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 634, Mon May 1 10:03:50 2000 UTC revision 818, Fri May 4 20:58:47 2001 UTC
# Line 13  Line 13 
13  Description:  Description:
14  ----------------------------------------------------------------------  ----------------------------------------------------------------------
15  Name: Matthias Blume  Name: Matthias Blume
16    Date: 2001/05/04 17:00:00 EDT
17    Tag: blume-20010504-cm-lsplit
18    Description:
19    
20    0. John merged pending changes to $/smlnj-lib.cm
21    
22    1. Allen's previous change accidentally backed out of one of Lal's
23       earlier changes.  I undid this mistake (re-introducing Lal's change).
24    
25    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
26       within the compiler where applicable.  There is some code simplification
27       because of that.
28    
29    3. The "split" phase (in FLINT) is now part of the default list of phases.
30       Compiler.Control.LambdaSplitting.* can be used to globally control the
31       lambda-splitting (cross-module-inlining) engine.  In addition to that,
32       it can now also be controlled on a per-source basis: CM has been taught
33       a new tool parameter applicable to ML source files.
34    
35       - To turn lambda-splitting off completely:
36            local open Compiler.Control.LambdaSplitting in
37                val _ = set Off
38            end
39       - To make "no lambda-splitting" the global default (but allow per-source
40         overriding); this is the initial setting:
41            local open Compiler.Control.LambdaSplitting in
42                val _ = set (Default NONE)
43            end
44       - To make "lambda-splitting with aggressiveness a" the global default
45         (and allow per-source overriding):
46            local open Compiler.Control.LambdaSplitting in
47                val _ = set (Default (SOME a))
48            end
49    
50       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
51         write (in the respective .cm-file):
52            a.sml (lambdasplitting:off)
53       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
54            a.sml (lambdasplitting:on)
55       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
56         <a> is a decimal non-negative integer):
57            a.sml (lambdasplitting:<a>)
58       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
59            a.sml (lambdasplitting:infinity)
60       - To use the global default for a.sml:
61            a.sml (lambdasplitting:default)
62         or simply
63            a.sml
64    
65    ----------------------------------------------------------------------
66    Name: Allen Leung
67    Date: 2001/05/04 01:57:00 EDT
68    Tag: leunga-20010504-sync
69    Description:
70    
71      MLRISC features.
72    
73      1. Fix to CMPXCHG instructions.
74      2. Changed RA interface to allow annotations in callbacks.
75      3. Added a new method to the stream interface to allow annotations updates.
76    
77    ----------------------------------------------------------------------
78    Name: Matthias Blume
79    Date: 2001/05/01 11:45:00 EDT
80    Tag: blume-20010501-pcedittmp
81    Description:
82    
83    Changed install.sh to use the current working directory instead of
84    /usr/tmp for a temporary file (pcedittmp).  The previous choice
85    of /usr/tmp caused trouble with MacOS X because of file premission
86    problems.
87    
88    ----------------------------------------------------------------------
89    Name: Matthias Blume
90    Date: 2001/04/20 11:10:00 EDT
91    Tag: blume-20010420-inMLflag
92    Description:
93    
94     - added vp_limitPtrMask to vproc-state.h
95       (for use by the raw-C-calls mechanism to implement proper interrupt
96        handling)
97     - made the ML compiler aware of various data-structure offsets so it
98       can generate code for accessing the vp_inML flag and vp_limitPtrMask
99     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
100       raw C-calls
101    
102    ----------------------------------------------------------------------
103    Name: Lal George
104    Date: 2001/04/20 09:15:28 EDT
105    Tag: george-20010420-macosX
106    Description:
107    
108     - Changes to port to Mac OS X; Darwin.
109    
110     - In the process I found that sqrt was broken on the PPC, because the
111       fsqrt instruction is not implemented.
112    
113    ----------------------------------------------------------------------
114    Name: Matthias Blume
115    Date: 2001/04/18 12:45:00 EDT
116    Tag: blume-20010418-ccalls
117    Description:
118    
119     - fixed two off-by-4 errors in the x86-specific c-calls implementation
120       (this bug prevented structure arguments containing pointers from being
121        passed correctly)
122     - changed the raw-C-call code in mlriscGen.sml in such a way that
123       structure arguments are represented as a pointer to the beginning
124       of the structure (instead of having a series of synthesized arguments,
125       one for each structure member)
126    
127     - made makeml script's verbosity level configurable via environment
128       variable (MAKEML_VERBOSITY)
129    
130     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
131       in rawmem-x86.sml; we are now using the real thing
132    
133    ----------------------------------------------------------------------
134    Name: Matthias Blume
135    Date: 2001/03/22 16:25:00 EST
136    Tag: blume-20010322-bootfiles
137    Description:
138    
139    Created a new set of bootfiles (for your automatic installation convenience).
140    
141    ----------------------------------------------------------------------
142    Name: Matthias Blume
143    Date: 2001/03/22 15:10:00 EST
144    Tag: blume-20010322-rawmem-parcm
145    Description:
146    
147    1. All "raw memory access" primitives for the new FFI are implemented now
148       (at least on the x86).
149    2. Some further cleanup of CM's parallel make mechanism.
150    
151    ----------------------------------------------------------------------
152    Name: Matthias Blume
153    Date: 2001/03/19 17:53:00 EST
154    Tag: blume-20010319-parallel
155    Description:
156    
157    Parallel make (using compile servers) now works again.
158    
159    To this end, CM.stabilize and CMB.make have been modified to work in
160    two passes when compile servers are attached:
161       1. Compile everything, do not perform stabilization; this pass
162          uses compile servers
163       2. Stabilize everything; this pass does not use compile servers
164    If there are no compile servers, the two passes are combined into one
165    (as before).  Splitting the passes increases the inherent parallelism
166    in the dependency graph because the entire graph including all
167    libraries is available at the same time.  This, in turn, improves
168    server utilization.  The downside is that the master process will
169    have to do some extra work after compilation is done (because for
170    technical reasons it must re-read all the binfiles during stabilization).
171    
172    ----------------------------------------------------------------------
173    Name: Matthias Blume
174    Date: 2001/03/16 12:22:00 EST
175    Tag: blume-20010316-bootfiles
176    Description:
177    
178    Created a new set of bootfiles (for your automatic installation convenience).
179    
180    ----------------------------------------------------------------------
181    Name: Matthias Blume
182    Date: 2001/03/16 11:00:00 EST
183    Tag: blume-20010316-MLTREE-fixup
184    Description:
185    
186    This is a minor fixup for an (untagged) earlier commit by Allen.
187    (A file was missing).
188    
189    ----------------------------------------------------------------------
190    Name: Allen Leung
191    Date: Mon Mar  5 18:54:57 EST 2001
192    Tag: leunga-20010305-cut-support
193    
194    1. New support for alternative control-flow in MLTREE.
195       Currently we support
196    
197          FLOW_TO(CALL ...., [k1,...,kn])
198    
199       This is needed for 'cuts to' in C-- and try/handle-like constructs
200       in Moby
201    
202       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
203    
204    2. Register Allocator
205    
206       Changes in interface [from Fermin, John]
207    
208    3. Alpha 8-bit SLL support [Fermin]
209    
210    4. All architectures
211    
212       A new module (ClusterExpandCopies) for expanding parallel copies.
213    
214    ----------------------------------------------------------------------
215    Name: Allen Leung
216    Date: 2001/02/27 23:07:00 EST
217    Tag: leunga-20010227-minor-stuff
218    
219    1. Alpha bug fix for CMOVNE
220    2. Handle mltree COND(..,FCMP ...,...)
