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 631, Fri Apr 28 08:30:52 2000 UTC revision 839, Thu Jun 7 20:28:44 2001 UTC
# Line 11  Line 11 
11  Date:  Date:
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15    ----------------------------------------------------------------------
16    Name: Matthias Blume
17    Date: 2001/06/07
18    Tag: blume-20010607-calls
19    Description:
20    
21    A number of internal changes related to C calls and calling conventions:
22    
23    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
24       number of bytes popped implicitly (by the callee).  In most cases
25       this field is 0 but on x86/win32 it is some non-zero value.  This
26       is information provided for the benefit of the "omit-frameptr" pass.
27    2. The CALL instruction on the x86 carries a similar "pops" field.
28       The instruction selection phase copies its value from the ML-Tree
29       CALL statement.
30    3. On all other architectures, the instruction selection phase checks
31       whether "pops=0" and complains if not.
32    4. The c-calls implementation for x86 now accepts two calling conventions:
33       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
34       up after the call and pops is set to 0.  For "stdcall", the caller
35       does nothing, leaving the cleanup to the callee; pops is set to
36       the number of bytes that were pushed onto the stack.
37    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
38       distinguish between "ccall" and "stdcall".
39    6. The UNIMP instruction has been added to the supported Sparc instruction
40       set. (This is needed for implementing the official C calling convention
41       on this architecture.)
42    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
43       work with the latest CM.
44    
45    ----------------------------------------------------------------------
46    Name: Matthias Blume
47    Date: 2001/06/05 15:10:00 EDT
48    Tag: blume-20010605-cm-index
49    Description:
50    
51    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
52    
53    1. CM can now generate "index files".  These are human-readable files
54       that list on a per-.cm-file basis each toplevel symbol defined or
55       imported.  The location of the index file for
56       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
57       To enable index-file generation, set CM.Control.generate_index to true
58       or export an environment-symbol: export CM_GENERATE_INDEX=true.
59    
60       The CM manual has been updated accordingly.
61    
62    2. I made some slight modifications to the c-calls API in MLRISC.
63    
64         a) There is now a callback to support saving/restoring of
65            dedicated but caller-save registers around the actual call
66            instruction.
67         b) One can optionally specify a comment-annotation for the
68            call instruction.
69    
70    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
71       (For example, the comment annotation shows the C prototype of
72        the function being called.)
73    
74    ----------------------------------------------------------------------
75    Name: Matthias Blume
76    Date: 2001/06/01 13:30:00 EDT
77    Tag: blume-20010601-nlffi-cleanup
78    Description:
79    
80    This is mostly a cleanup of MLFFI stuff:
81    
82       - some signature files have been put into a more exposed place
83       - the ugly 'f type parameter is gone (simplifies types tremendously!)
84       - ml-nlffigen changed accordingly
85       - tutorial updated
86    
87    Other changes:
88    
89       - author's affiliation in CM manual(s) updated
90       - some more recognized keywords added to Allen's sml.sty
91    
92    ----------------------------------------------------------------------
93    Name: Matthias Blume
94    Date: 2001/05/25 15:30:00 EDT
95    Tag: blume-20010525-iptr
96    Description:
97    
98      - put the official 110.33-README (as it appears on the ftp server) under
99        CVS
100      - fixed a small bug related to incomplete pointer types in
101        ml-nlffigen
102      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
103        (it does not need the 'f type parameter)
104    
105    ----------------------------------------------------------------------
106    Name: Matthias Blume
107    Date: 2001/05/23 14:30:00 EDT
108    Tag: Release_110_33
109    Description:
110    
111    New version number (110.33).  New bootfiles.
112    
113    ----------------------------------------------------------------------
114    Name: Matthias Blume
115    Date: 2001/05/22 18:06:00 EDT
116    Tag: blume-20010522-targets
117    Description:
118    
119    Made install.sh use file config/targets.customized if it exists, falling
120    back to config/targets if it doesn't.  This way one can have a customized
121    version of the targets file without touching the "real thing", thus
122    eliminating the constant fear of accidentally checking something bogus
123    back into the CVS repository...  (File config/targets.customized must
124    not be added to the repository!)
125    
126    ----------------------------------------------------------------------
127    Name: Matthias Blume
128    Date: 2001/05/22 16:30:00 EDT
129    Tag: blume-20010522-minitut
130    Description:
131    
132    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
133       struct returns.
134    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
135       incomplete, preliminary documentation for NLFFI.
