internal protocol encoding.)  Thus,
1395    what you really see is:
1396    
1397      $foo/bar/(baz.cm):a/b/c.sml
1398    
1399    I find this notation to be much more informative than before.
1400    
1401    Another new feature of the encoding is that special characters
1402    including parentheses, colons, (back)slashes, and white space are
1403    written as \ddd (where ddd is the decimal encoding of the character).
1404    
1405    *** The CM manual still needs to be updated.
1406    
1407    ----------------------------------------------------------------------
1408    Name: Allen Leung
1409    Date: 2000/06/15 00:38:00
1410    Tag: leunga-20000615-x86-peephole
1411    
1412    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1413    
1414    ----------------------------------------------------------------------
1415    Name: Matthias Blume
1416    Date: 2000/06/12 11:40:00
1417    Tag: blume-20000612-parmakefix
1418    Description:
1419    
1420    More cleanup after changing the file naming scheme: This time I
1421    repaired the parallel make mechanism for CMB.make which I broke earlier.
1422    
1423    ----------------------------------------------------------------------
1424    Name: Allen Leung
1425    Date: 2000/06/09 01:25:00
1426    Tag: leunga-20000609-various
1427    
1428    None of these things should affect normal SML/NJ operations
1429    
1430    1. Peephole improvements provided by Fermin (c--)
1431    2. New annotation DEFUSE for adding extra dependence (moby)
1432    3. New X86 LOCK instructions (moby)
1433    4. New machine description language for reservation tables (scheduling)
1434    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1435       trees etc.)
1436    6. I've changed the CM files so that they can work with versions
1437       110.0.6, 110.25 and 110.28
1438    
1439    ----------------------------------------------------------------------
1440    Name: Matthias Blume
1441    Date: 2000/06/09 12:40:00
1442    Tag: blume-20000609-log
1443    Description:
1444    
1445    - Removed all(?) remaining RCS Log entries from sources.
1446    
1447    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1448      anchored paths).
1449    
1450    ----------------------------------------------------------------------
1451    Name: Matthias Blume
1452    Date: 2000/06/07 17:00:00 JST
1453    Tag: blume-20000607-no-implicit-anchors
1454    Description:
1455    
1456    1. This update changes the default setting for
1457    CM.Control.implicit_anchors from true to false.  This means that
1458    implicit anchors are no longer permitted by default.  I also tried to
1459    make sure that nothing else still relies on implicit anchors.
1460    (This is the next step on the schedule towards a CM that does not even
1461    have the notion of implicit anchors anymore.)
1462    
1463    2. More CM manual updates.
1464    
1465    3. I managed to track down and fix the pickling bug I mentioned last
1466    time.  Because of the previously existing workaround, this entails no
1467    immediate practical changes.
1468    
1469    ----------------------------------------------------------------------
1470    Name: Matthias Blume
1471    Date: 2000/06/06 11:15:00 JST
1472    Tag: blume-20000606-lazierpickle
1473    Description:
1474    
1475    !!!! NEW BOOT FILES !!!!
1476    
1477    * The main purpose of this update is to make library pickles lazier in
1478    order to reduce the initial space penalty for autoloading a library.
1479    As a result, it is now possible to have $smlnj/compiler.cm
1480    pre-registered.  This should take care of the many complaints or
1481    inquiries about missing structure Compiler.  This required changes to
1482    CM's internal data structures and small tweaks to some algorithms.
1483    
1484    As a neat additional effect, it is no longer necessary (for the sake
1485    of lean heap image files) to distinguish between a "minimal" CM and a
1486    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1487    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1488    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1489    
1490    To make the life of compiler-hackers easier, "makeml" now also
1491    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1492    words, after you bootstrap a new sml for the first time, you will not
1493    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1494    around you will still have to do it, though.)
1495    
1496    * A second change consists of major updates to the CM manual.  There
1497    are now several appendices with summary information and also a full
1498    specification of the CM description file syntax.
1499    
1500    * In directory src/system I added the script "allcross".  This script
1501    invokes sml and cross-compiles the compiler for all supported
1502    architectures.  (Useful when providing a new set of boot files.)