221    3. Bug fix in simplifier
222    
223    ----------------------------------------------------------------------
224    Name: Matthias Blume
225    Date: 2001/01/30 17:50:00 EST
226    Tag: blume-20010130-sync
227    Description:
228    
229    This is just a minor update to sync my devel branch with the main brach.
230    The only visible change is the addition of some README files.
231    
232    ----------------------------------------------------------------------
233    Name: Matthias Blume
234    Date: 2001/01/12 23:30:00 JST
235    Tag: blume-20010112-bootfiles
236    Description:
237    
238    Made a new set of bootfiles that goes with the current state of the
239    repository.
240    
241    ----------------------------------------------------------------------
242    Name: Matthias Blume
243    Date: 2001/01/12 21:20:00 JST
244    Tag: blume-20010112-sync
245    Description:
246    
247    I am just flushing out some minor changes that had accumulated in
248    my private branch in order to sync with the main tree.  (This is
249    mainly because I had CVS trouble when trying to merge _into_ my
250    private branch.)
251    
252    Most people should be completely unaffected by this.
253    
254    ----------------------------------------------------------------------
255    Name: Allen Leung
256    Date: Thu Jan 11 21:03:00 EST 2001
257    Tag: leunga-20010111-labexp=mltree
258    Description:
259    
260    1.  Removed the type LabelExp and replace it by MLTree.
261    2.  Rewritten mltree-simplify with the pattern matcher tool.
262    3.  There were some bugs in alpha code generator which would break
263        64-bit code generation.
264    4.  Redo the tools to generate code with the
265    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
266        are now generated by perl scripts.
267    
268    ----------------------------------------------------------------------
269    Name: Matthias Blume
270    Date: 2001/01/10 21:55:00 JST
271    Tag: blume-20010110-rcc
272    Description:
273    
274    The RCC stuff now seems to work (but only on the x86).
275    This required hacking of the c-calls interface (and -implementation) in
276    MLRISC.
277    
278    Normal compiler users should be unaffected.
279    
280    ----------------------------------------------------------------------
281    Name: Matthias Blume
282    Date: 2001/01/09 01:20:00 JST
283    Tag: blume-20010109-rcc
284    Description:
285    
286    This is a fairly big patch, flushing out a large number of pending
287    changes that I made to my development copy over the last couple of days.
288    
289    Of practical relevance at this moment is a workaround for a pickling
290    bug that Allen ran into the other day.  The cause of the bug itself is
291    still unknown and it might be hard to fix it properly, but the
292    workaround has some merits of its own (namely somewhat reducing pickling
293    overhead for certain libraries).  Therefore, I think this solution should
294    be satisfactory at this time.
295    
296    The rest of the changes (i.e., the vast majority) has to do with my
297    ongoing efforts of providing direct support for C function calls from
298    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
299    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
300    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
301    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
302    handle RCC.
303    
304    The last part (the changes to mlriscGen.sml) are still known to be
305    wrong on the x86 and not implemented on all other architectures.  But
306    the infrastructure is in place. I had to change a few functor
307    signatures in the backend to be able to route the CCalls interface
308    from MLRISC there, and I had to specialize the mltree type (on the
309    x86) to include the necessary extensions. (The extensions themselves
310    were already there and redy to go in MLRISC/x86).
311    
312    Everything should be very happy as soon as someone helps me with
313    mlriscGen.sml...
314    
315    In any case, nothing of this should matter to anyone as long as the
316    new primop is not being used (which is going to be the case unless you
317    find it where I hid it :). The rest of the compiler is completely
318    unaffected.
319    
320    ----------------------------------------------------------------------
321    Name: Matthias Blume
322    Date: 2001/01/05 00:30:00 JST
323    Tag: blume-20010105-primops
324    Description:
325    
326    Added some experimental support for work that I am doing right now.
327    These changes mostly concern added primops, but there is also a new
328    experimental C library in the runtime system (but currently not enabled
329    anywhere except on Linux/X86).
330    
331    In the course of adding primops (and playing with them), I discovered that
332    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
333    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
334    To recover, I made type information in INL_PRIM mandatory and changed
335    prim.sml as well as built-in.sml accordingly.  The InLine structure now
336    has complete, correct type information (i.e., no bottom types).
337    
338    Since all these changes mean that we need new binfiles, I also bumped the
339    version number to 110.32.1.
340    
341    ----------------------------------------------------------------------
342    Name: Matthias Blume
343    Date: 2000/12/30 22:10:00 JST
344    Tag: blume-20001230-various
345    Description:
346    
347    Added proxy libraries for MLRISC and let MLRISC libraries refer
348    to each other using path anchors.  (See CM manual for explanation.)
349    
350    Updated CM documentation.
351    
352    Fixed some bugs in CM.
353    
354    Implemented "proxy" libraries (= syntactic sugar for CM).
355    
356    Added "-quiet" option to makeml and changed runtime system accordingly.
357    
358    Added cleanup handler for exportML to reset timers and compiler stats.
359    
360    ----------------------------------------------------------------------
361    Name: Lal George
362    Date: 2000/12/22 22:22:58 EST 2000
363    Tag: Release_110_32
364    Description:
365    
366            Infinite precision used throughout MLRISC.
367            see MLRISC/mltree/machine-int.sig
368    
369    ----------------------------------------------------------------------
370    Name: Matthias Blume
371    Date: 2000/12/22 23:16:00 JST
372    Tag: blume-20001222-warn
373    Description:
374    
375    Corrected wording and formatting of some CM warning message which I
376    broke in my previous patch.
377    
378    ----------------------------------------------------------------------
379    Name: Matthias Blume
380    Date: 2000/12/22 21:20:00 JST
381    Tag: blume-20001222-anchorenv
382    Description:
383    
384    Fixed CM's handling of anchor environments in connection with CMB.make.
385    
386    ----------------------------------------------------------------------
387    Name: Matthias Blume
388    Date: 2000/12/22 13:15:00 JST
389    Tag: blume-20001222-cleanup
390    Description:
391    
392    Removed src/cm/ffi which does not (and did not) belong here.
393    
394    ----------------------------------------------------------------------
395    Name: Matthias Blume
396    Date: 2000/12/21 23:55:00 JST
397    Tag: blume-20001221-exn
398    Description:
399    
400    Probably most important: CM no longer silently swallows all exceptions
401    in the compiler.
402    Plus: some other minor CM changes.  For example, CM now reports some
403    sizes for generated binfiles (code, data, envpickle, lambdapickle).
404    
405    ----------------------------------------------------------------------
406    Name: Matthias Blume
407    Date: 2000/12/15 00:01:05 JST
408    Tag: blume-20001215-dirtool
409    Description:
410    
411    - "dir" tool added.
412    - improvements and cleanup to Tools structure
413    - documentation updates
414    
415    ----------------------------------------------------------------------
416    Name: Allen Leung
417    Date: Thu Dec 14 03:45:24 EST 2000
418    Description:
419    Tag:  leunga-20001214-int-inf
420    Description:
421    
422       In IntInf, added these standard functions, which are missing from our
423    implementation:
424    
425        andb : int * int -> int
426        xorb : int * int -> int
427        orb  : int * int -> int
428        notb : int -> int
429         <<   : int * word -> int
430        ~>>  : int * word -> int
431    
432       Not tested, I hope they are correct.
433    
434    ----------------------------------------------------------------------
435    Name: Allen Leung
436    Date: Fri Dec  8 19:23:26 EST 2000
437    Description:
438    Tag:  leunga-20001208-nowhere
439    Description:
440    
441      Slight improvements to the 'nowhere' tool to handle OR-patterns,
442    to generate better error messages etc.  Plus a brief manual.