136    
137    ----------------------------------------------------------------------
138    Name: Matthias Blume
139    Date: 2001/05/14 11:30:00 EDT
140    Tag: blume-20010514-script
141    Description:
142    
143    Some bugs in install script fixed.
144    
145    In addition to that I also made a slight change to the NLFFI API:
146    Functors generated by ml-nlffigen now take the dynamic library as a
147    straight functor argument, not as a suspended one.  (The original
148    functor code used to force the suspension right away anyway, so there
149    was nothing gained by this complication of the interface.)
150    
151    ----------------------------------------------------------------------
152    Name: Matthias Blume
153    Date: 2001/05/11 14:35:00 EDT
154    Tag: blume-20010511-ml-nlffi
155    Description:
156    
157    I finally took the plunge and added my new FFI code to the main
158    repository.  For x86-linux it is now ready for prime-time.
159    
160    There are two new subdirectories of "src":
161    
162      - ml-nlffi-lib:
163           The utility library for programs using the FFI interface.
164           Here is the implementation of $/c.cm and its associated low-level
165           partners $/c-int.cm and $/memory.cm.
166      - ml-nlffigen:
167           A stand-alone program for generating ML glue code from C source
168           code.
169    
170    Building ml-nlffigen requires $/ckit-lib.cm.
171    
172    The config/install.sh script has been updates to do the Right Thing
173    (hopefully).
174    
175    Notice that the source tree for the C-Kit will not be put under "src"
176    but directly under the installation root directory.  (This is the
177    structure that currently exists on the CVS server when you check out
178    module "sml".)  Fortunately, config/install.sh knows about this oddity.
179    
180    Bugs: No documentation yet.
181    
182    ----------------------------------------------------------------------
183    Name: Matthias Blume
184    Date: 2001/05/09 16:35:00 EDT
185    Tag: blume-20010509-cpscontract
186    Description:
187    
188    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
189    wrapper/unwrapper elimination did not decrement usage counts and some
190    dead variables got overlooked by the dead-up logic.)
191    
192    ----------------------------------------------------------------------
193    Name: Lal George
194    Date: 2001/05/08  17:26:09 EDT
195    Tag: george-20010508-omit-frameptr
196    Description:
197    
198    Changes to implement the omit-frame-pointer optimization to support
199    raw C calls. For now, there is only support on the Intel x86, but
200    other architectures will follow as more experience is gained with this.
201    
202    
203    ----------------------------------------------------------------------
204    Name: Matthias Blume
205    Date: 2001/05/07 14:40:00 EDT
206    Tag: blume-20010507-proxies
207    Description:
208    
209    I made into "proxy libraries" all libraries that qualify for such a
210    change.  (A qualifying library is a library that has another library or
211    groups as its sole member and repeats that member's export list
212    verbatim.  A proxy library avoids this repetition by omitting its export
213    list, effectively inheriting the list that its (only) member exports.
214    See the CM manual for more explanation.)
215    The main effect is that explicit export lists for these libraries
216    do not have to be kepts in sync, making maintenance a bit easier.
217    
218    I also added copyright notices to many .cm-files.
219    
220    Last but not least, I made a new set of bootfiles.
221    
222    ----------------------------------------------------------------------
223    Name: Matthias Blume
224    Date: 2001/05/04 17:00:00 EDT
225    Tag: blume-20010504-cm-lsplit
226    Description:
227    
228    0. John merged pending changes to $/smlnj-lib.cm
229    
230    1. Allen's previous change accidentally backed out of one of Lal's
231       earlier changes.  I undid this mistake (re-introducing Lal's change).
232    
233    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
234       within the compiler where applicable.  There is some code simplification
235       because of that.
236    
237    3. The "split" phase (in FLINT) is now part of the default list of phases.
238       Compiler.Control.LambdaSplitting.* can be used to globally control the
239       lambda-splitting (cross-module-inlining) engine.  In addition to that,
240       it can now also be controlled on a per-source basis: CM has been taught
241       a new tool parameter applicable to ML source files.
