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 590, Sat Apr 1 02:24:08 2000 UTC revision 828, Fri May 11 18:32:07 2001 UTC
# Line 13  Line 13 
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Matthias Blume
17    Date: 2001/05/11 14:35:00 EDT
18    Tag: blume-20010511-ml-nlffi
19    Description:
20    
21    I finally took the plunge and added my new FFI code to the main
22    repository.  For x86-linux it is now ready for prime-time.
23    
24    There are two new subdirectories of "src":
25    
26      - ml-nlffi-lib:
27           The utility library for programs using the FFI interface.
28           Here is the implementation of $/c.cm and its associated low-level
29           partners $/c-int.cm and $/memory.cm.
30      - ml-nlffigen:
31           A stand-alone program for generating ML glue code from C source
32           code.
33    
34    Building ml-nlffigen requires $/ckit-lib.cm.
35    
36    The config/install.sh script has been updates to do the Right Thing
37    (hopefully).
38    
39    Notice that the source tree for the C-Kit will not be put under "src"
40    but directly under the installation root directory.  (This is the
41    structure that currently exists on the CVS server when you check out
42    module "sml".)  Fortunately, config/install.sh knows about this oddity.
43    
44    Bugs: No documentation yet.
45    
46    ----------------------------------------------------------------------
47    Name: Matthias Blume
48    Date: 2001/05/09 16:35:00 EDT
49    Tag: blume-20010509-cpscontract
50    Description:
51    
52    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
53    wrapper/unwrapper elimination did not decrement usage counts and some
54    dead variables got overlooked by the dead-up logic.)
55    
56    ----------------------------------------------------------------------
57    Name: Lal George
58    Date: 2001/05/08  17:26:09 EDT
59    Tag: george-20010508-omit-frameptr
60    Description:
61    
62    Changes to implement the omit-frame-pointer optimization to support
63    raw C calls. For now, there is only support on the Intel x86, but
64    other architectures will follow as more experience is gained with this.
65    
66    
67    ----------------------------------------------------------------------
68    Name: Matthias Blume
69    Date: 2001/05/07 14:40:00 EDT
70    Tag: blume-20010507-proxies
71    Description:
72    
73    I made into "proxy libraries" all libraries that qualify for such a
74    change.  (A qualifying library is a library that has another library or
75    groups as its sole member and repeats that member's export list
76    verbatim.  A proxy library avoids this repetition by omitting its export
77    list, effectively inheriting the list that its (only) member exports.
78    See the CM manual for more explanation.)
79    The main effect is that explicit export lists for these libraries
80    do not have to be kepts in sync, making maintenance a bit easier.
81    
82    I also added copyright notices to many .cm-files.
83    
84    Last but not least, I made a new set of bootfiles.
85    
86    ----------------------------------------------------------------------
87    Name: Matthias Blume
88    Date: 2001/05/04 17:00:00 EDT
89    Tag: blume-20010504-cm-lsplit
90    Description:
91    
92    0. John merged pending changes to $/smlnj-lib.cm
93    
94    1. Allen's previous change accidentally backed out of one of Lal's
95       earlier changes.  I undid this mistake (re-introducing Lal's change).
96    
97    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
98       within the compiler where applicable.  There is some code simplification
99       because of that.
100    
101    3. The "split" phase (in FLINT) is now part of the default list of phases.
102       Compiler.Control.LambdaSplitting.* can be used to globally control the
103       lambda-splitting (cross-module-inlining) engine.  In addition to that,
104       it can now also be controlled on a per-source basis: CM has been taught
105       a new tool parameter applicable to ML source files.
106    
107       - To turn lambda-splitting off completely:
108            local open Compiler.Control.LambdaSplitting in
109                val _ = set Off
110            end
111       - To make "no lambda-splitting" the global default (but allow per-source
112         overriding); this is the initial setting:
113            local open Compiler.Control.LambdaSplitting in
114                val _ = set (Default NONE)
115            end
116       - To make "lambda-splitting with aggressiveness a" the global default
117         (and allow per-source overriding):
118            local open Compiler.Control.LambdaSplitting in
119                val _ = set (Default (SOME a))
120            end
121    
122       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
123         write (in the respective .cm-file):
124            a.sml (lambdasplitting:off)
125       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
126            a.sml (lambdasplitting:on)
127       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
128         <a> is a decimal non-negative integer):
129            a.sml (lambdasplitting:<a>)
130       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
131            a.sml (lambdasplitting:infinity)
132       - To use the global default for a.sml:
133            a.sml (lambdasplitting:default)
134         or simply
135            a.sml
136    
137    ----------------------------------------------------------------------
138    Name: Allen Leung
139    Date: 2001/05/04 01:57:00 EDT
140    Tag: leunga-20010504-sync
141    Description:
142    
143      MLRISC features.
144    
145      1. Fix to CMPXCHG instructions.
146      2. Changed RA interface to allow annotations in callbacks.
147      3. Added a new method to the stream interface to allow annotations updates.
148    
149    ----------------------------------------------------------------------
150    Name: Matthias Blume
151    Date: 2001/05/01 11:45:00 EDT
152    Tag: blume-20010501-pcedittmp
153    Description:
154    
155    Changed install.sh to use the current working directory instead of
156    /usr/tmp for a temporary file (pcedittmp).  The previous choice
157    of /usr/tmp caused trouble with MacOS X because of file premission
158    problems.
159    
160    ----------------------------------------------------------------------
161    Name: Matthias Blume
162    Date: 2001/04/20 11:10:00 EDT
163    Tag: blume-20010420-inMLflag
164    Description:
165    
166     - added vp_limitPtrMask to vproc-state.h
167       (for use by the raw-C-calls mechanism to implement proper interrupt
168        handling)
169     - made the ML compiler aware of various data-structure offsets so it
170       can generate code for accessing the vp_inML flag and vp_limitPtrMask
171     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
172       raw C-calls
173    
174    ----------------------------------------------------------------------
175    Name: Lal George
176    Date: 2001/04/20 09:15:28 EDT
177    Tag: george-20010420-macosX
178    Description:
179    
180     - Changes to port to Mac OS X; Darwin.
181    
182     - In the process I found that sqrt was broken on the PPC, because the
183       fsqrt instruction is not implemented.
184    
185    ----------------------------------------------------------------------
186    Name: Matthias Blume
187    Date: 2001/04/18 12:45:00 EDT
188    Tag: blume-20010418-ccalls
189    Description:
190    
191     - fixed two off-by-4 errors in the x86-specific c-calls implementation
192       (this bug prevented structure arguments containing pointers from being
193        passed correctly)
194     - changed the raw-C-call code in mlriscGen.sml in such a way that
195       structure arguments are represented as a pointer to the beginning
196       of the structure (instead of having a series of synthesized arguments,
197       one for each structure member)
198    
199     - made makeml script's verbosity level configurable via environment
200       variable (MAKEML_VERBOSITY)
201    
202     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
203       in rawmem-x86.sml; we are now using the real thing
204    
205    ----------------------------------------------------------------------
206    Name: Matthias Blume
207    Date: 2001/03/22 16:25:00 EST
208    Tag: blume-20010322-bootfiles
209    Description:
210    
211    Created a new set of bootfiles (for your automatic installation convenience).
212    
213    ----------------------------------------------------------------------
214    Name: Matthias Blume
215    Date: 2001/03/22 15:10:00 EST
216    Tag: blume-20010322-rawmem-parcm
217    Description:
218    
219    1. All "raw memory access" primitives for the new FFI are implemented now
220       (at least on the x86).
221    2. Some further cleanup of CM's parallel make mechanism.
222    
223    ----------------------------------------------------------------------
224    Name: Matthias Blume
225    Date: 2001/03/19 17:53:00 EST
226    Tag: blume-20010319-parallel
227    Description:
228    
229    Parallel make (using compile servers) now works again.
230    
231    To this end, CM.stabilize and CMB.make have been modified to work in
232    two passes when compile servers are attached:
233       1. Compile everything, do not perform stabilization; this pass
234          uses compile servers
235       2. Stabilize everything; this pass does not use compile servers
236    If there are no compile servers, the two passes are combined into one
237    (as before).  Splitting the passes increases the inherent parallelism
238    in the dependency graph because the entire graph including all
239    libraries is available at the same time.  This, in turn, improves
240    server utilization.  The downside is that the master process will
241    have to do some extra work after compilation is done (because for
242    technical reasons it must re-read all the binfiles during stabilization).
243    
244    ----------------------------------------------------------------------
245    Name: Matthias Blume
246    Date: 2001/03/16 12:22:00 EST
247    Tag: blume-20010316-bootfiles
248    Description:
249    
250    Created a new set of bootfiles (for your automatic installation convenience).
251    
252    ----------------------------------------------------------------------
253    Name: Matthias Blume
254    Date: 2001/03/16 11:00:00 EST
255    Tag: blume-20010316-MLTREE-fixup
256    Description:
257    
258    This is a minor fixup for an (untagged) earlier commit by Allen.
259    (A file was missing).
260    
261    ----------------------------------------------------------------------
262    Name: Allen Leung
263    Date: Mon Mar  5 18:54:57 EST 2001
264    Tag: leunga-20010305-cut-support
265    
266    1. New support for alternative control-flow in MLTREE.