443    
444    ----------------------------------------------------------------------
445    Name: Lal George
446    Date: 2000/12/08 09:54:02 EST 2000
447    Tag: Release_110_31
448    Description:
449    
450    - Version 110.31
451    ----------------------------------------------------------------------
452    Name: Allen Leung
453    Date: Thu Dec  7 22:01:04 EST 2000
454    Tag:  leunga-20001207-cell-monster-hack
455    Description:
456    
457    Major MLRISC internal changes.  Affect all clients.
458    Summary:
459    
460    1.  Type CELLS.cell = int is now replaced by a datatype.
461        As a result, the old regmap is now gone.  Almost all interfaces
462        in MLRISC change as a consequence.
463    
464    2.  A new brand version of machine description tool (v3.0) that generates
465        modules expecting the new interface.  The old version is removed.
466    
467    3.  The RA interface has been further abstracted into two new functors.
468        RISC_RA and X86RA.  These functors have much simpler interfaces.
469        [See also directory MLRISC/demo.]
470    
471    4.  Some other new source->source code generation tools are available:
472    
473        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
474        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
475           I use this tool to generate the peephole optimizers---with the new
476           cell type changes, peephole rules are becoming difficult to write
477           without conditional pattern matching.
478    
479    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
480        cover the entire MLRISC source tree so many things broke.
481    
482    6.  CM files have been moved to the subdirectory MLRISC/cm.
483        They are moved because there are a lot of them and they clutter up the
484        root dir.
485    
486    7.  More detailed documentation to come...
487    
488        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
489        a makeml -rebuild first.  This is because of other other
490        changes that Matthias has made (see below).
491    
492    
493    ----------------------------------------------------------------------
494    Name: Matthias Blume
495    Date: 2000/11/30 23:12:00 JST
496    Tag: blume-20001130-filereorg
497    Description:
498    
499    Some manual updates and some file reorganizations in CM.
500    
501    ----------------------------------------------------------------------
502    Name: Matthias Blume
503    Date: 2000/11/24 17:45:00 JST
504    Tag: blume-20001124-link
505    Description:
506    
507    Drastically improved link traversal code for the case that the dynamic
508    value was already loaded at bootstrap time.  As a result, CM and CMB
509    now both load blazingly fast -- even on a very slow machine.  Also,
510    memory consumption has been further reduced by this.
511    
512    Warning: The format of the PIDMAP file has changed.  THerefore, to
513    bootstrap you have to do this:
514    
515    1. Run CMB.make
516    2. Make a symbolic link for the boot directory:
517         ln -s sml.boot.ARCH-OS xxx
518    3. "Rebuild" the boot directory:
519         ./makeml -boot xxx -rebuild sml ; rm xxx
520    4. Boot normally:
521          ./makeml
522    
523    ----------------------------------------------------------------------
524    Name: Matthias Blume
525    Date: 2000/11/21 21:20:00 JST
526    Tag: blume-20001121-tools
527    Description:
528    
529    Continued hacking on autoloading problem -- with success this time.
530    Also changed tool-plugin mechanism.  See new CM manual.
531    
532    ----------------------------------------------------------------------
533    Name: Matthias Blume
534    Date: 2000/11/19 14:30:00 JST
535    Tag:  blume-20001119-autoload
536    Description:
537    
538    Some hacking to make autoloading faster.  Success for CMB, no success
539    so far for CM.  There is a reduced structure CM' that autoloads faster.
540    (This is a temporary, non-documented hack to be eliminated again when
541    the general problem is solved.)
542    
543    ----------------------------------------------------------------------
544    Name: Matthias Blume
545    Date: 2000/11/17 14:10:00 JST
546    Tag: blume-20001117-pickle-lib
547    Description:
548    
549    1. Eliminated comp-lib.cm
550    2. Made pickle-lib.cm
551    3. Eliminated all uses of intset.sml (from comp-lib.cm)
552    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
553       equivalent constructs from smlnj-lib.cm (INtHashTable).
554    5. Point 4. also goes for those uses of intmap.* in MLRISC.
555       Duplicated intmap modules thrown out.
556    6. Hunted down all duplicated SCC code and replaced it with
557       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
558    7. Rewrote Feedback module.
559    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
560       should be thrown out and equivalent modules from smlnj-lib.cm
561       should be used (IntRedBlackSet, IntListSet, ...).
562    
563    Confirmed that compiler compiles to fixpoint.
564    
565    ----------------------------------------------------------------------
566    Name: Allen Leung
567    Date: 2000/11/10 18:00:00
568    Tag: leunga-20001110-new-x86-fp
569    
570    A new x86 floating point code generator has been added.
571    By default this is turned off.  To turn this on, do:
572    
573        CM.autoload "$smlnj/compiler.cm";
574        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
575    
576    Changes:
577    
578    1.  Changed FTAN to FPTAN so that the assembly output is correct.
579    2.  Changed the extension callback for FTANGENT to generate:
580    
581              fptan
582              fstp  %st(0)
583        instead of
584              fptan
585              fstpl ftempmem
586    
587    3.  Numerous assembly fixes for x86.
588    
589    5.  Cleaned up the machine code output module x86/x86MC.sml and added
590        support for a whole bunch of instructions and addressing modes:
591    
592          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
593          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
594          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
595          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
596          fxch %st(n)
597          fld %st(n)
598          fst %st(n)
599          fst mem
600          fstp %st(n)
601          fucom %st(n)
602          fucomp %st(n)
603    
604        All these are now generated when the fast fp mode is turned on.
605    
606    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
607    
608    ----------------------------------------------------------------------
609    Name: Matthias Blume
610    Date: 2000/11/09 11:20:00 JST
611    Tag: blume-20001109-scc
612    Description:
613    
614    Eliminated some code duplication:
615    
616    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
617       (Otherwise the functor is useless.)
618    2. Used GraphSCCFn where SCCUtilFun was used previously.
619    3. Got rid of SCCUtilFun (in comp-lib.cm).
620    
621    ----------------------------------------------------------------------
622    Name: Lal George
623    Date: 2000/11/06 09:02:21 EST 2000
624    Tag: Release_110_30
625    Description:
626    
627    - Version 110.30
628    ----------------------------------------------------------------------
629    Name: Matthias Blume
630    Date: 2000/11/04 14:45:00
631    Tag: blume-20001104-mlbuild
632    Description:
633    
634    - Made ml-build faster on startup.
635    - Documentation fixes.
636    
637    ----------------------------------------------------------------------
638    Name: Matthias Blume
639    Date: 2000/11/02 17:00:00 JST
640    Tag: blume-20001102-condcomp
641    Description:
642    
643    - Small tweaks to pickler -- new BOOTFILES!
644    - Version bumped to 110.29.2.
645    - Added conditional compilation facility to init.cmi (see comment there).
646    ----------------------------------------------------------------------
647    Name: Allen Leung
648    Date: 2000/10/23 19:31:00
649    Tag: leunga-20001023-demo-ra
650    
651    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
652    2. Test programs for the graph library updated
653    3. Some new MLRISC demo programs added
654    
655    ----------------------------------------------------------------------
656    Name: Matthias Blume
657    Date: 2000/08/31 22:15:00 JST
658    Tag: blume-20001017-errmsg
659    Description:
660    
661    More error message grief: Where there used to be no messages, there
662    now were some that had bogus error regions.  Fixed.