1503    
1504    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1505    added a small tweak to pickle-util.sml to work around this problem,
1506    but it is not a proper fix yet.  I will investigate further.  (The
1507    effect of the bug was an inflation of library pickle size.)
1508    
1509    * Version number increased to 110.28.1 (to avoid compatibility problems).
1510    
1511    ----------------------------------------------------------------------
1512    Name: Allen Leung
1513    Date: 2000/05/25 17:28 EDT
1514    Tag: leunga-20000525-ra
1515    Description:
1516    
1517      Fixed a bug in freezing phase of the register allocator.
1518    
1519    ----------------------------------------------------------------------
1520    Name: Allen Leung
1521    Date: 2000/05/15 22:53 EDT
1522    Tag: leunga-20000515-alpha-x86-ra
1523    Description:
1524    
1525      1. Alpha
1526    
1527          Slight cleanup.  Removed the instruction SGNXL
1528    
1529      2. X86
1530    
1531          Added the following instructions to the instruction set:
1532    
1533            ROLx, RORx,
1534            BTx, BTSx, BTLx, BTRx,
1535            XCHGx, and variants with the LOCK prefix
1536    
1537      3. Register Allocation
1538    
1539          The module ra-rewrite-with-renaming has been improved.
1540    
1541      These have no effect on SML/NJ.
1542    
1543    ----------------------------------------------------------------------
1544    Name: Matthias Blume
1545    Date: 2000/05/15 16:20:00 JST
1546    Tag: blume-20000515-lightrebuild
1547    Description:
1548    
1549    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1550       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1551       defined.  In effect, the command will not build any cross-compiler
1552       backends and therefore finish more quickly.
1553    
1554       The "fixpt" script also takes a "-light" switch to be able to use
1555       this new facility while compiling for a fixpoint.
1556    
1557    2. I replaced all mentions of anchored paths in group owner specifications
1558       with simple relative paths (usually starting with "..").
1559       The rationale is that a library's internal workings should not be
1560       compromised by the lack of some anchor.  (An anchor is necessary
1561       for someone who wants to refer to the library by an anchored path,
1562       but it should not be necessary to build the same library in the first
1563       place.)
1564    
1565    3. I changed the way CM's tool mechanism determines the shell command
1566       string used for things like ml-yacc etc. so that it does not break
1567       when CM.Control.implicit_anchors is turned off.
1568    
1569    ----------------------------------------------------------------------
1570    Name: Matthias Blume
1571    Date: 2000/05/12 18:20:00 JST
1572    Tag: blume-20000512-ml-build
1573    Description:
1574    
1575    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1576    getting installed properly (by config/install.sh).
1577    
1578    ----------------------------------------------------------------------
1579    Name: Matthias Blume
1580    Date: 2000/05/12 17:30:00 JST
1581    Tag: blume-20000512-anchors
1582    Description:
1583    
1584    !!! NEW BOOT FILES !!!
1585    
1586    This change is in preparation of fading out support for "implicitly
1587    anchored path names".  I went through all sources and used the
1588    explicit (and relatively new) $-notation.  See system/README and the
1589    CM manual for more info on this.
1590    
1591    I also modified the anchoring scheme for some things such as "smlnj",
1592    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1593    anchors are more expressive: anchor name and first arc do not have to
1594    coincide.  This entails the following user-visible change:
1595    
1596    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1597    particular, when you fire up sml with a command-line argument, say,
1598    e.g.:
1599    
1600       sml '$smlnj/cmb.cm'
1601    
1602    At the ML toplevel prompt:
1603    
1604       CM.autoload "$smlnj/cmb.cm";
1605    
1606    There is also a new controller in CM.Control that can be used to turn
1607    off all remaining support for implicit anchors by saying:
1608    
1609        CM.autoload "$smlnj/
1610        #set CM.Control.implicit_anchors false;
1611    
1612    This causes CM to reject implicitly anchored paths.  This is (for the
1613    time being) less permissive than the "final" version where there will
1614    be no more such implicit anchors and relative paths will be just that:
1615    relative.