267       Currently we support
268    
269          FLOW_TO(CALL ...., [k1,...,kn])
270    
271       This is needed for 'cuts to' in C-- and try/handle-like constructs
272       in Moby
273    
274       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
275    
276    2. Register Allocator
277    
278       Changes in interface [from Fermin, John]
279    
280    3. Alpha 8-bit SLL support [Fermin]
281    
282    4. All architectures
283    
284       A new module (ClusterExpandCopies) for expanding parallel copies.
285    
286    ----------------------------------------------------------------------
287    Name: Allen Leung
288    Date: 2001/02/27 23:07:00 EST
289    Tag: leunga-20010227-minor-stuff
290    
291    1. Alpha bug fix for CMOVNE
292    2. Handle mltree COND(..,FCMP ...,...)
293    3. Bug fix in simplifier
294    
295    ----------------------------------------------------------------------
296    Name: Matthias Blume
297    Date: 2001/01/30 17:50:00 EST
298    Tag: blume-20010130-sync
299    Description:
300    
301    This is just a minor update to sync my devel branch with the main brach.
302    The only visible change is the addition of some README files.
303    
304    ----------------------------------------------------------------------
305    Name: Matthias Blume
306    Date: 2001/01/12 23:30:00 JST
307    Tag: blume-20010112-bootfiles
308    Description:
309    
310    Made a new set of bootfiles that goes with the current state of the
311    repository.
312    
313    ----------------------------------------------------------------------
314    Name: Matthias Blume
315    Date: 2001/01/12 21:20:00 JST
316    Tag: blume-20010112-sync
317    Description:
318    
319    I am just flushing out some minor changes that had accumulated in
320    my private branch in order to sync with the main tree.  (This is
321    mainly because I had CVS trouble when trying to merge _into_ my
322    private branch.)
323    
324    Most people should be completely unaffected by this.
325    
326    ----------------------------------------------------------------------
327    Name: Allen Leung
328    Date: Thu Jan 11 21:03:00 EST 2001
329    Tag: leunga-20010111-labexp=mltree
330    Description:
331    
332    1.  Removed the type LabelExp and replace it by MLTree.
333    2.  Rewritten mltree-simplify with the pattern matcher tool.
334    3.  There were some bugs in alpha code generator which would break
335        64-bit code generation.
336    4.  Redo the tools to generate code with the
337    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
338        are now generated by perl scripts.
339    
340    ----------------------------------------------------------------------
341    Name: Matthias Blume
342    Date: 2001/01/10 21:55:00 JST
343    Tag: blume-20010110-rcc
344    Description:
345    
346    The RCC stuff now seems to work (but only on the x86).
347    This required hacking of the c-calls interface (and -implementation) in
348    MLRISC.
349    
350    Normal compiler users should be unaffected.
351    
352    ----------------------------------------------------------------------
353    Name: Matthias Blume
354    Date: 2001/01/09 01:20:00 JST
355    Tag: blume-20010109-rcc
356    Description:
357    
358    This is a fairly big patch, flushing out a large number of pending
359    changes that I made to my development copy over the last couple of days.
360    
361    Of practical relevance at this moment is a workaround for a pickling
362    bug that Allen ran into the other day.  The cause of the bug itself is
363    still unknown and it might be hard to fix it properly, but the
364    workaround has some merits of its own (namely somewhat reducing pickling
365    overhead for certain libraries).  Therefore, I think this solution should
366    be satisfactory at this time.
367    
368    The rest of the changes (i.e., the vast majority) has to do with my
369    ongoing efforts of providing direct support for C function calls from
370    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
371    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
372    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
373    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
374    handle RCC.
375    
376    The last part (the changes to mlriscGen.sml) are still known to be
377    wrong on the x86 and not implemented on all other architectures.  But
378    the infrastructure is in place. I had to change a few functor
379    signatures in the backend to be able to route the CCalls interface
380    from MLRISC there, and I had to specialize the mltree type (on the
381    x86) to include the necessary extensions. (The extensions themselves
382    were already there and redy to go in MLRISC/x86).
383    
384    Everything should be very happy as soon as someone helps me with
385    mlriscGen.sml...
386    
387    In any case, nothing of this should matter to anyone as long as the
388    new primop is not being used (which is going to be the case unless you
389    find it where I hid it :). The rest of the compiler is completely
390    unaffected.
391    
392    ----------------------------------------------------------------------
393    Name: Matthias Blume
394    Date: 2001/01/05 00:30:00 JST
395    Tag: blume-20010105-primops
396    Description:
397    
398    Added some experimental support for work that I am doing right now.
399    These changes mostly concern added primops, but there is also a new
400    experimental C library in the runtime system (but currently not enabled
401    anywhere except on Linux/X86).
402    
403    In the course of adding primops (and playing with them), I discovered that
404    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
405    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
406    To recover, I made type information in INL_PRIM mandatory and changed
407    prim.sml as well as built-in.sml accordingly.  The InLine structure now
408    has complete, correct type information (i.e., no bottom types).
409    
410    Since all these changes mean that we need new binfiles, I also bumped the
411    version number to 110.32.1.
412    
413    ----------------------------------------------------------------------
414    Name: Matthias Blume
415    Date: 2000/12/30 22:10:00 JST
416    Tag: blume-20001230-various
417    Description:
418    
419    Added proxy libraries for MLRISC and let MLRISC libraries refer
420    to each other using path anchors.  (See CM manual for explanation.)
421    
422    Updated CM documentation.
423    
424    Fixed some bugs in CM.
425    
426    Implemented "proxy" libraries (= syntactic sugar for CM).
427    
428    Added "-quiet" option to makeml and changed runtime system accordingly.
429    
430    Added cleanup handler for exportML to reset timers and compiler stats.
431    
432    ----------------------------------------------------------------------
433    Name: Lal George
434    Date: 2000/12/22 22:22:58 EST 2000
435    Tag: Release_110_32
436    Description:
437    
438            Infinite precision used throughout MLRISC.
439            see MLRISC/mltree/machine-int.sig
440    
441    ----------------------------------------------------------------------
442    Name: Matthias Blume
443    Date: 2000/12/22 23:16:00 JST
444    Tag: blume-20001222-warn
445    Description:
446    
447    Corrected wording and formatting of some CM warning message which I
448    broke in my previous patch.
449    
450    ----------------------------------------------------------------------
451    Name: Matthias Blume
452    Date: 2000/12/22 21:20:00 JST
453    Tag: blume-20001222-anchorenv
454    Description:
455    
456    Fixed CM's handling of anchor environments in connection with CMB.make.
457    
458    ----------------------------------------------------------------------
459    Name: Matthias Blume
460    Date: 2000/12/22 13:15:00 JST
461    Tag: blume-20001222-cleanup
462    Description:
463    
464    Removed src/cm/ffi which does not (and did not) belong here.
465    
466    ----------------------------------------------------------------------
467    Name: Matthias Blume
468    Date: 2000/12/21 23:55:00 JST
469    Tag: blume-20001221-exn
470    Description:
471    
472    Probably most important: CM no longer silently swallows all exceptions
473    in the compiler.
474    Plus: some other minor CM changes.  For example, CM now reports some
475    sizes for generated binfiles (code, data, envpickle, lambdapickle).
476    
477    ----------------------------------------------------------------------
478    Name: Matthias Blume
479    Date: 2000/12/15 00:01:05 JST
480    Tag: blume-20001215-dirtool
481    Description:
482    
483    - "dir" tool added.
484    - improvements and cleanup to Tools structure
485    - documentation updates
486    
487    ----------------------------------------------------------------------
488    Name: Allen Leung
489    Date: Thu Dec 14 03:45:24 EST 2000
490    Description:
491    Tag:  leunga-20001214-int-inf
492    Description:
493    
494       In IntInf, added these standard functions, which are missing from our
495    implementation:
496    
497        andb : int * int -> int
498        xorb : int * int -> int
499        orb  : int * int -> int
500        notb : int -> int
501         <<   : int * word -> int
502        ~>>  : int * word -> int
503    
504       Not tested, I hope they are correct.
505    
506    ----------------------------------------------------------------------
507    Name: Allen Leung
508    Date: Fri Dec  8 19:23:26 EST 2000
509    Description:
510    Tag:  leunga-20001208-nowhere
511    Description:
512    
513      Slight improvements to the 'nowhere' tool to handle OR-patterns,
514    to generate better error messages etc.  Plus a brief manual.
515    
516    ----------------------------------------------------------------------
517    Name: Lal George
518    Date: 2000/12/08 09:54:02 EST 2000
519    Tag: Release_110_31
520    Description:
521    
522    - Version 110.31
523    ----------------------------------------------------------------------
524    Name: Allen Leung
525    Date: Thu Dec  7 22:01:04 EST 2000
526    Tag:  leunga-20001207-cell-monster-hack
527    Description:
528    
529    Major MLRISC internal changes.  Affect all clients.
530    Summary:
531    
532    1.  Type CELLS.cell = int is now replaced by a datatype.
533        As a result, the old regmap is now gone.  Almost all interfaces
534        in MLRISC change as a consequence.
535    
536    2.  A new brand version of machine description tool (v3.0) that generates
537        modules expecting the new interface.  The old version is removed.
538    
539    3.  The RA interface has been further abstracted into two new functors.
540        RISC_RA and X86RA.  These functors have much simpler interfaces.
541        [See also directory MLRISC/demo.]