663    
664    ----------------------------------------------------------------------
665    Name: Matthias Blume
666    Date: 2000/08/31 17:30:00 JST
667    Tag: blume-20001017-v110p29p1
668    Description:
669    
670    I made a version 110.29.1 with new bootfiles.
671    
672    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
673              CM documentation changes and a small bugfix in CM's error reporting.
674    
675    ----------------------------------------------------------------------
676    Name: Lal George
677    Date: 2000/09/27 14:42:35 EDT
678    Tag: george-20000927-nodestatus
679    Description:
680    
681    Changed the type of the nodestatus, so that:
682    
683            SPILLED(~1)             is now SPILLED
684            SPILLED(m) where m>=0   is now MEMREG(m)
685            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
686    
687    ----------------------------------------------------------------------
688    Name: Matthias Blume
689    Date: 2000/09/07 14:45:00 JST
690    Tag: blume-20000907-cmerrmsg
691    Description:
692    
693    Small tweak to CM to avoid getting ML syntax error messages twice.
694    
695    ----------------------------------------------------------------------
696    Name: Matthias Blume
697    Date: 2000/08/31 18:00:00 JST
698    Tag: blume-20000831-cvsbootfiles
699    Description:
700    
701    New URL for boot files (because the 110.29 files on the BL server do
702    now work correctly with my updated install scripts for yacc and lex).
703    
704    ----------------------------------------------------------------------
705    Name: Matthias Blume
706    Date: 2000/08/08 12:33:00 JST
707    Tag: blume-20000808-manual
708    Description:
709    
710    Tiny update to CM manual.
711    
712    ----------------------------------------------------------------------
713    Name: Allen Leung
714    Date: 2000/08/7 19:31:00
715    Tag: leunga-20000807-a-whole-bunch-of-stuff
716    
717      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
718    and Mobdy.
719    
720    1.  x86
721    
722       a.  Fixes to peephole module by John and Dan.
723       b.  Assembly fix to SETcc by Allen.
724       c.  Fix to c-call by John.
725       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
726       e.  Instruction selection optimization to SETcc as suggested by John.
727    
728           For example,
729    
730            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
731    
732           should generate:
733    
734            MOVL a, x
735            SUBL b, x
736            SHRL 31, x
737    
738    2.  IR stuff
739    
740         A bunch of new DJ-graph related algorithms added.  These
741         speed up SSA construction.
742    
743    3.  SSA + Scheduling
744    
745         Added code for SSA and scheduling to the repository
746    
747    ----------------------------------------------------------------------
748    Name: Lal George
749    Date: 2000/07/27 11:53:14 EDT
750    
751    Tag: lal-20000727-linux-ppc
752    Description:
753    
754     Made changes to support Linux PPC.
755     p.s. I have confirmation that the 110.29 boot files work fine.
756    
757    ----------------------------------------------------------------------
758    Name: Matthias Blume
759    Date: 2000/07/27 17:40:00 JST
760    Tag: blume-20000727-scripts
761    Description:
762    
763    !!!! WARNING !!!!
764    You must recompile the runtime system!
765    !!!! WARNING !!!!
766    
767    This is basically another round of script-enhancements:
768    
769    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
770       and undefine CM preprocessor symbols.
771    
772    2. ml-build avoids generating a new heap image if it finds that the
773       existing one is still ok.  (The condition is that no ML file had to
774       be recompiled and all ML files are found to be older that the heap
775       file.)
776    
777       To make this work smoothly, I also hacked the runtime system as
778       well as SMLofNJ.SysInfo to get access to the heap image suffix
779       (.sparc-solaris, ...) that is currently being used.
780    
781       Moreover, the signature of CM.mk_standalone has changed.  See the
782       CM manual.
783    
784    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
785       CM manual for details.)
786    
787    4. More CM manual updates:
788        - all of the above has been documented.
789        - there is now a section describing the (CM-related) command line
790          arguments that are accepted by the "sml" command
791    
792    ----------------------------------------------------------------------
793    Name: Matthias Blume
794    Date: 2000/07/25 16:20:00 JST
795    Tag: blume-20000725-makedepend
796    Description:
797    
798    Added a script called ml-makedepend.  This can be used in makefiles
799    for Unix' make in a way very similar to the "makedepend" command for
800    C.
801    
802    The script internally uses function CM.sources.
803    
804    Synopsis:
805    
806        ml-makedepend [-f makefile] cmfile targetname
807    
808    The default for the makefile is "makefile" (or "Makefile" should
809    "makefile" not exist).
810    
811    ml-makedepend adds a cmfile/targetname-specific section to this
812    makefile (after removing the previous version of this section).  The
813    section contains a single dependency specification with targetname on
814    the LHS (targetname is an arbitrary name), and a list of files derived
815    from the cmfile on the RHS.  Some of the files on the RHS are
816    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
817    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
818    path names.  The makefile writer is responsible for making sure that
819    these variables have correct at the time "make" is invoked.
820    
821    ----------------------------------------------------------------------
822    Name: Matthias Blume
823    Date: 2000/07/22 23:30:00 JST
824    Tag: blume-20000722-urlupdate
825    Description:
826    
827    Changed BOOT and config/srcarchiveurl to point to BL server:
828    
829        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
830    
831    ----------------------------------------------------------------------
832    Name: Matthias Blume
833    Date: 2000/07/18 18:00:00 JST
834    Tag: blume-20000718-Version_110_29
835    Description:
836    
837    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
838    
839    2. Updated config/version to 110.29
840    
841    3. Updated config/srcarchiveurl
842    
843    3. New boot files!
844       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
845    
846    ----------------------------------------------------------------------
847    Name: Matthias Blume
848    Date: 2000/07/11 13:58:00 JST
849    Tag: blume-20000711-doctypo
850    Description:
851    
852    Fixed a few typos in CM manual.
853    
854    ----------------------------------------------------------------------
855    Name: Allen Leung
856    Date: 2000/06/15 00:38:00
857    Tag: leunga-20000704-sparc-x86
858    
859    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
860    2. fix to x86 RET bug [found by Dan Grossman]
861    3. sparc assembly bug fix for ticc instructions [found by Fermin]
862    
863       Affects c-- and moby only
864    
865    ----------------------------------------------------------------------
866    Name: Matthias Blume
867    Date: 2000/07/04 15:26:00
868    Tag: blume-20000704-trigger
869    Description:
870    
871    1. Improvements to CM manual.
872    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
873       of getting a back-trace.  The function, when called, raises an
874       internal exception which explicitly carries the full back-trace history,
875       so it is unaffected by any intervening handle-raise pairs ("trivial"
876       or not).  The interactive loop will print that history once it arrives
877       at top level.
878       Short of having all exceptions implicitly carry the full history, the
879       recommended way of using this facility is:
880         - compile your program with instrumentation "on"
881         - run it, when it raises an exception, look at the history
882         - if the history is "cut off" because of some handler, go and modify
883           your program so that it explicitly calls BTrace.trigger
884         - recompile (still instrumented), and rerun; look at the full history
885    
886    ----------------------------------------------------------------------
887    Name: Matthias Blume
888    Date: 2000/07/03 15:36:00 JST
889    Tag: blume-20000702-manual
890    Description:
891    
892    Small corrections and updates to CM manual.
893    
894    ----------------------------------------------------------------------
895    Name: Matthias Blume
896    Date: 2000/06/29 16:04:00 JST
897    Tag: blume-20000629-yacctool
898    Description:
899    
900    Changes:
901    
902    1. Class "mlyacc" now takes separate arguments to pass options to
903       generated .sml- and .sig-files independently.