242    
243       - To turn lambda-splitting off completely:
244            local open Compiler.Control.LambdaSplitting in
245                val _ = set Off
246            end
247       - To make "no lambda-splitting" the global default (but allow per-source
248         overriding); this is the initial setting:
249            local open Compiler.Control.LambdaSplitting in
250                val _ = set (Default NONE)
251            end
252       - To make "lambda-splitting with aggressiveness a" the global default
253         (and allow per-source overriding):
254            local open Compiler.Control.LambdaSplitting in
255                val _ = set (Default (SOME a))
256            end
257    
258       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
259         write (in the respective .cm-file):
260            a.sml (lambdasplitting:off)
261       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
262            a.sml (lambdasplitting:on)
263       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
264         <a> is a decimal non-negative integer):
265            a.sml (lambdasplitting:<a>)
266       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
267            a.sml (lambdasplitting:infinity)
268       - To use the global default for a.sml:
269            a.sml (lambdasplitting:default)
270         or simply
271            a.sml
272    
273    ----------------------------------------------------------------------
274    Name: Allen Leung
275    Date: 2001/05/04 01:57:00 EDT
276    Tag: leunga-20010504-sync
277    Description:
278    
279      MLRISC features.
280    
281      1. Fix to CMPXCHG instructions.
282      2. Changed RA interface to allow annotations in callbacks.
283      3. Added a new method to the stream interface to allow annotations updates.
284    
285    ----------------------------------------------------------------------
286    Name: Matthias Blume
287    Date: 2001/05/01 11:45:00 EDT
288    Tag: blume-20010501-pcedittmp
289    Description:
290    
291    Changed install.sh to use the current working directory instead of
292    /usr/tmp for a temporary file (pcedittmp).  The previous choice
293    of /usr/tmp caused trouble with MacOS X because of file premission
294    problems.
295    
296    ----------------------------------------------------------------------
297    Name: Matthias Blume
298    Date: 2001/04/20 11:10:00 EDT
299    Tag: blume-20010420-inMLflag
300    Description:
301    
302     - added vp_limitPtrMask to vproc-state.h
303       (for use by the raw-C-calls mechanism to implement proper interrupt
304        handling)
305     - made the ML compiler aware of various data-structure offsets so it
306       can generate code for accessing the vp_inML flag and vp_limitPtrMask
307     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
308       raw C-calls
309    
310    ----------------------------------------------------------------------
311    Name: Lal George
312    Date: 2001/04/20 09:15:28 EDT
313    Tag: george-20010420-macosX
314    Description:
315    
316     - Changes to port to Mac OS X; Darwin.
317    
318     - In the process I found that sqrt was broken on the PPC, because the
319       fsqrt instruction is not implemented.
320    
321    ----------------------------------------------------------------------
322    Name: Matthias Blume
323    Date: 2001/04/18 12:45:00 EDT
324    Tag: blume-20010418-ccalls
325    Description:
326    
327     - fixed two off-by-4 errors in the x86-specific c-calls implementation
328       (this bug prevented structure arguments containing pointers from being
329        passed correctly)
330     - changed the raw-C-call code in mlriscGen.sml in such a way that
331       structure arguments are represented as a pointer to the beginning
332       of the structure (instead of having a series of synthesized arguments,
333       one for each structure member)
334    
335     - made makeml script's verbosity level configurable via environment
336       variable (MAKEML_VERBOSITY)
337    
338     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
339       in rawmem-x86.sml; we are now using the real thing
340    
341    ----------------------------------------------------------------------
342    Name: Matthias Blume
343    Date: 2001/03/22 16:25:00 EST
344    Tag: blume-20010322-bootfiles
345    Description:
346    
347    Created a new set of bootfiles (for your automatic installation convenience).
348    
349    ----------------------------------------------------------------------
350    Name: Matthias Blume
351    Date: 2001/03/22 15:10:00 EST
352    Tag: blume-20010322-rawmem-parcm
353    Description:
354    
355    1. All "raw memory access" primitives for the new FFI are implemented now
356       (at least on the x86).
357    2. Some further cleanup of CM's parallel make mechanism.
358    
359    ----------------------------------------------------------------------
360    Name: Matthias Blume
361    Date: 2001/03/19 17:53:00 EST
362    Tag: blume-20010319-parallel
363    Description:
364    
365    Parallel make (using compile servers) now works again.
366    
367    To this end, CM.stabilize and CMB.make have been modified to work in
368    two passes when compile servers are attached:
369       1. Compile everything, do not perform stabilization; this pass
370          uses compile servers
371       2. Stabilize everything; this pass does not use compile servers
372    If there are no compile servers, the two passes are combined into one
373    (as before).  Splitting the passes increases the inherent parallelism
374    in the dependency graph because the entire graph including all
375    libraries is available at the same time.  This, in turn, improves
376    server utilization.  The downside is that the master process will
377    have to do some extra work after compilation is done (because for
378    technical reasons it must re-read all the binfiles during stabilization).