542    
543    4.  Some other new source->source code generation tools are available:
544    
545        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
546        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
547           I use this tool to generate the peephole optimizers---with the new
548           cell type changes, peephole rules are becoming difficult to write
549           without conditional pattern matching.
550    
551    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
552        cover the entire MLRISC source tree so many things broke.
553    
554    6.  CM files have been moved to the subdirectory MLRISC/cm.
555        They are moved because there are a lot of them and they clutter up the
556        root dir.
557    
558    7.  More detailed documentation to come...
559    
560        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
561        a makeml -rebuild first.  This is because of other other
562        changes that Matthias has made (see below).
563    
564    
565    ----------------------------------------------------------------------
566    Name: Matthias Blume
567    Date: 2000/11/30 23:12:00 JST
568    Tag: blume-20001130-filereorg
569    Description:
570    
571    Some manual updates and some file reorganizations in CM.
572    
573    ----------------------------------------------------------------------
574    Name: Matthias Blume
575    Date: 2000/11/24 17:45:00 JST
576    Tag: blume-20001124-link
577    Description:
578    
579    Drastically improved link traversal code for the case that the dynamic
580    value was already loaded at bootstrap time.  As a result, CM and CMB
581    now both load blazingly fast -- even on a very slow machine.  Also,
582    memory consumption has been further reduced by this.
583    
584    Warning: The format of the PIDMAP file has changed.  THerefore, to
585    bootstrap you have to do this:
586    
587    1. Run CMB.make
588    2. Make a symbolic link for the boot directory:
589         ln -s sml.boot.ARCH-OS xxx
590    3. "Rebuild" the boot directory:
591         ./makeml -boot xxx -rebuild sml ; rm xxx
592    4. Boot normally:
593          ./makeml
594    
595    ----------------------------------------------------------------------
596    Name: Matthias Blume
597    Date: 2000/11/21 21:20:00 JST
598    Tag: blume-20001121-tools
599    Description:
600    
601    Continued hacking on autoloading problem -- with success this time.
602    Also changed tool-plugin mechanism.  See new CM manual.
603    
604    ----------------------------------------------------------------------
605    Name: Matthias Blume
606    Date: 2000/11/19 14:30:00 JST
607    Tag:  blume-20001119-autoload
608    Description:
609    
610    Some hacking to make autoloading faster.  Success for CMB, no success
611    so far for CM.  There is a reduced structure CM' that autoloads faster.
612    (This is a temporary, non-documented hack to be eliminated again when
613    the general problem is solved.)
614    
615    ----------------------------------------------------------------------
616    Name: Matthias Blume
617    Date: 2000/11/17 14:10:00 JST
618    Tag: blume-20001117-pickle-lib
619    Description:
620    
621    1. Eliminated comp-lib.cm
622    2. Made pickle-lib.cm
623    3. Eliminated all uses of intset.sml (from comp-lib.cm)
624    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
625       equivalent constructs from smlnj-lib.cm (INtHashTable).
626    5. Point 4. also goes for those uses of intmap.* in MLRISC.
627       Duplicated intmap modules thrown out.
628    6. Hunted down all duplicated SCC code and replaced it with
629       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
630    7. Rewrote Feedback module.
631    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
632       should be thrown out and equivalent modules from smlnj-lib.cm
633       should be used (IntRedBlackSet, IntListSet, ...).
634    
635    Confirmed that compiler compiles to fixpoint.
636    
637    ----------------------------------------------------------------------
638    Name: Allen Leung
639    Date: 2000/11/10 18:00:00
640    Tag: leunga-20001110-new-x86-fp
641    
642    A new x86 floating point code generator has been added.
643    By default this is turned off.  To turn this on, do:
644    
645        CM.autoload "$smlnj/compiler.cm";
646        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
647    
648    Changes:
649    
650    1.  Changed FTAN to FPTAN so that the assembly output is correct.
651    2.  Changed the extension callback for FTANGENT to generate:
652    
653              fptan
654              fstp  %st(0)
655        instead of
656              fptan
657              fstpl ftempmem
658    
659    3.  Numerous assembly fixes for x86.
660    
661    5.  Cleaned up the machine code output module x86/x86MC.sml and added
662        support for a whole bunch of instructions and addressing modes:
663    
664          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
665          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
666          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
667          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
668          fxch %st(n)
669          fld %st(n)
670          fst %st(n)
671          fst mem
672          fstp %st(n)
673          fucom %st(n)
674          fucomp %st(n)
675    
676        All these are now generated when the fast fp mode is turned on.
677    
678    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
679    
680    ----------------------------------------------------------------------
681    Name: Matthias Blume
682    Date: 2000/11/09 11:20:00 JST
683    Tag: blume-20001109-scc
684    Description:
685    
686    Eliminated some code duplication:
687    
688    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
689       (Otherwise the functor is useless.)
690    2. Used GraphSCCFn where SCCUtilFun was used previously.
691    3. Got rid of SCCUtilFun (in comp-lib.cm).
692    
693    ----------------------------------------------------------------------
694    Name: Lal George
695    Date: 2000/11/06 09:02:21 EST 2000
696    Tag: Release_110_30
697    Description:
698    
699    - Version 110.30
700    ----------------------------------------------------------------------
701    Name: Matthias Blume
702    Date: 2000/11/04 14:45:00
703    Tag: blume-20001104-mlbuild
704    Description:
705    
706    - Made ml-build faster on startup.
707    - Documentation fixes.
708    
709    ----------------------------------------------------------------------
710    Name: Matthias Blume
711    Date: 2000/11/02 17:00:00 JST
712    Tag: blume-20001102-condcomp
713    Description:
714    
715    - Small tweaks to pickler -- new BOOTFILES!
716    - Version bumped to 110.29.2.
717    - Added conditional compilation facility to init.cmi (see comment there).
718    ----------------------------------------------------------------------
719    Name: Allen Leung
720    Date: 2000/10/23 19:31:00
721    Tag: leunga-20001023-demo-ra
722    
723    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
724    2. Test programs for the graph library updated
725    3. Some new MLRISC demo programs added
726    
727    ----------------------------------------------------------------------
728    Name: Matthias Blume
729    Date: 2000/08/31 22:15:00 JST
730    Tag: blume-20001017-errmsg
731    Description:
732    
733    More error message grief: Where there used to be no messages, there
734    now were some that had bogus error regions.  Fixed.
735    
736    ----------------------------------------------------------------------
737    Name: Matthias Blume
738    Date: 2000/08/31 17:30:00 JST
739    Tag: blume-20001017-v110p29p1
740    Description:
741    
742    I made a version 110.29.1 with new bootfiles.
743    
744    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
745              CM documentation changes and a small bugfix in CM's error reporting.
746    
747    ----------------------------------------------------------------------
748    Name: Lal George
749    Date: 2000/09/27 14:42:35 EDT
750    Tag: george-20000927-nodestatus
751    Description:
752    
753    Changed the type of the nodestatus, so that:
754    
755            SPILLED(~1)             is now SPILLED
756            SPILLED(m) where m>=0   is now MEMREG(m)
757            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
758    
759    ----------------------------------------------------------------------
760    Name: Matthias Blume
761    Date: 2000/09/07 14:45:00 JST
762    Tag: blume-20000907-cmerrmsg
763    Description:
764    
765    Small tweak to CM to avoid getting ML syntax error messages twice.
766    
767    ----------------------------------------------------------------------
768    Name: Matthias Blume
769    Date: 2000/08/31 18:00:00 JST
770    Tag: blume-20000831-cvsbootfiles
771    Description:
772    
773    New URL for boot files (because the 110.29 files on the BL server do
774    now work correctly with my updated install scripts for yacc and lex).
775    
776    ----------------------------------------------------------------------
777    Name: Matthias Blume
778    Date: 2000/08/08 12:33:00 JST
779    Tag: blume-20000808-manual
780    Description:
781    
782    Tiny update to CM manual.
783    
784    ----------------------------------------------------------------------
785    Name: Allen Leung
786    Date: 2000/08/7 19:31:00
787    Tag: leunga-20000807-a-whole-bunch-of-stuff
788    
789      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
790    and Mobdy.
791    
792    1.  x86
793    
794       a.  Fixes to peephole module by John and Dan.
795       b.  Assembly fix to SETcc by Allen.
796       c.  Fix to c-call by John.
797       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
798       e.  Instruction selection optimization to SETcc as suggested by John.
799    
800           For example,
801    
802            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
803    
804           should generate:
805    
806            MOVL a, x
807            SUBL b, x
808            SHRL 31, x
809    
810    2.  IR stuff
811    
812         A bunch of new DJ-graph related algorithms added.  These
813         speed up SSA construction.
814    
815    3.  SSA + Scheduling
816    
817         Added code for SSA and scheduling to the repository
818    
819    ----------------------------------------------------------------------
820    Name: Lal George
821    Date: 2000/07/27 11:53:14 EDT
822    
823    Tag: lal-20000727-linux-ppc
824    Description:
825    
826     Made changes to support Linux PPC.
827     p.s. I have confirmation that the 110.29 boot files work fine.
828    
829    ----------------------------------------------------------------------
830    Name: Matthias Blume
831    Date: 2000/07/27 17:40:00 JST
832    Tag: blume-20000727-scripts
833    Description:
834    
835    !!!! WARNING !!!!