1616    
1617    The next step (version after next version?) will be to make the
1618    default for CM.Control.implicit_anchors false.  After the dust has
1619    settled, I can then produce the "final" version of this...
1620    
1621    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1622    
1623    ----------------------------------------------------------------------
1624    Name: Matthias Blume
1625    Date: 2000/05/11 16:30:00 JST
1626    Tag: blume-20000511-sources
1627    Description:
1628    
1629    The main change is that I added function CM.sources as a generalized
1630    version of the earlier CM.makedepend.  This entails the following
1631    additional changes:
1632    
1633      - CM.makedepend has been dropped.
1634    
1635      - CM manual has been updated.
1636    
1637      - TOOLS signature and API have been changed.
1638    
1639    ----------------------------------------------------------------------
1640    Name: Allen Leung
1641    Date: 2000/05/10 21:17 EDT
1642    Tag: leunga-20000510-moby-c--ssa
1643    Description:
1644    
1645      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1646    None of these affect SML/NJ.
1647    
1648    1. Register Allocation
1649    
1650        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1651           This module tries to remove local (i.e. basic block level) redundancies
1652           during spilling.
1653    
1654        b. A new framework for performing region based register allocation.
1655           Not yet entirely functional.
1656    
1657    2. X86
1658    
1659       a. DefUse for POP was missing the stack pointer [found by Lal]
1660       b. Reload for CALL was incorrect in X86Spill [found by John]
1661       c. Various fixes in X86Spill so that it can be used correctly for
1662          the new spilling module.
1663    
1664    3. SSA/IR
1665    
1666       a. New module ir/dj-dataflow.sml implements elimination based
1667          data flow analysis.
1668    
1669    4. MLRiscGen
1670    
1671       a. Fix for gc type annotation
1672    
1673    5. MDGen
1674    
1675       Various fixes for machine description -> ml code translation.  For ssa
1676       only.
1677    
1678    ----------------------------------------------------------------------
1679    Name: Allen Leung
1680    Date: 2000/05/08 22:17 EDT
1681    Tag: leunga-20000508-labexp
1682    Description:
1683    
1684      Fermin has found a few assembly problems with constant expressions
1685      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1686      which choke on dumb assemblers.  This is his fix.
1687    
1688    ----------------------------------------------------------------------
1689    Name: Dave MacQueen
1690    Date: 2000/04/09 14:00 EDT
1691    Tag: dbm-20000502-Version_110_28
1692    Description:
1693    
1694    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1695    
1696    2. Updated config/version to 110.28
1697    
1698    3. Updated config/srcarchiveurl
1699    
1700    3. New boot files!
1701       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1702    
1703    ----------------------------------------------------------------------
1704    Name: Matthias Blume
1705    Date: 2000/05/01 19:05:00 JST
1706    Tag: blume-20000501-noweb
1707    Description:
1708    
1709    A new noweb tool has been added.  The existing system is entirely
1710    unaffected by this, but some CM users have asked for renewed noweb
1711    support.  Everything is documented in the CM manual.
1712    
1713    New (plugin) libraries:
1714    
1715       noweb-tool.cm
1716       nw-ext.cm
1717    
1718    ----------------------------------------------------------------------
1719    Name: Dave MacQueen
1720    Date: 2000/04/30 12:40PM EDT
1721    Tag: dbm-20000430-bug_fixes
1722    Description:
1723    
1724    1. Fix for bug 1498
1725       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1726       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1727         added toRealArray function
1728       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1729         added check for tag Obj.RealArray to array printing case in ppObj
1730    
1731    2. Fix for bug 1510
1732       smlnj/src/compiler/Semant/types/typesutil.sml
1733         fixed definition of dummyargs (used by equalTycon) so that
1734         dummy args are distinct types
1735    
1736    ----------------------------------------------------------------------
1737    Name: Matthias Blume
1738    Date: 2000/04/30 01:00:00 JST
1739    Tag: blume-20000430-versions
1740    Description:
1741    
1742    1. CM version numbering added.  This is an implementation of Lal's
1743       proposal for adding version numbers and version checking to .cm
1744       files.  Lal said that his proposal was just that -- a proposal.