379    
380    ----------------------------------------------------------------------
381    Name: Matthias Blume
382    Date: 2001/03/16 12:22:00 EST
383    Tag: blume-20010316-bootfiles
384    Description:
385    
386    Created a new set of bootfiles (for your automatic installation convenience).
387    
388    ----------------------------------------------------------------------
389    Name: Matthias Blume
390    Date: 2001/03/16 11:00:00 EST
391    Tag: blume-20010316-MLTREE-fixup
392    Description:
393    
394    This is a minor fixup for an (untagged) earlier commit by Allen.
395    (A file was missing).
396    
397    ----------------------------------------------------------------------
398    Name: Allen Leung
399    Date: Mon Mar  5 18:54:57 EST 2001
400    Tag: leunga-20010305-cut-support
401    
402    1. New support for alternative control-flow in MLTREE.
403       Currently we support
404    
405          FLOW_TO(CALL ...., [k1,...,kn])
406    
407       This is needed for 'cuts to' in C-- and try/handle-like constructs
408       in Moby
409    
410       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
411    
412    2. Register Allocator
413    
414       Changes in interface [from Fermin, John]
415    
416    3. Alpha 8-bit SLL support [Fermin]
417    
418    4. All architectures
419    
420       A new module (ClusterExpandCopies) for expanding parallel copies.
421    
422    ----------------------------------------------------------------------
423    Name: Allen Leung
424    Date: 2001/02/27 23:07:00 EST
425    Tag: leunga-20010227-minor-stuff
426    
427    1. Alpha bug fix for CMOVNE
428    2. Handle mltree COND(..,FCMP ...,...)
429    3. Bug fix in simplifier
430    
431    ----------------------------------------------------------------------
432    Name: Matthias Blume
433    Date: 2001/01/30 17:50:00 EST
434    Tag: blume-20010130-sync
435    Description:
436    
437    This is just a minor update to sync my devel branch with the main brach.
438    The only visible change is the addition of some README files.
439    
440    ----------------------------------------------------------------------
441    Name: Matthias Blume
442    Date: 2001/01/12 23:30:00 JST
443    Tag: blume-20010112-bootfiles
444    Description:
445    
446    Made a new set of bootfiles that goes with the current state of the
447    repository.
448    
449    ----------------------------------------------------------------------
450    Name: Matthias Blume
451    Date: 2001/01/12 21:20:00 JST
452    Tag: blume-20010112-sync
453    Description:
454    
455    I am just flushing out some minor changes that had accumulated in
456    my private branch in order to sync with the main tree.  (This is
457    mainly because I had CVS trouble when trying to merge _into_ my
458    private branch.)
459    
460    Most people should be completely unaffected by this.
461    
462    ----------------------------------------------------------------------
463    Name: Allen Leung
464    Date: Thu Jan 11 21:03:00 EST 2001
465    Tag: leunga-20010111-labexp=mltree
466    Description:
467    
468    1.  Removed the type LabelExp and replace it by MLTree.
469    2.  Rewritten mltree-simplify with the pattern matcher tool.
470    3.  There were some bugs in alpha code generator which would break
471        64-bit code generation.
472    4.  Redo the tools to generate code with the
473    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
474        are now generated by perl scripts.
475    
476    ----------------------------------------------------------------------
477    Name: Matthias Blume
478    Date: 2001/01/10 21:55:00 JST
479    Tag: blume-20010110-rcc
480    Description:
481    
482    The RCC stuff now seems to work (but only on the x86).
483    This required hacking of the c-calls interface (and -implementation) in
484    MLRISC.
485    
486    Normal compiler users should be unaffected.
487    
488    ----------------------------------------------------------------------
489    Name: Matthias Blume
490    Date: 2001/01/09 01:20:00 JST
491    Tag: blume-20010109-rcc
492    Description:
493    
494    This is a fairly big patch, flushing out a large number of pending
495    changes that I made to my development copy over the last couple of days.
496    
497    Of practical relevance at this moment is a workaround for a pickling
498    bug that Allen ran into the other day.  The cause of the bug itself is
499    still unknown and it might be hard to fix it properly, but the
500    workaround has some merits of its own (namely somewhat reducing pickling
501    overhead for certain libraries).  Therefore, I think this solution should
502    be satisfactory at this time.
503    
504    The rest of the changes (i.e., the vast majority) has to do with my
505    ongoing efforts of providing direct support for C function calls from
506    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
507    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
508    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
509    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
510    handle RCC.