836    You must recompile the runtime system!
837    !!!! WARNING !!!!
838    
839    This is basically another round of script-enhancements:
840    
841    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
842       and undefine CM preprocessor symbols.
843    
844    2. ml-build avoids generating a new heap image if it finds that the
845       existing one is still ok.  (The condition is that no ML file had to
846       be recompiled and all ML files are found to be older that the heap
847       file.)
848    
849       To make this work smoothly, I also hacked the runtime system as
850       well as SMLofNJ.SysInfo to get access to the heap image suffix
851       (.sparc-solaris, ...) that is currently being used.
852    
853       Moreover, the signature of CM.mk_standalone has changed.  See the
854       CM manual.
855    
856    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
857       CM manual for details.)
858    
859    4. More CM manual updates:
860        - all of the above has been documented.
861        - there is now a section describing the (CM-related) command line
862          arguments that are accepted by the "sml" command
863    
864    ----------------------------------------------------------------------
865    Name: Matthias Blume
866    Date: 2000/07/25 16:20:00 JST
867    Tag: blume-20000725-makedepend
868    Description:
869    
870    Added a script called ml-makedepend.  This can be used in makefiles
871    for Unix' make in a way very similar to the "makedepend" command for
872    C.
873    
874    The script internally uses function CM.sources.
875    
876    Synopsis:
877    
878        ml-makedepend [-f makefile] cmfile targetname
879    
880    The default for the makefile is "makefile" (or "Makefile" should
881    "makefile" not exist).
882    
883    ml-makedepend adds a cmfile/targetname-specific section to this
884    makefile (after removing the previous version of this section).  The
885    section contains a single dependency specification with targetname on
886    the LHS (targetname is an arbitrary name), and a list of files derived
887    from the cmfile on the RHS.  Some of the files on the RHS are
888    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
889    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
890    path names.  The makefile writer is responsible for making sure that
891    these variables have correct at the time "make" is invoked.
892    
893    ----------------------------------------------------------------------
894    Name: Matthias Blume
895    Date: 2000/07/22 23:30:00 JST
896    Tag: blume-20000722-urlupdate
897    Description:
898    
899    Changed BOOT and config/srcarchiveurl to point to BL server:
900    
901        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
902    
903    ----------------------------------------------------------------------
904    Name: Matthias Blume
905    Date: 2000/07/18 18:00:00 JST
906    Tag: blume-20000718-Version_110_29
907    Description:
908    
909    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
910    
911    2. Updated config/version to 110.29
912    
913    3. Updated config/srcarchiveurl
914    
915    3. New boot files!
916       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
917    
918    ----------------------------------------------------------------------
919    Name: Matthias Blume
920    Date: 2000/07/11 13:58:00 JST
921    Tag: blume-20000711-doctypo
922    Description:
923    
924    Fixed a few typos in CM manual.
925    
926    ----------------------------------------------------------------------
927    Name: Allen Leung
928    Date: 2000/06/15 00:38:00
929    Tag: leunga-20000704-sparc-x86
930    
931    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
932    2. fix to x86 RET bug [found by Dan Grossman]
933    3. sparc assembly bug fix for ticc instructions [found by Fermin]
934    
935       Affects c-- and moby only
936    
937    ----------------------------------------------------------------------
938    Name: Matthias Blume
939    Date: 2000/07/04 15:26:00
940    Tag: blume-20000704-trigger
941    Description:
942    
943    1. Improvements to CM manual.
944    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
945       of getting a back-trace.  The function, when called, raises an
946       internal exception which explicitly carries the full back-trace history,
947       so it is unaffected by any intervening handle-raise pairs ("trivial"
948       or not).  The interactive loop will print that history once it arrives
949       at top level.
950       Short of having all exceptions implicitly carry the full history, the
951       recommended way of using this facility is:
952         - compile your program with instrumentation "on"
953         - run it, when it raises an exception, look at the history
954         - if the history is "cut off" because of some handler, go and modify
955           your program so that it explicitly calls BTrace.trigger
956         - recompile (still instrumented), and rerun; look at the full history
957    
958    ----------------------------------------------------------------------
959    Name: Matthias Blume
960    Date: 2000/07/03 15:36:00 JST
961    Tag: blume-20000702-manual
962    Description:
963    
964    Small corrections and updates to CM manual.
965    
966    ----------------------------------------------------------------------
967    Name: Matthias Blume
968    Date: 2000/06/29 16:04:00 JST
969    Tag: blume-20000629-yacctool
970    Description:
971    
972    Changes:
973    
974    1. Class "mlyacc" now takes separate arguments to pass options to
975       generated .sml- and .sig-files independently.
976    2. Corresponding CM manual updates.
977    3. BTrace module now also reports call sites.  (However, for loop clusters
978       it only shows from where the cluster was entered.)  There are associated
979       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
980    
981    ----------------------------------------------------------------------
982    Name: Matthias Blume
983    Date: 2000/06/27 16:51:00 JST
984    Tag: blume-20000627-noweb
985    Description:
986    
987    Changes:
988    
989     1. Implemented "subdir" and "witness" options for noweb tool.
990        This caused some slight internal changes in CM's tool implementation.
991     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
992        some remaining issues from earlier path anchor changes.
993     3. Updated CM manual accordingly.
994    
995     4. Changed implementation of back-tracing so that I now consider it
996        ready for prime-time.
997    
998        In particular, you don't have to explicitly trigger the back-trace
999        anymore.  Instead, if you are running BTrace-instrumented code and
1000        there is an uncaught exception (regardless of whether or not it was
1001        raised in instrumented code), the top-level evalloop will print
1002        the back-trace.
1003    
1004        Features:
1005    
1006          - Instrumented and uninstrumented code work together seemlessly.
1007            (Of course, uninstrumented code is never mentioned in actual
1008             back-traces.)
1009    
1010          - Asymptotic time- and space-complexity of instrumented code is
1011            equal to that of uninstrumented code.  (This means that
1012            tail-recursion is preserved by the instrumentation phase.)
1013    
1014          - Modules whose code has been instrumented in different sessions
1015            work together without problem.
1016    
1017          - There is no penalty whatsoever on uninstrumented code.
1018    
1019          - There is no penalty on "raise" expressions, even in
1020            instrumented code.
1021    
1022        A potential bug (or perhaps it is a feature, too):
1023    
1024          A back-trace reaches no further than the outermost instrumented
1025          non-trivial "raise".  Here, a "trivial" raise is one that is the
1026          sole RHS of a "handle" rule.  Thus, back-traces reach trough
1027    
1028               <exp> handle e => raise e
1029    
1030          and even
1031    
1032               <exp> handle Foo => raise Bar
1033    
1034          and, of course, through
1035    
1036               <exp> handle Foo => ...
1037    
1038         if the exception was not Foo.
1039    
1040         Back-traces always reach right through any un-instrumented code
1041         including any of its "handle" expressions, trivial or not.
1042    
1043       To try this out, do the following:
1044    
1045         - Erase all existing binfiles for your program.
1046           (You may keep binfiles for those modules where you think you
1047            definitely don't need back-tracing.)
1048         - Turn on back-trace instrumentation:
1049              SMLofNJ.Internals.BTrace.mode (SOME true);
1050         - Recompile your program.  (I.e., run "CM.make" or "use".)
1051         - You may now turn instrumentation off again (if you want):
1052              SMLofNJ.Internals.BTrace.mode (SOME false);
1053         - Run your program as usual.  If it raises an exception that
1054           reaches the interactive toplevel, then a back-trace will
1055           automatically be printed.  After that, the toplevel loop
1056           will print the exception history as usual.
1057    
1058    ----------------------------------------------------------------------
1059    Name: Matthias Blume
1060    Date: 2000/06/26 09:56:46 JST
1061    Tag: blume-20000626-setup
1062    Description:
1063    
1064    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1065          ML code before and after compiling a file (e.g., to set compiler
1066          flags)
1067    
1068    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1069              - associated changes to btrace.sml (BTrace instrumentation pass)
1070              - cleaner implementation of btimp.sml (BTrace tracing and report
1071                module)
1072    
1073    CM manual: * new path encoding documented
1074               * description of setup-parameter to "sml" added
1075    
1076    The biggest user-visible change to back-tracing is that it is no
1077    longer necessary to compile all traced modules within the same
1078    session.  (This was a real limitation.)
1079    
1080    ----------------------------------------------------------------------
1081    Name: Matthias Blume
1082    Date: 2000/06/24 12:40:00 JST
1083    Tag: blume-20000624-startup
1084    Description:
1085    
1086    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1087    bit too often -- to put it mildly. :)
1088    
1089    ----------------------------------------------------------------------
1090    Name: Matthias Blume
1091    Date: 2000/06/23 18:20:00 JST
1092    Tag: blume-20000623-btrace
1093    Description:
1094    
1095    This updates adds a backtrace facility to aid programmers in debugging
1096    their programs.  This involves the following changes:
1097    
1098    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1099       keeping track of the current call stack.  When programs are compiled
1100       in a special mode, the compiler will insert calls to these hooks
1101       into the user program.
1102       "Hook" means that it is possible for different implementations of
1103       back-tracing to register themselves (at different times).