904    2. Corresponding CM manual updates.
905    3. BTrace module now also reports call sites.  (However, for loop clusters
906       it only shows from where the cluster was entered.)  There are associated
907       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
908    
909    ----------------------------------------------------------------------
910    Name: Matthias Blume
911    Date: 2000/06/27 16:51:00 JST
912    Tag: blume-20000627-noweb
913    Description:
914    
915    Changes:
916    
917     1. Implemented "subdir" and "witness" options for noweb tool.
918        This caused some slight internal changes in CM's tool implementation.
919     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
920        some remaining issues from earlier path anchor changes.
921     3. Updated CM manual accordingly.
922    
923     4. Changed implementation of back-tracing so that I now consider it
924        ready for prime-time.
925    
926        In particular, you don't have to explicitly trigger the back-trace
927        anymore.  Instead, if you are running BTrace-instrumented code and
928        there is an uncaught exception (regardless of whether or not it was
929        raised in instrumented code), the top-level evalloop will print
930        the back-trace.
931    
932        Features:
933    
934          - Instrumented and uninstrumented code work together seemlessly.
935            (Of course, uninstrumented code is never mentioned in actual
936             back-traces.)
937    
938          - Asymptotic time- and space-complexity of instrumented code is
939            equal to that of uninstrumented code.  (This means that
940            tail-recursion is preserved by the instrumentation phase.)
941    
942          - Modules whose code has been instrumented in different sessions
943            work together without problem.
944    
945          - There is no penalty whatsoever on uninstrumented code.
946    
947          - There is no penalty on "raise" expressions, even in
948            instrumented code.
949    
950        A potential bug (or perhaps it is a feature, too):
951    
952          A back-trace reaches no further than the outermost instrumented
953          non-trivial "raise".  Here, a "trivial" raise is one that is the
954          sole RHS of a "handle" rule.  Thus, back-traces reach trough
955    
956               <exp> handle e => raise e
957    
958          and even
959    
960               <exp> handle Foo => raise Bar
961    
962          and, of course, through
963    
964               <exp> handle Foo => ...
965    
966         if the exception was not Foo.
967    
968         Back-traces always reach right through any un-instrumented code
969         including any of its "handle" expressions, trivial or not.
970    
971       To try this out, do the following:
972    
973         - Erase all existing binfiles for your program.
974           (You may keep binfiles for those modules where you think you
975            definitely don't need back-tracing.)
976         - Turn on back-trace instrumentation:
977              SMLofNJ.Internals.BTrace.mode (SOME true);
978         - Recompile your program.  (I.e., run "CM.make" or "use".)
979         - You may now turn instrumentation off again (if you want):
980              SMLofNJ.Internals.BTrace.mode (SOME false);
981         - Run your program as usual.  If it raises an exception that
982           reaches the interactive toplevel, then a back-trace will
983           automatically be printed.  After that, the toplevel loop
984           will print the exception history as usual.
985    
986    ----------------------------------------------------------------------
987    Name: Matthias Blume
988    Date: 2000/06/26 09:56:46 JST
989    Tag: blume-20000626-setup
990    Description:
991    
992    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
993          ML code before and after compiling a file (e.g., to set compiler
994          flags)
995    
996    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
997              - associated changes to btrace.sml (BTrace instrumentation pass)
998              - cleaner implementation of btimp.sml (BTrace tracing and report
999                module)
1000    
1001    CM manual: * new path encoding documented
1002               * description of setup-parameter to "sml" added
1003    
1004    The biggest user-visible change to back-tracing is that it is no
1005    longer necessary to compile all traced modules within the same
1006    session.  (This was a real limitation.)
1007    
1008    ----------------------------------------------------------------------
1009    Name: Matthias Blume
1010    Date: 2000/06/24 12:40:00 JST
1011    Tag: blume-20000624-startup
1012    Description:
1013    
1014    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1015    bit too often -- to put it mildly. :)
1016    
1017    ----------------------------------------------------------------------
1018    Name: Matthias Blume
1019    Date: 2000/06/23 18:20:00 JST
1020    Tag: blume-20000623-btrace
1021    Description:
1022    
1023    This updates adds a backtrace facility to aid programmers in debugging
1024    their programs.  This involves the following changes:
1025    
1026    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1027       keeping track of the current call stack.  When programs are compiled
1028       in a special mode, the compiler will insert calls to these hooks
1029       into the user program.
1030       "Hook" means that it is possible for different implementations of
1031       back-tracing to register themselves (at different times).
1032    
1033    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1034       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1035       It can be turned on using this call:
1036         SMLofNJ.Internals.BTrace.mode (SOME true);
1037       Turning it off again:
1038         SMLofNJ.Internals.BTrace.mode (SOME false);
1039       Querying the current status:
1040         SMLofNJ.Internals.BTrace.mode NONE;
1041       Annotated programs are about twice as big as normal ones, and they
1042       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1043       where all hooks do nothing).  The slowdown with a plugin that is
1044       actually useful (such as the one supplied by default) is even greater,
1045       but in the case of the default plugin it is still only an constant
1046       factor (amortized).
1047    
1048    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1049       with a sub-structure BTrace for controlling back-tracing.  In particular,
1050       the above-mentioned function "mode" controls whether the annotation
1051       phase is invoked by the compiler.  Another important function is
1052       "trigger": when called it aborts the current execution and causes
1053       the top-level loop to print a full back-trace.
1054    
1055    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1056       for back-tracing.  It keeps track of the dynamic call stack and in
1057       addition to that it keeps a partial history at each "level" of that
1058       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1059       d and b (at separate times, dynamically), then the report will show:
1060    
1061       GOTO   d
1062             /c
1063       GOTO  \b
1064       CALL   a
1065    
1066       This shows that there was an initial non-tail call of a, then a
1067       tail-call to b or c, looping behavior in a cluster of functions that
1068       consist of b and c, and then a goto from that cluster (i.e., either from
1069       b or from c) to d.
1070    
1071       Note that (depending on the user program) the amount of information
1072       that the back-trace module has to keep track of at each level is bounded
1073       by a constant.  Thus, the whole implementation has the same asymptotical
1074       complexity as the original program (both in space and in time).
1075    
1076    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1077       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1078       which is raised by the "trigger" function mentioned above.
1079    
1080    Notes on usage:
1081    
1082    - Annotated code works well together with unannotated code:
1083    Unannotated calls simply do not show up at all in the backtrace.
1084    
1085    - It is not a good idea to let modules that were annotated during
1086    different sessions run at the same time.  This is because the compiler
1087    chooses small integers to identify individual functions, and there
1088    will be clashes if different modules were compiled in separate sessions.
1089    (Nothing will crash, and you will even be told about the clashes, but
1090    back-trace information will in general not be useful.)
1091    
1092    - Back-tracing can be confused by callcc and capture.
1093    
1094    - The only way of getting a back-trace right now is to explicitly
1095    invoke the "trigger" function from your user program.  Eventually, we
1096    should make every exception carry back-trace information (if
1097    available).  But since this creates more overhead at "raise"-time
1098    (similar to the current exnHistory overhead), I have not yet
1099    implemented this.  (The implementation will be rather easy.)  With
1100    exceptions carrying back-trace information, this facility will be even
1101    more useful because users don't need to modify their programs...