1745       For the time being I went ahead and implemented it so that people
1746       can comment on it.  Everything is completely backward-compatible
1747       (except for the stable library format, i.e., new bootfiles!).
1748    
1749       As usual, see the CM manual for details.
1750    
1751    2. An alternative syntax for anchored paths has been implemented.
1752       Dave has recently voiced the same concerns that I had when I did
1753       this, so there should be some support.  My take is that eventually
1754       I will let support for the current syntax (where anchors are
1755       "implicit") fade out in favor of the new, explicit syntax.
1756       In order to be backward-compatible, both old and new syntax are
1757       currently supported.
1758    
1759       Again, see the CM manual for details.
1760    
1761    3. Parallel make is trying to be slightly smarter:  When the master
1762       process finds a "bottleneck", i.e., when there is only one
1763       compilation unit that can be compiled and everybody else is
1764       waiting on it, then it will simply compile it directly instead
1765       of clumsily telling one of the slaves to do it.
1766    
1767    4. Support for "unsharing" added.  This is necessary in order to be
1768       able to have two different versions of the same library running
1769       at the same time (e.g., for trying out a new MLRISC while still
1770       having the old MLRISC linked into the current compiler, etc.)
1771       See the CM manual.
1772    
1773    5. Simple "makedepend" functionality added for generating Makefile
1774       dependency information.  (This is rather crude at the moment.
1775       Expect some changes here in the future.)
1776    
1777    6. ".fun" added as a recognized suffix for ML files. Also documented
1778       explicitly in the manual that the fallback behavior (unknown suffix
1779       -> ML file) is not an official feature!
1780    
1781    7. Small changes to the pickler for stable libraries.
1782    
1783    8. Several internal changes to CM (for cleanup/improvement).
1784    
1785    
1786    !!!! NEW BINFILES !!!!
1787    
1788    ----------------------------------------------------------------------
1789    Name: Matthias Blume
1790    Date: 2000/04/28 17:30:00 JST
1791    Tag: blume-20000428-pathconfig
1792    Description:
1793    
1794    1. I changed config/install.sh to remove duplicate entries from the
1795       lib/pathconfig file at the end.  Moreover, the final version of
1796       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1797       in src/system/installml.
1798    
1799    2. The config/install.sh script now consistently uses relative
1800       pathnames in lib/pathconfig whenever the anchor is in the lib
1801       directory.  (So far this was true for the libraries that come
1802       pre-compiled and bundled as part of the bootfiles but not for
1803       libraries that are compiled by the script itself.)
1804    
1805    ----------------------------------------------------------------------
1806    Name: Matthias Blume
1807    Date: 2000/04/26 13:10:00 JST
1808    Tag: blume-20000426-fun_suffix
1809    Description:
1810    
1811    Added ".fun" as a recognized file name suffix (for ML code).
1812    
1813    ----------------------------------------------------------------------
1814    Name: Allen Leung
1815    Date: 2000/04/25 17:00:00 EST
1816    Tag: leunga-20000425-alpha-ra
1817    Description:
1818    
1819    1. Alpha
1820    
1821        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1822    in C--.
1823    
1824    2. RA
1825    
1826       Added a flag "ra-dump-size" to print out the size of the flowgraph
1827       and the interference graph.
1828    
1829    ----------------------------------------------------------------------
1830    Name: Dave MacQueen
1831    Date: 2000/04/25/
1832    Tag: dbm-20000425-mlyacc_doc_examples
1833    Description:
1834      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1835      Updated all three examples in src/ml-yacc/examples to run
1836      under 110.* using CM.make.
1837    
1838    ----------------------------------------------------------------------
1839    Name: Allen Leung
1840    Date: 2000/04/20 23:04:00 EST
1841    Tag: leunga-20000420-ssa-c---stuff
1842    Description:
1843    
1844      This update synchronizes my repository with Yale's.  Most of these
1845    changes, however, do not affect SML/NJ at all (the RA is an exception).