1104    
1105    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1106       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1107       It can be turned on using this call:
1108         SMLofNJ.Internals.BTrace.mode (SOME true);
1109       Turning it off again:
1110         SMLofNJ.Internals.BTrace.mode (SOME false);
1111       Querying the current status:
1112         SMLofNJ.Internals.BTrace.mode NONE;
1113       Annotated programs are about twice as big as normal ones, and they
1114       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1115       where all hooks do nothing).  The slowdown with a plugin that is
1116       actually useful (such as the one supplied by default) is even greater,
1117       but in the case of the default plugin it is still only an constant
1118       factor (amortized).
1119    
1120    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1121       with a sub-structure BTrace for controlling back-tracing.  In particular,
1122       the above-mentioned function "mode" controls whether the annotation
1123       phase is invoked by the compiler.  Another important function is
1124       "trigger": when called it aborts the current execution and causes
1125       the top-level loop to print a full back-trace.
1126    
1127    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1128       for back-tracing.  It keeps track of the dynamic call stack and in
1129       addition to that it keeps a partial history at each "level" of that
1130       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1131       d and b (at separate times, dynamically), then the report will show:
1132    
1133       GOTO   d
1134             /c
1135       GOTO  \b
1136       CALL   a
1137    
1138       This shows that there was an initial non-tail call of a, then a
1139       tail-call to b or c, looping behavior in a cluster of functions that
1140       consist of b and c, and then a goto from that cluster (i.e., either from
1141       b or from c) to d.
1142    
1143       Note that (depending on the user program) the amount of information
1144       that the back-trace module has to keep track of at each level is bounded
1145       by a constant.  Thus, the whole implementation has the same asymptotical
1146       complexity as the original program (both in space and in time).
1147    
1148    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1149       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1150       which is raised by the "trigger" function mentioned above.
1151    
1152    Notes on usage:
1153    
1154    - Annotated code works well together with unannotated code:
1155    Unannotated calls simply do not show up at all in the backtrace.
1156    
1157    - It is not a good idea to let modules that were annotated during
1158    different sessions run at the same time.  This is because the compiler
1159    chooses small integers to identify individual functions, and there
1160    will be clashes if different modules were compiled in separate sessions.
1161    (Nothing will crash, and you will even be told about the clashes, but
1162    back-trace information will in general not be useful.)
1163    
1164    - Back-tracing can be confused by callcc and capture.
1165    
1166    - The only way of getting a back-trace right now is to explicitly
1167    invoke the "trigger" function from your user program.  Eventually, we
1168    should make every exception carry back-trace information (if
1169    available).  But since this creates more overhead at "raise"-time
1170    (similar to the current exnHistory overhead), I have not yet
1171    implemented this.  (The implementation will be rather easy.)  With
1172    exceptions carrying back-trace information, this facility will be even
1173    more useful because users don't need to modify their programs...
1174    
1175    - While it is possible to compile the compiler with back-trace
1176    annotations turned on (I did it to get some confidence in
1177    correctness), you must make absolutely sure that core.sml and
1178    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1179    be compiled with annotation because there is no core access yet, but
1180    if you compile btimp.sml with annotation, then the system will go into
1181    an infinite recursion and crash.)
1182    Since CM currently does not know about BTrace, the only way to turn
1183    annotations on and off for different modules of the compiler is to
1184    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1185    this is awkward and clumsy.
1186    
1187    Sample sessions:
1188    
1189    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1190    - SMLofNJ.Internals.BTrace.mode (SOME true);
1191    [autoloading]
1192    [autoloading done]
1193    val it = false : bool
1194    - structure X = struct
1195    -     fun main n = let
1196    -         fun a (x, 0) = d x
1197    -           | a (x, n) = b (x, n - 1)
1198    -         and b (x, n) = c (x, n)
1199    -         and c (x, n) = a (x, n)
1200    -         and d x = e (x, 3)
1201    -         and e (x, 0) = f x
1202    -           | e (x, n) = e (x, n - 1)
1203    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1204    -           | f n = n * g (n - 1)
1205    -         and g n = a (n, 3)
1206    -     in
1207    -         f n
1208    -     end
1209    - end;
1210    structure X : sig val main : int -> int end
1211    - X.main 3;
1212    *** BACK-TRACE ***
1213    GOTO   stdIn:4.2-13.20: X.main[2].f
1214    GOTO-( stdIn:4.2-13.20: X.main[2].e
1215    GOTO   stdIn:4.2-13.20: X.main[2].d
1216         / stdIn:4.2-13.20: X.main[2].a
1217         | stdIn:4.2-13.20: X.main[2].b
1218    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1219    CALL   stdIn:4.2-13.20: X.main[2].g
1220    GOTO   stdIn:4.2-13.20: X.main[2].f
1221    GOTO-( stdIn:4.2-13.20: X.main[2].e
1222    GOTO   stdIn:4.2-13.20: X.main[2].d
1223         / stdIn:4.2-13.20: X.main[2].a
1224         | stdIn:4.2-13.20: X.main[2].b
1225    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1226    CALL   stdIn:4.2-13.20: X.main[2].g
1227    GOTO   stdIn:4.2-13.20: X.main[2].f
1228    GOTO-( stdIn:4.2-13.20: X.main[2].e
1229    GOTO   stdIn:4.2-13.20: X.main[2].d
1230         / stdIn:4.2-13.20: X.main[2].a
1231         | stdIn:4.2-13.20: X.main[2].b
1232    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1233    CALL   stdIn:4.2-13.20: X.main[2].g
1234    GOTO   stdIn:4.2-13.20: X.main[2].f
1235    CALL   stdIn:2.15-17.4: X.main[2]
1236    -
1237    
1238    (Note that because of a FLINt bug the above code currently does not
1239    compile without BTrace turned on.)
1240    
1241    Here is another example, using my modified Tiger compiler:
1242    
1243    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1244    - SMLofNJ.Internals.BTrace.mode (SOME true);
1245    [autoloading]
1246    [autoloading done]
1247    val it = false : bool
1248    - CM.make "sources.cm";
1249    [autoloading]
1250    ...
1251    [autoloading done]
1252    [scanning sources.cm]
1253    [parsing (sources.cm):parse.sml]
1254    [creating directory CM/SKEL ...]
1255    [parsing (sources.cm):tiger.lex.sml]
1256    ...
1257    [wrote CM/sparc-unix/semant.sml]
1258    [compiling (sources.cm):main.sml]
1259    [wrote CM/sparc-unix/main.sml]
1260    [New bindings added.]
1261    val it = true : bool
1262    - Main.compile ("../testcases/merge.tig", "foo.out");
1263    *** BACK-TRACE ***
1264    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
1265    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1266    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1267    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1268    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1269    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1270    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
1271         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
1272    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
1273    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1274    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
1275    CALL   main.sml:1.18-118.4: Main.compile[2]
1276    -
1277    
1278    ----------------------------------------------------------------------
1279    Name: Matthias Blumen
1280    Date: 2000/06/21 18:00:00 JST
1281    Tag: blume-20000621-manual
1282    Description:
1283    
1284    CM manual update: Path environments documented.
1285    
1286    ----------------------------------------------------------------------
1287    Name: Matthias Blume
1288    Date: 2000/06/19 13:40:00
1289    Tag: blume-20000619-manual
1290    Description:
1291    
1292    CM manual and system/README update.  This only covers the fact that
1293    there are no more implicit anchors.  (Path environments and the "bind"
1294    option to "cm" have yet to be documented.)
1295    
1296    ----------------------------------------------------------------------
1297    Name: Matthias Blume
1298    Date: 2000/06/19 11:05:00 JST
1299    Tag: blume-20000619-chdir-bugfix
1300    Description:
1301    
1302    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1303    
1304    ----------------------------------------------------------------------
1305    Name: Matthias Blume
1306    Date: 2000/06/18 22:00:10 JST
1307    Tag: blume-20000618-implicit-anchors-really-gone
1308    Description:
1309    
1310    I updates the previous HISTORY entry where I forgot to mention that
1311    implicit anchors are no longer with us.
1312    
1313    The current update also gets rid of the (now useless) controller
1314    CM.Control.implicit_anchors.
1315    
1316    ----------------------------------------------------------------------
1317    Name: Matthias Blume
1318    Date: 2000/06/16 17:30:00 JST
1319    Tag: blume-20000616-anchorenv
1320    Description:
1321    
1322    This patch implements the long anticipated (just kidding :) "anchor
1323    environment" mechanism.  In the course of doing this, I also
1324    re-implemented CM's internal "SrcPath" module from scratch.  The new
1325    one should be more robust in certain boundary cases.  In any case, it
1326    is a lot cleaner than its predecessor (IMHO).
1327    
1328    This time, although there is yet another boot file format change, I
1329    kept the unpickler backward-compatible.  As a result, no new bootfiles
1330    are necessary and bootstrapping is straightforward.  (You cannot read
1331    new bootfiles into an old system, but the other way around is no
1332    problem.)
1333    
1334    Visible changes:
1335    
1336    ** 0. Implicit path anchors (without the leading $-symbol) are no
1337    longer recognized at all. This means that such path names are not
1338    illegal either.  For example, the name basis.cm simply refers to a
1339    local file called "basis.cm" (i.e, the name is an ordinary path
1340    relative to .cm-files directory).  Or, to put it differently, only
1341    names that start with $ are anchored paths.
1342    
1343    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1344    vanished.