1102    
1103    - While it is possible to compile the compiler with back-trace
1104    annotations turned on (I did it to get some confidence in
1105    correctness), you must make absolutely sure that core.sml and
1106    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1107    be compiled with annotation because there is no core access yet, but
1108    if you compile btimp.sml with annotation, then the system will go into
1109    an infinite recursion and crash.)
1110    Since CM currently does not know about BTrace, the only way to turn
1111    annotations on and off for different modules of the compiler is to
1112    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1113    this is awkward and clumsy.
1114    
1115    Sample sessions:
1116    
1117    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1118    - SMLofNJ.Internals.BTrace.mode (SOME true);
1119    [autoloading]
1120    [autoloading done]
1121    val it = false : bool
1122    - structure X = struct
1123    -     fun main n = let
1124    -         fun a (x, 0) = d x
1125    -           | a (x, n) = b (x, n - 1)
1126    -         and b (x, n) = c (x, n)
1127    -         and c (x, n) = a (x, n)
1128    -         and d x = e (x, 3)
1129    -         and e (x, 0) = f x
1130    -           | e (x, n) = e (x, n - 1)
1131    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1132    -           | f n = n * g (n - 1)
1133    -         and g n = a (n, 3)
1134    -     in
1135    -         f n
1136    -     end
1137    - end;
1138    structure X : sig val main : int -> int end
1139    - X.main 3;
1140    *** BACK-TRACE ***
1141    GOTO   stdIn:4.2-13.20: X.main[2].f
1142    GOTO-( stdIn:4.2-13.20: X.main[2].e
1143    GOTO   stdIn:4.2-13.20: X.main[2].d
1144         / stdIn:4.2-13.20: X.main[2].a
1145         | stdIn:4.2-13.20: X.main[2].b
1146    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1147    CALL   stdIn:4.2-13.20: X.main[2].g
1148    GOTO   stdIn:4.2-13.20: X.main[2].f
1149    GOTO-( stdIn:4.2-13.20: X.main[2].e
1150    GOTO   stdIn:4.2-13.20: X.main[2].d
1151         / stdIn:4.2-13.20: X.main[2].a
1152         | stdIn:4.2-13.20: X.main[2].b
1153    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1154    CALL   stdIn:4.2-13.20: X.main[2].g
1155    GOTO   stdIn:4.2-13.20: X.main[2].f
1156    GOTO-( stdIn:4.2-13.20: X.main[2].e
1157    GOTO   stdIn:4.2-13.20: X.main[2].d
1158         / stdIn:4.2-13.20: X.main[2].a
1159         | stdIn:4.2-13.20: X.main[2].b
1160    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1161    CALL   stdIn:4.2-13.20: X.main[2].g
1162    GOTO   stdIn:4.2-13.20: X.main[2].f
1163    CALL   stdIn:2.15-17.4: X.main[2]
1164    -
1165    
1166    (Note that because of a FLINt bug the above code currently does not
1167    compile without BTrace turned on.)
1168    
1169    Here is another example, using my modified Tiger compiler:
1170    
1171    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1172    - SMLofNJ.Internals.BTrace.mode (SOME true);
1173    [autoloading]
1174    [autoloading done]
1175    val it = false : bool
1176    - CM.make "sources.cm";
1177    [autoloading]
1178    ...
1179    [autoloading done]
1180    [scanning sources.cm]
1181    [parsing (sources.cm):parse.sml]
1182    [creating directory CM/SKEL ...]
1183    [parsing (sources.cm):tiger.lex.sml]
1184    ...
1185    [wrote CM/sparc-unix/semant.sml]
1186    [compiling (sources.cm):main.sml]
1187    [wrote CM/sparc-unix/main.sml]
1188    [New bindings added.]
1189    val it = true : bool
1190    - Main.compile ("../testcases/merge.tig", "foo.out");
1191    *** BACK-TRACE ***
1192    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
1193    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1194    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1195    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1196    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1197    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1198    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
1199         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
1200    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
1201    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1202    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
1203    CALL   main.sml:1.18-118.4: Main.compile[2]
1204    -
1205    
1206    ----------------------------------------------------------------------
1207    Name: Matthias Blumen
1208    Date: 2000/06/21 18:00:00 JST
1209    Tag: blume-20000621-manual
1210    Description:
1211    
1212    CM manual update: Path environments documented.
1213    
1214    ----------------------------------------------------------------------
1215    Name: Matthias Blume
1216    Date: 2000/06/19 13:40:00
1217    Tag: blume-20000619-manual
1218    Description:
1219    
1220    CM manual and system/README update.  This only covers the fact that
1221    there are no more implicit anchors.  (Path environments and the "bind"
1222    option to "cm" have yet to be documented.)
1223    
1224    ----------------------------------------------------------------------
1225    Name: Matthias Blume
1226    Date: 2000/06/19 11:05:00 JST
1227    Tag: blume-20000619-chdir-bugfix
1228    Description:
1229    
1230    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1231    
1232    ----------------------------------------------------------------------
1233    Name: Matthias Blume
1234    Date: 2000/06/18 22:00:10 JST
1235    Tag: blume-20000618-implicit-anchors-really-gone
1236    Description:
1237    
1238    I updates the previous HISTORY entry where I forgot to mention that
1239    implicit anchors are no longer with us.
1240    
1241    The current update also gets rid of the (now useless) controller
1242    CM.Control.implicit_anchors.
1243    
1244    ----------------------------------------------------------------------
1245    Name: Matthias Blume
1246    Date: 2000/06/16 17:30:00 JST
1247    Tag: blume-20000616-anchorenv
1248    Description:
1249    
1250    This patch implements the long anticipated (just kidding :) "anchor
1251    environment" mechanism.  In the course of doing this, I also
1252    re-implemented CM's internal "SrcPath" module from scratch.  The new
1253    one should be more robust in certain boundary cases.  In any case, it
1254    is a lot cleaner than its predecessor (IMHO).
1255    
1256    This time, although there is yet another boot file format change, I
1257    kept the unpickler backward-compatible.  As a result, no new bootfiles
1258    are necessary and bootstrapping is straightforward.  (You cannot read
1259    new bootfiles into an old system, but the other way around is no
1260    problem.)
1261    
1262    Visible changes:
1263    
1264    ** 0. Implicit path anchors (without the leading $-symbol) are no
1265    longer recognized at all. This means that such path names are not
1266    illegal either.  For example, the name basis.cm simply refers to a
1267    local file called "basis.cm" (i.e, the name is an ordinary path
1268    relative to .cm-files directory).  Or, to put it differently, only
1269    names that start with $ are anchored paths.
1270    
1271    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1272    vanished.
1273    
1274    John (Reppy) had critizised this as soon as I originally proposed and
1275    implemented it, but at that time I did not really deeply believe
1276    him. :) Now I came full-circle because I need the $<singlearc> syntax
1277    in another place where it cannot be seen as an abbreviation for
1278    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1279    seems to mean (i.e., it "expands" into the corresponding anchor
1280    value).
1281    
1282    However, when paths are used as members in CM description files, it
1283    continues to be true that there must be at least another arc after the
1284    anchor.  This is now enforced separately during semantic analysis
1285    (i.e., from a lexical/syntactical point of view, the notation is ok.)
1286    
1287    ** 2. The "cm" class now accepts an option "bind".  The option's value
1288    is a sub-option list of precisely two items -- one labeled "anchor"
1289    and the other one labeled "value".  As you might expect, "anchor" is
1290    used to specify an anchor name to be bound, and "value" specifies what
1291    the anchor is being bound to.
1292    
1293    The value must be a directory name and can be given in either standard
1294    syntax (including the possibility that it is itself an anchored path)
1295    or native syntax.