1846    
1847    1. Register Allocator
1848    
1849       a. An improvement in the interference graph construction:
1850          Given a copy
1851    
1852                s <- t
1853    
1854          no interference edge between s and t is added for this definition of s.
1855    
1856       b. I've added two new spill heuristic modules that Fermin and I developed
1857          (in the new library RA.cm). These are unused in SML/NJ but maybe
1858          useful for others (Moby?)
1859    
1860    2. X86
1861    
1862       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1863    
1864    3. Alpha
1865    
1866       a. Added the BSR instruction and code generation that goes with it [C--]
1867       b. Other fixes too numerous to recount provided by Fermin [C--]
1868    
1869    4. Regmaps
1870    
1871       a. The regmaps are not initialized with the identity physical bindings
1872          at creation time.  This is unneeded.
1873    
1874    5. MLRISC Optimizations
1875    
1876       a. The DJ-Graph module can now compute the iterated dominance frontiers
1877          intersects with liveness incrementally in linear time! Woohoo!
1878          This is now used in my new SSA construction algorithm.
1879    
1880       b. THe branch reorganization module is now smarter about linear chains of
1881          basic blocks.
1882    
1883    
1884    ----------------------------------------------------------------------
1885    Name: Matthias Blume
1886    Date: 2000/04/12 13:52:00 JST
1887    Tag: blume_main_v110p27_1
1888    Description:
1889    
1890    Changed install.sh script to handle archive files without version number
1891    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1892    name of the boot file archive.
1893    
1894    ----------------------------------------------------------------------
1895    Name: Dave MacQueen
1896    Date: 2000/04/09 14:00 EDT
1897    Tag: dbm-20000410-Version_110_27
1898    Description:
1899    
1900    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1901    
1902    2. Updated src/config/version to 110.27
1903    
1904    3. New boot files!
1905    
1906    ----------------------------------------------------------------------
1907    Name: Allen Leung
1908    Date: 2000/04/09 19:09:00 EST
1909    Tag: leunga-20000409-misc
1910    Description:
1911    
1912    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1913    
1914    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1915    
1916    ----------------------------------------------------------------------
1917    Name: Stefan
1918    Date: 2000/04/07 10:00:00 EDT
1919    Tag: monnier-20000406-branch-handling
1920    Description:
1921    
1922    Improved handling of branches (mostly those generated from
1923    polymorphic equality), removed switchoff and changed the
1924    default optimization settings (more cpsopt and less flintopt).
1925    
1926    ----------------------------------------------------------------------
1927    Name: Allen Leung
1928    Date: 2000/04/06 01:30:00 EST
1929    Tag: leunga-20000406-peephole-x86-SSA-2
1930    Description:
1931    
1932       Forgot a few files.
1933    
1934    ----------------------------------------------------------------------
1935    Name: Allen Leung
1936    Date: 2000/04/06 00:36:00 EST
1937    Tag: leunga-20000406-peephole-x86-SSA
1938    Description:
1939    
1940    1.  New Peephole code
1941    
1942    2.  Minor improvement to X86 instruction selection
1943    
1944    3.  Various fixes to SSA and machine description -> code translator
1945    
1946    ----------------------------------------------------------------------
1947    Name: Matthias Blume
1948    Date: 2000/04/05 12:30:00 JST
1949    Tag: blume_main_v110p26p2_3
1950    Description:
1951    
1952    This update just merges three minor cosmetic updates to CM's sources
1953    to get ready for the 110.27 code freeze on Friday.  No functionality
1954    has changed.
1955    
1956    ----------------------------------------------------------------------
1957    Name: Allen Leung
1958    Date: 2000/04/04 19:39:00 EST
1959    Tag: leunga-20000404-x86-asm
1960    Description:
1961    
1962    1.  Fixed a problem in X86 assembly.