1345    
1346    John (Reppy) had critizised this as soon as I originally proposed and
1347    implemented it, but at that time I did not really deeply believe
1348    him. :) Now I came full-circle because I need the $<singlearc> syntax
1349    in another place where it cannot be seen as an abbreviation for
1350    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1351    seems to mean (i.e., it "expands" into the corresponding anchor
1352    value).
1353    
1354    However, when paths are used as members in CM description files, it
1355    continues to be true that there must be at least another arc after the
1356    anchor.  This is now enforced separately during semantic analysis
1357    (i.e., from a lexical/syntactical point of view, the notation is ok.)
1358    
1359    ** 2. The "cm" class now accepts an option "bind".  The option's value
1360    is a sub-option list of precisely two items -- one labeled "anchor"
1361    and the other one labeled "value".  As you might expect, "anchor" is
1362    used to specify an anchor name to be bound, and "value" specifies what
1363    the anchor is being bound to.
1364    
1365    The value must be a directory name and can be given in either standard
1366    syntax (including the possibility that it is itself an anchored path)
1367    or native syntax.
1368    
1369    Examples:
1370    
1371       foo.cm (bind:(anchor:bar value:$mystuff/bar))
1372       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1373    
1374    and so on.
1375    
1376    The meaning of this is that the .cm-file will be processed with an
1377    augmented anchor environment where the given anchor(s) is/are bound to
1378    the given values(s).
1379    
1380    The rationale for having this feature is this: Suppose you are trying
1381    to use two different (already stable) libraries a.cm and b.cm (that
1382    you perhaps didn't write yourself).  Further, suppose each of these
1383    two libraries internally uses its own auxiliary library $aux/lib.cm.
1384    Normally you would now have a problem because the anchor "lib" can not
1385    be bound to more than one value globally.  Therefore, the project that
1386    uses both a.cm and b.cm must locally redirect the anchor to some other
1387    place:
1388    
1389       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1390       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1391    
1392    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1393    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
1394    
1395    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1396    you when you do so at the time of CM.stabilize).  Therefore, you can
1397    also use an anchored path as the value:
1398    
1399      a.cm (bind:(anchor:lib value:$a-lib))
1400      b.cm (bind:(anchor:lib value:$b-lib))
1401    
1402    Now you can globally configure (using the usual CM.Anchor.anchor or
1403    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1404    itself is always locally bound, setting it globally is no longer
1405    meaningful or necessary (but it does not hurt either).  In fact, "lib"
1406    can still be used as a global anchor for separate purposes.  As a
1407    matter of fact, one can locally define "lib" in terms of a global
1408    "lib":
1409    
1410      a.cm (bind:(anchor:lib value:$lib/a))
1411      b.cm (bind:(anchor:lib value:$lib/b))
1412    
1413    ** 3: The encoding of path names has changed.  This affects the way
1414    path names are shown in CM's progress report and also the internal
1415    protocol encoding used for parallel make.
1416    
1417    The encoding now uses one or more ':'-separated segments.  Each
1418    segments corresponds to a file that has been specified relative to the
1419    file given by its preceding segment.  The first segment is either
1420    relative to the CWD, absolute, or anchored.  Each segment itself is
1421    basically a Unix pathname; all segments but the first are relative.
1422    
1423    Example:
1424    
1425       $foo/bar/baz.cm:a/b/c.sml
1426    
1427    This path denotes the file bar/a/b/c.sml relative to the directory
1428    denoted by anchor "foo".  Notice that the encoding also includes
1429    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1430    paths are resolved relative to the .cm-files directory, so baz.cm must
1431    be ignored to get the "real" pathname.
1432    
1433    To make this fact more obvious, CM puts the names of such "virtual
1434    arcs" into parentheses when they appear in progress reports. (No
1435    parentheses will appear in the internal protocol encoding.)  Thus,
1436    what you really see is:
1437    
1438      $foo/bar/(baz.cm):a/b/c.sml
1439    
1440    I find this notation to be much more informative than before.
1441    
1442    Another new feature of the encoding is that special characters
1443    including parentheses, colons, (back)slashes, and white space are
1444    written as \ddd (where ddd is the decimal encoding of the character).
1445    
1446    *** The CM manual still needs to be updated.
1447    
1448    ----------------------------------------------------------------------
1449    Name: Allen Leung
1450    Date: 2000/06/15 00:38:00
1451    Tag: leunga-20000615-x86-peephole
1452    
1453    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1454    
1455    ----------------------------------------------------------------------
1456    Name: Matthias Blume
1457    Date: 2000/06/12 11:40:00
1458    Tag: blume-20000612-parmakefix
1459    Description:
1460    
1461    More cleanup after changing the file naming scheme: This time I
1462    repaired the parallel make mechanism for CMB.make which I broke earlier.
1463    
1464    ----------------------------------------------------------------------
1465    Name: Allen Leung
1466    Date: 2000/06/09 01:25:00
1467    Tag: leunga-20000609-various
1468    
1469    None of these things should affect normal SML/NJ operations
1470    
1471    1. Peephole improvements provided by Fermin (c--)
1472    2. New annotation DEFUSE for adding extra dependence (moby)
1473    3. New X86 LOCK instructions (moby)
1474    4. New machine description language for reservation tables (scheduling)
1475    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1476       trees etc.)
1477    6. I've changed the CM files so that they can work with versions
1478       110.0.6, 110.25 and 110.28
1479    
1480    ----------------------------------------------------------------------
1481    Name: Matthias Blume
1482    Date: 2000/06/09 12:40:00
1483    Tag: blume-20000609-log
1484    Description:
1485    
1486    - Removed all(?) remaining RCS Log entries from sources.
1487    
1488    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1489      anchored paths).
1490    
1491    ----------------------------------------------------------------------
1492    Name: Matthias Blume
1493    Date: 2000/06/07 17:00:00 JST
1494    Tag: blume-20000607-no-implicit-anchors
1495    Description:
1496    
1497    1. This update changes the default setting for
1498    CM.Control.implicit_anchors from true to false.  This means that
1499    implicit anchors are no longer permitted by default.  I also tried to
1500    make sure that nothing else still relies on implicit anchors.
1501    (This is the next step on the schedule towards a CM that does not even
1502    have the notion of implicit anchors anymore.)
1503    
1504    2. More CM manual updates.
1505    
1506    3. I managed to track down and fix the pickling bug I mentioned last
1507    time.  Because of the previously existing workaround, this entails no
1508    immediate practical changes.
1509    
1510    ----------------------------------------------------------------------
1511    Name: Matthias Blume
1512    Date: 2000/06/06 11:15:00 JST
1513    Tag: blume-20000606-lazierpickle
1514    Description:
1515    
1516    !!!! NEW BOOT FILES !!!!
1517    
1518    * The main purpose of this update is to make library pickles lazier in
1519    order to reduce the initial space penalty for autoloading a library.
1520    As a result, it is now possible to have $smlnj/compiler.cm
1521    pre-registered.  This should take care of the many complaints or
1522    inquiries about missing structure Compiler.  This required changes to
1523    CM's internal data structures and small tweaks to some algorithms.
1524    
1525    As a neat additional effect, it is no longer necessary (for the sake
1526    of lean heap image files) to distinguish between a "minimal" CM and a
1527    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1528    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1529    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1530    
1531    To make the life of compiler-hackers easier, "makeml" now also
1532    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1533    words, after you bootstrap a new sml for the first time, you will not
1534    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1535    around you will still have to do it, though.)
1536    
1537    * A second change consists of major updates to the CM manual.  There
1538    are now several appendices with summary information and also a full
1539    specification of the CM description file syntax.
1540    
1541    * In directory src/system I added the script "allcross".  This script
1542    invokes sml and cross-compiles the compiler for all supported
1543    architectures.  (Useful when providing a new set of boot files.)
1544    
1545    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1546    added a small tweak to pickle-util.sml to work around this problem,
1547    but it is not a proper fix yet.  I will investigate further.  (The
1548    effect of the bug was an inflation of library pickle size.)
1549    
1550    * Version number increased to 110.28.1 (to avoid compatibility problems).
1551    
1552    ----------------------------------------------------------------------
1553    Name: Allen Leung
1554    Date: 2000/05/25 17:28 EDT
1555    Tag: leunga-20000525-ra
1556    Description:
1557    
1558      Fixed a bug in freezing phase of the register allocator.
1559    
1560    ----------------------------------------------------------------------
1561    Name: Allen Leung
1562    Date: 2000/05/15 22:53 EDT
1563    Tag: leunga-20000515-alpha-x86-ra
1564    Description:
1565    
1566      1. Alpha
1567    
1568          Slight cleanup.  Removed the instruction SGNXL
1569    
1570      2. X86
1571    
1572          Added the following instructions to the instruction set:
1573    
1574            ROLx, RORx,
1575            BTx, BTSx, BTLx, BTRx,
1576            XCHGx, and variants with the LOCK prefix
1577    
1578      3. Register Allocation
1579    
1580          The module ra-rewrite-with-renaming has been improved.
1581    
1582      These have no effect on SML/NJ.
1583    
1584    ----------------------------------------------------------------------
1585    Name: Matthias Blume
1586    Date: 2000/05/15 16:20:00 JST
1587    Tag: blume-20000515-lightrebuild
1588    Description:
1589    
1590    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1591       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1592       defined.  In effect, the command will not build any cross-compiler
1593       backends and therefore finish more quickly.