1296    
1297    Examples:
1298    
1299       foo.cm (bind:(anchor:bar value:$mystuff/bar))
1300       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1301    
1302    and so on.
1303    
1304    The meaning of this is that the .cm-file will be processed with an
1305    augmented anchor environment where the given anchor(s) is/are bound to
1306    the given values(s).
1307    
1308    The rationale for having this feature is this: Suppose you are trying
1309    to use two different (already stable) libraries a.cm and b.cm (that
1310    you perhaps didn't write yourself).  Further, suppose each of these
1311    two libraries internally uses its own auxiliary library $aux/lib.cm.
1312    Normally you would now have a problem because the anchor "lib" can not
1313    be bound to more than one value globally.  Therefore, the project that
1314    uses both a.cm and b.cm must locally redirect the anchor to some other
1315    place:
1316    
1317       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1318       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1319    
1320    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1321    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
1322    
1323    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1324    you when you do so at the time of CM.stabilize).  Therefore, you can
1325    also use an anchored path as the value:
1326    
1327      a.cm (bind:(anchor:lib value:$a-lib))
1328      b.cm (bind:(anchor:lib value:$b-lib))
1329    
1330    Now you can globally configure (using the usual CM.Anchor.anchor or
1331    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1332    itself is always locally bound, setting it globally is no longer
1333    meaningful or necessary (but it does not hurt either).  In fact, "lib"
1334    can still be used as a global anchor for separate purposes.  As a
1335    matter of fact, one can locally define "lib" in terms of a global
1336    "lib":
1337    
1338      a.cm (bind:(anchor:lib value:$lib/a))
1339      b.cm (bind:(anchor:lib value:$lib/b))
1340    
1341    ** 3: The encoding of path names has changed.  This affects the way
1342    path names are shown in CM's progress report and also the internal
1343    protocol encoding used for parallel make.
1344    
1345    The encoding now uses one or more ':'-separated segments.  Each
1346    segments corresponds to a file that has been specified relative to the
1347    file given by its preceding segment.  The first segment is either
1348    relative to the CWD, absolute, or anchored.  Each segment itself is
1349    basically a Unix pathname; all segments but the first are relative.
1350    
1351    Example:
1352    
1353       $foo/bar/baz.cm:a/b/c.sml
1354    
1355    This path denotes the file bar/a/b/c.sml relative to the directory
1356    denoted by anchor "foo".  Notice that the encoding also includes
1357    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1358    paths are resolved relative to the .cm-files directory, so baz.cm must
1359    be ignored to get the "real" pathname.
1360    
1361    To make this fact more obvious, CM puts the names of such "virtual
1362    arcs" into parentheses when they appear in progress reports. (No
1363    parentheses will appear in the internal protocol encoding.)  Thus,
1364    what you really see is:
1365    
1366      $foo/bar/(baz.cm):a/b/c.sml
1367    
1368    I find this notation to be much more informative than before.
1369    
1370    Another new feature of the encoding is that special characters
1371    including parentheses, colons, (back)slashes, and white space are
1372    written as \ddd (where ddd is the decimal encoding of the character).
1373    
1374    *** The CM manual still needs to be updated.
1375    
1376    ----------------------------------------------------------------------
1377    Name: Allen Leung
1378    Date: 2000/06/15 00:38:00
1379    Tag: leunga-20000615-x86-peephole
1380    
1381    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1382    
1383    ----------------------------------------------------------------------
1384    Name: Matthias Blume
1385    Date: 2000/06/12 11:40:00
1386    Tag: blume-20000612-parmakefix
1387    Description:
1388    
1389    More cleanup after changing the file naming scheme: This time I
1390    repaired the parallel make mechanism for CMB.make which I broke earlier.
1391    
1392    ----------------------------------------------------------------------
1393    Name: Allen Leung
1394    Date: 2000/06/09 01:25:00
1395    Tag: leunga-20000609-various
1396    
1397    None of these things should affect normal SML/NJ operations
1398    
1399    1. Peephole improvements provided by Fermin (c--)
1400    2. New annotation DEFUSE for adding extra dependence (moby)
1401    3. New X86 LOCK instructions (moby)
1402    4. New machine description language for reservation tables (scheduling)
1403    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1404       trees etc.)
1405    6. I've changed the CM files so that they can work with versions
1406       110.0.6, 110.25 and 110.28
1407    
1408    ----------------------------------------------------------------------
1409    Name: Matthias Blume
1410    Date: 2000/06/09 12:40:00
1411    Tag: blume-20000609-log
1412    Description:
1413    
1414    - Removed all(?) remaining RCS Log entries from sources.
1415    
1416    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1417      anchored paths).
1418    
1419    ----------------------------------------------------------------------
1420    Name: Matthias Blume
1421    Date: 2000/06/07 17:00:00 JST
1422    Tag: blume-20000607-no-implicit-anchors
1423    Description:
1424    
1425    1. This update changes the default setting for
1426    CM.Control.implicit_anchors from true to false.  This means that
1427    implicit anchors are no longer permitted by default.  I also tried to
1428    make sure that nothing else still relies on implicit anchors.
1429    (This is the next step on the schedule towards a CM that does not even
1430    have the notion of implicit anchors anymore.)
1431    
1432    2. More CM manual updates.
1433    
1434    3. I managed to track down and fix the pickling bug I mentioned last
1435    time.  Because of the previously existing workaround, this entails no
1436    immediate practical changes.
1437    
1438    ----------------------------------------------------------------------
1439    Name: Matthias Blume
1440    Date: 2000/06/06 11:15:00 JST
1441    Tag: blume-20000606-lazierpickle
1442    Description:
1443    
1444    !!!! NEW BOOT FILES !!!!
1445    
1446    * The main purpose of this update is to make library pickles lazier in
1447    order to reduce the initial space penalty for autoloading a library.
1448    As a result, it is now possible to have $smlnj/compiler.cm
1449    pre-registered.  This should take care of the many complaints or
1450    inquiries about missing structure Compiler.  This required changes to
1451    CM's internal data structures and small tweaks to some algorithms.
1452    
1453    As a neat additional effect, it is no longer necessary (for the sake
1454    of lean heap image files) to distinguish between a "minimal" CM and a
1455    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1456    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1457    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1458    
1459    To make the life of compiler-hackers easier, "makeml" now also
1460    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1461    words, after you bootstrap a new sml for the first time, you will not
1462    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1463    around you will still have to do it, though.)
1464    
1465    * A second change consists of major updates to the CM manual.  There
1466    are now several appendices with summary information and also a full
1467    specification of the CM description file syntax.
1468    
1469    * In directory src/system I added the script "allcross".  This script
1470    invokes sml and cross-compiles the compiler for all supported
1471    architectures.  (Useful when providing a new set of boot files.)
1472    
1473    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1474    added a small tweak to pickle-util.sml to work around this problem,
1475    but it is not a proper fix yet.  I will investigate further.  (The
1476    effect of the bug was an inflation of library pickle size.)
1477    
1478    * Version number increased to 110.28.1 (to avoid compatibility problems).