511    
512    The last part (the changes to mlriscGen.sml) are still known to be
513    wrong on the x86 and not implemented on all other architectures.  But
514    the infrastructure is in place. I had to change a few functor
515    signatures in the backend to be able to route the CCalls interface
516    from MLRISC there, and I had to specialize the mltree type (on the
517    x86) to include the necessary extensions. (The extensions themselves
518    were already there and redy to go in MLRISC/x86).
519    
520    Everything should be very happy as soon as someone helps me with
521    mlriscGen.sml...
522    
523    In any case, nothing of this should matter to anyone as long as the
524    new primop is not being used (which is going to be the case unless you
525    find it where I hid it :). The rest of the compiler is completely
526    unaffected.
527    
528    ----------------------------------------------------------------------
529    Name: Matthias Blume
530    Date: 2001/01/05 00:30:00 JST
531    Tag: blume-20010105-primops
532    Description:
533    
534    Added some experimental support for work that I am doing right now.
535    These changes mostly concern added primops, but there is also a new
536    experimental C library in the runtime system (but currently not enabled
537    anywhere except on Linux/X86).
538    
539    In the course of adding primops (and playing with them), I discovered that
540    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
541    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
542    To recover, I made type information in INL_PRIM mandatory and changed
543    prim.sml as well as built-in.sml accordingly.  The InLine structure now
544    has complete, correct type information (i.e., no bottom types).
545    
546    Since all these changes mean that we need new binfiles, I also bumped the
547    version number to 110.32.1.
548    
549    ----------------------------------------------------------------------
550    Name: Matthias Blume
551    Date: 2000/12/30 22:10:00 JST
552    Tag: blume-20001230-various
553    Description:
554    
555    Added proxy libraries for MLRISC and let MLRISC libraries refer
556    to each other using path anchors.  (See CM manual for explanation.)
557    
558    Updated CM documentation.
559    
560    Fixed some bugs in CM.
561    
562    Implemented "proxy" libraries (= syntactic sugar for CM).
563    
564    Added "-quiet" option to makeml and changed runtime system accordingly.
565    
566    Added cleanup handler for exportML to reset timers and compiler stats.
567    
568    ----------------------------------------------------------------------
569    Name: Lal George
570    Date: 2000/12/22 22:22:58 EST 2000
571    Tag: Release_110_32
572    Description:
573    
574            Infinite precision used throughout MLRISC.
575            see MLRISC/mltree/machine-int.sig
576    
577    ----------------------------------------------------------------------
578    Name: Matthias Blume
579    Date: 2000/12/22 23:16:00 JST
580    Tag: blume-20001222-warn
581    Description:
582    
583    Corrected wording and formatting of some CM warning message which I
584    broke in my previous patch.
585    
586    ----------------------------------------------------------------------
587    Name: Matthias Blume
588    Date: 2000/12/22 21:20:00 JST
589    Tag: blume-20001222-anchorenv
590    Description:
591    
592    Fixed CM's handling of anchor environments in connection with CMB.make.
593    
594    ----------------------------------------------------------------------
595    Name: Matthias Blume
596    Date: 2000/12/22 13:15:00 JST
597    Tag: blume-20001222-cleanup
598    Description:
599    
600    Removed src/cm/ffi which does not (and did not) belong here.
601    
602    ----------------------------------------------------------------------
603    Name: Matthias Blume
604    Date: 2000/12/21 23:55:00 JST
605    Tag: blume-20001221-exn
606    Description:
607    
608    Probably most important: CM no longer silently swallows all exceptions
609    in the compiler.
610    Plus: some other minor CM changes.  For example, CM now reports some
611    sizes for generated binfiles (code, data, envpickle, lambdapickle).
612    
613    ----------------------------------------------------------------------
614    Name: Matthias Blume
615    Date: 2000/12/15 00:01:05 JST
616    Tag: blume-20001215-dirtool
617    Description:
618    
619    - "dir" tool added.