1594    
1595       The "fixpt" script also takes a "-light" switch to be able to use
1596       this new facility while compiling for a fixpoint.
1597    
1598    2. I replaced all mentions of anchored paths in group owner specifications
1599       with simple relative paths (usually starting with "..").
1600       The rationale is that a library's internal workings should not be
1601       compromised by the lack of some anchor.  (An anchor is necessary
1602       for someone who wants to refer to the library by an anchored path,
1603       but it should not be necessary to build the same library in the first
1604       place.)
1605    
1606    3. I changed the way CM's tool mechanism determines the shell command
1607       string used for things like ml-yacc etc. so that it does not break
1608       when CM.Control.implicit_anchors is turned off.
1609    
1610    ----------------------------------------------------------------------
1611    Name: Matthias Blume
1612    Date: 2000/05/12 18:20:00 JST
1613    Tag: blume-20000512-ml-build
1614    Description:
1615    
1616    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1617    getting installed properly (by config/install.sh).
1618    
1619    ----------------------------------------------------------------------
1620    Name: Matthias Blume
1621    Date: 2000/05/12 17:30:00 JST
1622    Tag: blume-20000512-anchors
1623    Description:
1624    
1625    !!! NEW BOOT FILES !!!
1626    
1627    This change is in preparation of fading out support for "implicitly
1628    anchored path names".  I went through all sources and used the
1629    explicit (and relatively new) $-notation.  See system/README and the
1630    CM manual for more info on this.
1631    
1632    I also modified the anchoring scheme for some things such as "smlnj",
1633    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1634    anchors are more expressive: anchor name and first arc do not have to
1635    coincide.  This entails the following user-visible change:
1636    
1637    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1638    particular, when you fire up sml with a command-line argument, say,
1639    e.g.:
1640    
1641       sml '$smlnj/cmb.cm'
1642    
1643    At the ML toplevel prompt:
1644    
1645       CM.autoload "$smlnj/cmb.cm";
1646    
1647    There is also a new controller in CM.Control that can be used to turn
1648    off all remaining support for implicit anchors by saying:
1649    
1650        CM.autoload "$smlnj/
1651        #set CM.Control.implicit_anchors false;
1652    
1653    This causes CM to reject implicitly anchored paths.  This is (for the
1654    time being) less permissive than the "final" version where there will
1655    be no more such implicit anchors and relative paths will be just that:
1656    relative.
1657    
1658    The next step (version after next version?) will be to make the
1659    default for CM.Control.implicit_anchors false.  After the dust has
1660    settled, I can then produce the "final" version of this...
1661    
1662    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1663    
1664    ----------------------------------------------------------------------
1665    Name: Matthias Blume
1666    Date: 2000/05/11 16:30:00 JST
1667    Tag: blume-20000511-sources
1668    Description:
1669    
1670    The main change is that I added function CM.sources as a generalized
1671    version of the earlier CM.makedepend.  This entails the following
1672    additional changes:
1673    
1674      - CM.makedepend has been dropped.
1675    
1676      - CM manual has been updated.
1677    
1678      - TOOLS signature and API have been changed.
1679    
1680    ----------------------------------------------------------------------
1681    Name: Allen Leung
1682    Date: 2000/05/10 21:17 EDT
1683    Tag: leunga-20000510-moby-c--ssa
1684    Description:
1685    
1686      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1687    None of these affect SML/NJ.
1688    
1689    1. Register Allocation
1690    
1691        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1692           This module tries to remove local (i.e. basic block level) redundancies
1693           during spilling.
1694    
1695        b. A new framework for performing region based register allocation.
1696           Not yet entirely functional.
1697    
1698    2. X86
1699    
1700       a. DefUse for POP was missing the stack pointer [found by Lal]
1701       b. Reload for CALL was incorrect in X86Spill [found by John]
1702       c. Various fixes in X86Spill so that it can be used correctly for
1703          the new spilling module.
1704    
1705    3. SSA/IR
1706    
1707       a. New module ir/dj-dataflow.sml implements elimination based
1708          data flow analysis.
1709    
1710    4. MLRiscGen
1711    
1712       a. Fix for gc type annotation
1713    
1714    5. MDGen
1715    
1716       Various fixes for machine description -> ml code translation.  For ssa
1717       only.
1718    
1719    ----------------------------------------------------------------------
1720    Name: Allen Leung
1721    Date: 2000/05/08 22:17 EDT
1722    Tag: leunga-20000508-labexp
1723    Description:
1724    
1725      Fermin has found a few assembly problems with constant expressions
1726      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1727      which choke on dumb assemblers.  This is his fix.
1728    
1729    ----------------------------------------------------------------------
1730    Name: Dave MacQueen
1731    Date: 2000/04/09 14:00 EDT
1732    Tag: dbm-20000502-Version_110_28
1733    Description:
1734    
1735    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1736    
1737    2. Updated config/version to 110.28
1738    
1739    3. Updated config/srcarchiveurl
1740    
1741    3. New boot files!
1742       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1743    
1744    ----------------------------------------------------------------------
1745    Name: Matthias Blume
1746    Date: 2000/05/01 19:05:00 JST
1747    Tag: blume-20000501-noweb
1748    Description:
1749    
1750    A new noweb tool has been added.  The existing system is entirely
1751    unaffected by this, but some CM users have asked for renewed noweb
1752    support.  Everything is documented in the CM manual.
1753    
1754    New (plugin) libraries:
1755    
1756       noweb-tool.cm
1757       nw-ext.cm
1758    
1759    ----------------------------------------------------------------------
1760    Name: Dave MacQueen
1761    Date: 2000/04/30 12:40PM EDT
1762    Tag: dbm-20000430-bug_fixes
1763    Description:
1764    
1765    1. Fix for bug 1498
1766       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1767       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1768         added toRealArray function
1769       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1770         added check for tag Obj.RealArray to array printing case in ppObj
1771    
1772    2. Fix for bug 1510
1773       smlnj/src/compiler/Semant/types/typesutil.sml
1774         fixed definition of dummyargs (used by equalTycon) so that
1775         dummy args are distinct types
1776    
1777    ----------------------------------------------------------------------
1778    Name: Matthias Blume
1779    Date: 2000/04/30 01:00:00 JST
1780    Tag: blume-20000430-versions
1781    Description:
1782    
1783    1. CM version numbering added.  This is an implementation of Lal's
1784       proposal for adding version numbers and version checking to .cm
1785       files.  Lal said that his proposal was just that -- a proposal.
1786       For the time being I went ahead and implemented it so that people
1787       can comment on it.  Everything is completely backward-compatible
1788       (except for the stable library format, i.e., new bootfiles!).
1789    
1790       As usual, see the CM manual for details.
1791    
1792    2. An alternative syntax for anchored paths has been implemented.
1793       Dave has recently voiced the same concerns that I had when I did
1794       this, so there should be some support.  My take is that eventually
1795       I will let support for the current syntax (where anchors are
1796       "implicit") fade out in favor of the new, explicit syntax.
1797       In order to be backward-compatible, both old and new syntax are
1798       currently supported.
1799    
1800       Again, see the CM manual for details.
1801    
1802    3. Parallel make is trying to be slightly smarter:  When the master
1803       process finds a "bottleneck", i.e., when there is only one
1804       compilation unit that can be compiled and everybody else is
1805       waiting on it, then it will simply compile it directly instead
1806       of clumsily telling one of the slaves to do it.
1807    
1808    4. Support for "unsharing" added.  This is necessary in order to be
1809       able to have two different versions of the same library running
1810       at the same time (e.g., for trying out a new MLRISC while still
1811       having the old MLRISC linked into the current compiler, etc.)
1812       See the CM manual.
1813    
1814    5. Simple "makedepend" functionality added for generating Makefile
1815       dependency information.  (This is rather crude at the moment.
1816       Expect some changes here in the future.)
1817    
1818    6. ".fun" added as a recognized suffix for ML files. Also documented
1819       explicitly in the manual that the fallback behavior (unknown suffix
1820       -> ML file) is not an official feature!
1821    
1822    7. Small changes to the pickler for stable libraries.
1823    
1824    8. Several internal changes to CM (for cleanup/improvement).
1825    
1826    
1827    !!!! NEW BINFILES !!!!
1828    
1829    ----------------------------------------------------------------------
1830    Name: Matthias Blume
1831    Date: 2000/04/28 17:30:00 JST
1832    Tag: blume-20000428-pathconfig
1833    Description:
1834    
1835    1. I changed config/install.sh to remove duplicate entries from the
1836       lib/pathconfig file at the end.  Moreover, the final version of
1837       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1838       in src/system/installml.
1839    
1840    2. The config/install.sh script now consistently uses relative
1841       pathnames in lib/pathconfig whenever the anchor is in the lib
1842       directory.  (So far this was true for the libraries that come
1843       pre-compiled and bundled as part of the bootfiles but not for
1844       libraries that are compiled by the script itself.)
1845    
1846    ----------------------------------------------------------------------
1847    Name: Matthias Blume
1848    Date: 2000/04/26 13:10:00 JST
1849    Tag: blume-20000426-fun_suffix
1850    Description:
1851    
1852    Added ".fun" as a recognized file name suffix (for ML code).
1853    
1854    ----------------------------------------------------------------------
1855    Name: Allen Leung
1856    Date: 2000/04/25 17:00:00 EST
1857    Tag: leunga-20000425-alpha-ra
1858    Description:
1859    
1860    1. Alpha
1861    
1862        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1863    in C--.