1479    
1480    ----------------------------------------------------------------------
1481    Name: Allen Leung
1482    Date: 2000/05/25 17:28 EDT
1483    Tag: leunga-20000525-ra
1484    Description:
1485    
1486      Fixed a bug in freezing phase of the register allocator.
1487    
1488    ----------------------------------------------------------------------
1489    Name: Allen Leung
1490    Date: 2000/05/15 22:53 EDT
1491    Tag: leunga-20000515-alpha-x86-ra
1492    Description:
1493    
1494      1. Alpha
1495    
1496          Slight cleanup.  Removed the instruction SGNXL
1497    
1498      2. X86
1499    
1500          Added the following instructions to the instruction set:
1501    
1502            ROLx, RORx,
1503            BTx, BTSx, BTLx, BTRx,
1504            XCHGx, and variants with the LOCK prefix
1505    
1506      3. Register Allocation
1507    
1508          The module ra-rewrite-with-renaming has been improved.
1509    
1510      These have no effect on SML/NJ.
1511    
1512    ----------------------------------------------------------------------
1513    Name: Matthias Blume
1514    Date: 2000/05/15 16:20:00 JST
1515    Tag: blume-20000515-lightrebuild
1516    Description:
1517    
1518    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1519       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1520       defined.  In effect, the command will not build any cross-compiler
1521       backends and therefore finish more quickly.
1522    
1523       The "fixpt" script also takes a "-light" switch to be able to use
1524       this new facility while compiling for a fixpoint.
1525    
1526    2. I replaced all mentions of anchored paths in group owner specifications
1527       with simple relative paths (usually starting with "..").
1528       The rationale is that a library's internal workings should not be
1529       compromised by the lack of some anchor.  (An anchor is necessary
1530       for someone who wants to refer to the library by an anchored path,
1531       but it should not be necessary to build the same library in the first
1532       place.)
1533    
1534    3. I changed the way CM's tool mechanism determines the shell command
1535       string used for things like ml-yacc etc. so that it does not break
1536       when CM.Control.implicit_anchors is turned off.
1537    
1538    ----------------------------------------------------------------------
1539    Name: Matthias Blume
1540    Date: 2000/05/12 18:20:00 JST
1541    Tag: blume-20000512-ml-build
1542    Description:
1543    
1544    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1545    getting installed properly (by config/install.sh).
1546    
1547    ----------------------------------------------------------------------
1548    Name: Matthias Blume
1549    Date: 2000/05/12 17:30:00 JST
1550    Tag: blume-20000512-anchors
1551    Description:
1552    
1553    !!! NEW BOOT FILES !!!
1554    
1555    This change is in preparation of fading out support for "implicitly
1556    anchored path names".  I went through all sources and used the
1557    explicit (and relatively new) $-notation.  See system/README and the
1558    CM manual for more info on this.
1559    
1560    I also modified the anchoring scheme for some things such as "smlnj",
1561    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1562    anchors are more expressive: anchor name and first arc do not have to
1563    coincide.  This entails the following user-visible change:
1564    
1565    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1566    particular, when you fire up sml with a command-line argument, say,
1567    e.g.:
1568    
1569       sml '$smlnj/cmb.cm'
1570    
1571    At the ML toplevel prompt:
1572    
1573       CM.autoload "$smlnj/cmb.cm";
1574    
1575    There is also a new controller in CM.Control that can be used to turn
1576    off all remaining support for implicit anchors by saying:
1577    
1578        CM.autoload "$smlnj/
1579        #set CM.Control.implicit_anchors false;
1580    
1581    This causes CM to reject implicitly anchored paths.  This is (for the
1582    time being) less permissive than the "final" version where there will
1583    be no more such implicit anchors and relative paths will be just that:
1584    relative.
1585    
1586    The next step (version after next version?) will be to make the
1587    default for CM.Control.implicit_anchors false.  After the dust has
1588    settled, I can then produce the "final" version of this...
1589    
1590    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1591    
1592    ----------------------------------------------------------------------
1593    Name: Matthias Blume
1594    Date: 2000/05/11 16:30:00 JST
1595    Tag: blume-20000511-sources
1596    Description:
1597    
1598    The main change is that I added function CM.sources as a generalized
1599    version of the earlier CM.makedepend.  This entails the following
1600    additional changes:
1601    
1602      - CM.makedepend has been dropped.
1603    
1604      - CM manual has been updated.
1605    
1606      - TOOLS signature and API have been changed.
1607    
1608    ----------------------------------------------------------------------
1609    Name: Allen Leung
1610    Date: 2000/05/10 21:17 EDT
1611    Tag: leunga-20000510-moby-c--ssa
1612    Description:
1613    
1614      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1615    None of these affect SML/NJ.
1616    
1617    1. Register Allocation
1618    
1619        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1620           This module tries to remove local (i.e. basic block level) redundancies
1621           during spilling.
1622    
1623        b. A new framework for performing region based register allocation.
1624           Not yet entirely functional.
1625    
1626    2. X86
1627    
1628       a. DefUse for POP was missing the stack pointer [found by Lal]
1629       b. Reload for CALL was incorrect in X86Spill [found by John]
1630       c. Various fixes in X86Spill so that it can be used correctly for
1631          the new spilling module.
1632    
1633    3. SSA/IR
1634    
1635       a. New module ir/dj-dataflow.sml implements elimination based
1636          data flow analysis.
1637    
1638    4. MLRiscGen
1639    
1640       a. Fix for gc type annotation
1641    
1642    5. MDGen
1643    
1644       Various fixes for machine description -> ml code translation.  For ssa
1645       only.
1646    
1647    ----------------------------------------------------------------------
1648    Name: Allen Leung
1649    Date: 2000/05/08 22:17 EDT
1650    Tag: leunga-20000508-labexp
1651    Description:
1652    
1653      Fermin has found a few assembly problems with constant expressions
1654      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1655      which choke on dumb assemblers.  This is his fix.
1656    
1657    ----------------------------------------------------------------------
1658    Name: Dave MacQueen
1659    Date: 2000/04/09 14:00 EDT
1660    Tag: dbm-20000502-Version_110_28
1661    Description:
1662    
1663    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1664    
1665    2. Updated config/version to 110.28
1666    
1667    3. Updated config/srcarchiveurl
1668    
1669    3. New boot files!
1670       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1671    
1672    ----------------------------------------------------------------------
1673    Name: Matthias Blume
1674  Date: 2000/05/01 19:05:00 JST  Date: 2000/05/01 19:05:00 JST
1675  Tag: blume-20000501-noweb  Tag: blume-20000501-noweb
1676  Description:  Description:
# Line 310  Line 1968 
1968    
1969  3.  Assembly  3.  Assembly
1970    
1971      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
1972      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
1973      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
1974    
# Line 333  Line 1991 
1991    
1992  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
1993  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
1994  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
1995  legal surface-syntax symbol).  legal surface-syntax symbol).
1996    
1997  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 343  Line 2001 
2001  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
2002    
2003  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2004  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
2005  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
2006  more details.  more details.
2007    
# Line 458  Line 2116 
2116     (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
2117     used).     used).
2118    
2119     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
2120     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
2121     Makefiles or Shell scripts.     Makefiles or Shell scripts.
2122    
2123     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 471  Line 2129 
2129    
2130     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
2131    
2132     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
2133     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
2134    
2135  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 605  Line 2263 
2263    
2264       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
2265       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
2266       testl is perferable to andl.       testl is preferable to andl.
2267    
2268  7. RA (x86 only)  7. RA (x86 only)
2269    
# Line 787  Line 2445 
2445    
2446  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2447  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2448  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
2449  explanation.  explanation.
2450    
2451  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2452  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2453  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2454  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 834  Line 2492 
2492  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
2493  <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,
2494  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
2495  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2496  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2497  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2498    
# Line 884  Line 2542 
2542    it from that remote directory.    it from that remote directory.
2543    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2544    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2545    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2546    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2547    
2548    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.634  
changed lines
  Added in v.818

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