1963    
1964        Things like
1965    
1966           jmp %eax
1967           jmp (%eax)
1968    
1969        should be output as
1970    
1971           jmp *%eax
1972           jmp *(%eax)
1973    
1974    2.  Assembly output
1975    
1976          Added a new flag
1977    
1978              "asm-indent-copies" (default to false)
1979    
1980          When this flag is on, parallel copies will be indented an extra level.
1981    
1982    ----------------------------------------------------------------------
1983    Name: Allen Leung
1984    Date: 2000/04/04 03:18:00 EST
1985    Tag: leunga-20000404-C--Moby
1986    Description:
1987    
1988        All of these fixes are related to C--, Moby, and my own optimization
1989        stuff; so they shouldn't affect SML/NJ.
1990    
1991    1.  X86
1992    
1993        Various fixes related floating point, and extensions.
1994    
1995    2.  Alpha
1996    
1997        Some extra patterns related to loads with signed/zero extension
1998        provided by Fermin.
1999    
2000    3.  Assembly
2001    
2002        When generating assembly, resolve the value of client defined constants,
2003        instead of generating symbolic values.  This is controlled by the
2004        new flag "asm-resolve-constants", which is default to true.
2005    
2006    4.  Machine Descriptions
2007    
2008        a. The precedence parser was slightly broken when parsing infixr symbols.
2009        b. The type generalizing code had the bound variables reversed, resulting
2010           in a problem during arity raising.
2011        c. Various fixes in machine descriptions.
2012    
2013    ----------------------------------------------------------------------
2014    Name: Matthias Blume
2015    Date: 2000/04/03 16:05:00 JST
2016    Tag: blume_main_v110p26p2_2
2017    Description:
2018    
2019    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
2020    now done via the ordinary static environment that is context to each
2021    compilation unit.
2022    
2023    To this end, I arranged that instead of "structure Core" as "structure
2024    _Core" is bound in the pervasive environment.  Core access is done via
2025    _Core (which can never be accidentally rebound because _Core is not a
2026    legal surface-syntax symbol).
2027    
2028    The current solution is much cleaner because the core environment is
2029    now simply part of the pervasive environment which is part of every
2030    compilation unit's context anyway.  In particular, this eliminates all
2031    special-case handling that was necessary until now in order to deal
2032    with dynamic and symbolic parts of the core environment.
2033    
2034    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2035    compilation manager's bootstrap compiler (actually: in the "init group"
2036    handling).  See the comments in src/system/smlnj/init/init.cmi for
2037    more details.
2038    
2039    I also tried to track down all mentions of "Core" (as string argument
2040    to Symbol.strSymbol) in the compiler and replaced them with a
2041    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
2042    name appears in one place only.
2043    
2044    Binfile and bootfile format have not changed, but the switchover from
2045    the old "init.cmi" to the new one is a bit tricky, so I supplied new
2046    bootfiles anyway.
2047    
2048  ----------------------------------------------------------------------  ----------------------------------------------------------------------
2049  Name: Allen Leung  Name: Allen Leung
2050  Date: 2000/04/02 21:17:00 EST  Date: 2000/04/02 21:17:00 EST
# Line 113  Line 2147 
2147     (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
2148     used).     used).
2149    
2150     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
2151     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
2152     Makefiles or Shell scripts.     Makefiles or Shell scripts.
2153    
2154     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 126  Line 2160 
2160    
2161     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
2162    
2163     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
2164     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
2165    
2166  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 260  Line 2294 
2294    
2295       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
2296       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
2297       testl is perferable to andl.       testl is preferable to andl.
2298    
2299  7. RA (x86 only)  7. RA (x86 only)
2300    
# Line 442  Line 2476 
2476    
2477  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2478  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2479  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
2480  explanation.  explanation.
2481    
2482  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2483  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2484  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2485  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 489  Line 2523 
2523  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
2524  <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,
2525  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
2526  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2527  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2528  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2529    
# Line 539  Line 2573 
2573    it from that remote directory.    it from that remote directory.
2574    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2575    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2576    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2577    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2578    
2579    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.591  
changed lines
  Added in v.824

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