1864    
1865    2. RA
1866    
1867       Added a flag "ra-dump-size" to print out the size of the flowgraph
1868       and the interference graph.
1869    
1870    ----------------------------------------------------------------------
1871    Name: Dave MacQueen
1872    Date: 2000/04/25/
1873    Tag: dbm-20000425-mlyacc_doc_examples
1874    Description:
1875      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1876      Updated all three examples in src/ml-yacc/examples to run
1877      under 110.* using CM.make.
1878    
1879    ----------------------------------------------------------------------
1880    Name: Allen Leung
1881    Date: 2000/04/20 23:04:00 EST
1882    Tag: leunga-20000420-ssa-c---stuff
1883    Description:
1884    
1885      This update synchronizes my repository with Yale's.  Most of these
1886    changes, however, do not affect SML/NJ at all (the RA is an exception).
1887    
1888    1. Register Allocator
1889    
1890       a. An improvement in the interference graph construction:
1891          Given a copy
1892    
1893                s <- t
1894    
1895          no interference edge between s and t is added for this definition of s.
1896    
1897       b. I've added two new spill heuristic modules that Fermin and I developed
1898          (in the new library RA.cm). These are unused in SML/NJ but maybe
1899          useful for others (Moby?)
1900    
1901    2. X86
1902    
1903       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1904    
1905    3. Alpha
1906    
1907       a. Added the BSR instruction and code generation that goes with it [C--]
1908       b. Other fixes too numerous to recount provided by Fermin [C--]
1909    
1910    4. Regmaps
1911    
1912       a. The regmaps are not initialized with the identity physical bindings
1913          at creation time.  This is unneeded.
1914    
1915    5. MLRISC Optimizations
1916    
1917       a. The DJ-Graph module can now compute the iterated dominance frontiers
1918          intersects with liveness incrementally in linear time! Woohoo!
1919          This is now used in my new SSA construction algorithm.
1920    
1921       b. THe branch reorganization module is now smarter about linear chains of
1922          basic blocks.
1923    
1924    
1925    ----------------------------------------------------------------------
1926    Name: Matthias Blume
1927    Date: 2000/04/12 13:52:00 JST
1928    Tag: blume_main_v110p27_1
1929    Description:
1930    
1931    Changed install.sh script to handle archive files without version number
1932    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1933    name of the boot file archive.
1934    
1935    ----------------------------------------------------------------------
1936    Name: Dave MacQueen
1937    Date: 2000/04/09 14:00 EDT
1938    Tag: dbm-20000410-Version_110_27
1939    Description:
1940    
1941    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1942    
1943    2. Updated src/config/version to 110.27
1944    
1945    3. New boot files!
1946    
1947    ----------------------------------------------------------------------
1948    Name: Allen Leung
1949    Date: 2000/04/09 19:09:00 EST
1950    Tag: leunga-20000409-misc
1951    Description:
1952    
1953    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1954    
1955    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1956    
1957    ----------------------------------------------------------------------
1958    Name: Stefan
1959    Date: 2000/04/07 10:00:00 EDT
1960    Tag: monnier-20000406-branch-handling
1961    Description:
1962    
1963    Improved handling of branches (mostly those generated from
1964    polymorphic equality), removed switchoff and changed the
1965    default optimization settings (more cpsopt and less flintopt).
1966    
1967    ----------------------------------------------------------------------
1968    Name: Allen Leung
1969    Date: 2000/04/06 01:30:00 EST
1970    Tag: leunga-20000406-peephole-x86-SSA-2
1971    Description:
1972    
1973       Forgot a few files.
1974    
1975    ----------------------------------------------------------------------
1976    Name: Allen Leung
1977    Date: 2000/04/06 00:36:00 EST
1978    Tag: leunga-20000406-peephole-x86-SSA
1979    Description:
1980    
1981    1.  New Peephole code
1982    
1983    2.  Minor improvement to X86 instruction selection
1984    
1985    3.  Various fixes to SSA and machine description -> code translator
1986    
1987    ----------------------------------------------------------------------
1988    Name: Matthias Blume
1989    Date: 2000/04/05 12:30:00 JST
1990    Tag: blume_main_v110p26p2_3
1991    Description:
1992    
1993    This update just merges three minor cosmetic updates to CM's sources
1994    to get ready for the 110.27 code freeze on Friday.  No functionality
1995    has changed.
1996    
1997    ----------------------------------------------------------------------
1998    Name: Allen Leung
1999    Date: 2000/04/04 19:39:00 EST
2000    Tag: leunga-20000404-x86-asm
2001    Description:
2002    
2003    1.  Fixed a problem in X86 assembly.
2004    
2005        Things like
2006    
2007           jmp %eax
2008           jmp (%eax)
2009    
2010        should be output as
2011    
2012           jmp *%eax
2013           jmp *(%eax)
2014    
2015    2.  Assembly output
2016    
2017          Added a new flag
2018    
2019              "asm-indent-copies" (default to false)
2020    
2021          When this flag is on, parallel copies will be indented an extra level.
2022    
2023    ----------------------------------------------------------------------
2024    Name: Allen Leung
2025    Date: 2000/04/04 03:18:00 EST
2026    Tag: leunga-20000404-C--Moby
2027    Description:
2028    
2029        All of these fixes are related to C--, Moby, and my own optimization
2030        stuff; so they shouldn't affect SML/NJ.
2031    
2032    1.  X86
2033    
2034        Various fixes related floating point, and extensions.
2035    
2036    2.  Alpha
2037    
2038        Some extra patterns related to loads with signed/zero extension
2039        provided by Fermin.
2040    
2041    3.  Assembly
2042    
2043        When generating assembly, resolve the value of client defined constants,
2044        instead of generating symbolic values.  This is controlled by the
2045        new flag "asm-resolve-constants", which is default to true.
2046    
2047    4.  Machine Descriptions
2048    
2049        a. The precedence parser was slightly broken when parsing infixr symbols.
2050        b. The type generalizing code had the bound variables reversed, resulting
2051           in a problem during arity raising.
2052        c. Various fixes in machine descriptions.
2053    
2054    ----------------------------------------------------------------------
2055    Name: Matthias Blume
2056    Date: 2000/04/03 16:05:00 JST
2057    Tag: blume_main_v110p26p2_2
2058    Description:
2059    
2060    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
2061    now done via the ordinary static environment that is context to each
2062    compilation unit.
2063    
2064    To this end, I arranged that instead of "structure Core" as "structure
2065    _Core" is bound in the pervasive environment.  Core access is done via
2066    _Core (which can never be accidentally rebound because _Core is not a
2067    legal surface-syntax symbol).
2068    
2069    The current solution is much cleaner because the core environment is
2070    now simply part of the pervasive environment which is part of every
2071    compilation unit's context anyway.  In particular, this eliminates all
2072    special-case handling that was necessary until now in order to deal
2073    with dynamic and symbolic parts of the core environment.
2074    
2075    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2076    compilation manager's bootstrap compiler (actually: in the "init group"
2077    handling).  See the comments in src/system/smlnj/init/init.cmi for
2078    more details.
2079    
2080    I also tried to track down all mentions of "Core" (as string argument
2081    to Symbol.strSymbol) in the compiler and replaced them with a
2082    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
2083    name appears in one place only.
2084    
2085    Binfile and bootfile format have not changed, but the switchover from
2086    the old "init.cmi" to the new one is a bit tricky, so I supplied new
2087    bootfiles anyway.
2088    
2089    ----------------------------------------------------------------------
2090    Name: Allen Leung
2091    Date: 2000/04/02 21:17:00 EST
2092    Tag: leunga-20000402-mltree
2093    Description:
2094    
2095       1. Renamed the constructor CALL in MLTREE by popular demand.
2096       2. Added a bunch of files from my repository.  These are currently
2097          used by other non-SMLNJ backends.
2098    
2099    ----------------------------------------------------------------------
2100  Name: Allen Leung  Name: Allen Leung
2101  Date: 2000/03/31 21:15:00 EST  Date: 2000/03/31 21:15:00 EST
2102  Tag: leunga-20000331-aliasing  Tag: leunga-20000331-aliasing
# Line 104  Line 2188 
2188     (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
2189     used).     used).
2190    
2191     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
2192     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
2193     Makefiles or Shell scripts.     Makefiles or Shell scripts.
2194    
2195     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 117  Line 2201 
2201    
2202     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
2203    
2204     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
2205     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
2206    
2207  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 251  Line 2335 
2335    
2336       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
2337       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
2338       testl is perferable to andl.       testl is preferable to andl.
2339    
2340  7. RA (x86 only)  7. RA (x86 only)
2341    
# Line 433  Line 2517 
2517    
2518  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2519  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2520  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
2521  explanation.  explanation.
2522    
2523  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2524  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2525  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2526  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 480  Line 2564 
2564  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
2565  <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,
2566  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
2567  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2568  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2569  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2570    
# Line 530  Line 2614 
2614    it from that remote directory.    it from that remote directory.
2615    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2616    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2617    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2618    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2619    
2620    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.590  
changed lines
  Added in v.828

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