620    - improvements and cleanup to Tools structure
621    - documentation updates
622    
623    ----------------------------------------------------------------------
624    Name: Allen Leung
625    Date: Thu Dec 14 03:45:24 EST 2000
626    Description:
627    Tag:  leunga-20001214-int-inf
628    Description:
629    
630       In IntInf, added these standard functions, which are missing from our
631    implementation:
632    
633        andb : int * int -> int
634        xorb : int * int -> int
635        orb  : int * int -> int
636        notb : int -> int
637         <<   : int * word -> int
638        ~>>  : int * word -> int
639    
640       Not tested, I hope they are correct.
641    
642    ----------------------------------------------------------------------
643    Name: Allen Leung
644    Date: Fri Dec  8 19:23:26 EST 2000
645    Description:
646    Tag:  leunga-20001208-nowhere
647    Description:
648    
649      Slight improvements to the 'nowhere' tool to handle OR-patterns,
650    to generate better error messages etc.  Plus a brief manual.
651    
652    ----------------------------------------------------------------------
653    Name: Lal George
654    Date: 2000/12/08 09:54:02 EST 2000
655    Tag: Release_110_31
656    Description:
657    
658    - Version 110.31
659    ----------------------------------------------------------------------
660    Name: Allen Leung
661    Date: Thu Dec  7 22:01:04 EST 2000
662    Tag:  leunga-20001207-cell-monster-hack
663    Description:
664    
665    Major MLRISC internal changes.  Affect all clients.
666    Summary:
667    
668    1.  Type CELLS.cell = int is now replaced by a datatype.
669        As a result, the old regmap is now gone.  Almost all interfaces
670        in MLRISC change as a consequence.
671    
672    2.  A new brand version of machine description tool (v3.0) that generates
673        modules expecting the new interface.  The old version is removed.
674    
675    3.  The RA interface has been further abstracted into two new functors.
676        RISC_RA and X86RA.  These functors have much simpler interfaces.
677        [See also directory MLRISC/demo.]
678    
679    4.  Some other new source->source code generation tools are available:
680    
681        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
682        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
683           I use this tool to generate the peephole optimizers---with the new
684           cell type changes, peephole rules are becoming difficult to write
685           without conditional pattern matching.
686    
687    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
688        cover the entire MLRISC source tree so many things broke.
689    
690    6.  CM files have been moved to the subdirectory MLRISC/cm.
691        They are moved because there are a lot of them and they clutter up the
692        root dir.
693    
694    7.  More detailed documentation to come...
695    
696        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
697        a makeml -rebuild first.  This is because of other other
698        changes that Matthias has made (see below).
699    
700    
701    ----------------------------------------------------------------------
702    Name: Matthias Blume
703    Date: 2000/11/30 23:12:00 JST
704    Tag: blume-20001130-filereorg
705    Description:
706    
707    Some manual updates and some file reorganizations in CM.
708    
709    ----------------------------------------------------------------------
710    Name: Matthias Blume
711    Date: 2000/11/24 17:45:00 JST
712    Tag: blume-20001124-link
713    Description:
714    
715    Drastically improved link traversal code for the case that the dynamic
716    value was already loaded at bootstrap time.  As a result, CM and CMB
717    now both load blazingly fast -- even on a very slow machine.  Also,
718    memory consumption has been further reduced by this.
719    
720    Warning: The format of the PIDMAP file has changed.  THerefore, to
721    bootstrap you have to do this:
722    
723    1. Run CMB.make
724    2. Make a symbolic link for the boot directory:
725         ln -s sml.boot.ARCH-OS xxx
726    3. "Rebuild" the boot directory:
727         ./makeml -boot xxx -rebuild sml ; rm xxx
728    4. Boot normally:
729          ./makeml
730    
731    ----------------------------------------------------------------------
732    Name: Matthias Blume
733    Date: 2000/11/21 21:20:00 JST
734    Tag: blume-20001121-tools
735    Description:
736    
737    Continued hacking on autoloading problem -- with success this time.
738    Also changed tool-plugin mechanism.  See new CM manual.
739    
740    ----------------------------------------------------------------------
741    Name: Matthias Blume
742    Date: 2000/11/19 14:30:00 JST
743    Tag:  blume-20001119-autoload
744    Description:
745    
746    Some hacking to make autoloading faster.  Success for CMB, no success
747    so far for CM.  There is a reduced structure CM' that autoloads faster.
748    (This is a temporary, non-documented hack to be eliminated again when
749    the general problem is solved.)
750    
751    ----------------------------------------------------------------------
752    Name: Matthias Blume
753    Date: 2000/11/17 14:10:00 JST
754    Tag: blume-20001117-pickle-lib
755    Description:
756    
757    1. Eliminated comp-lib.cm
758    2. Made pickle-lib.cm
759    3. Eliminated all uses of intset.sml (from comp-lib.cm)
760    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
761       equivalent constructs from smlnj-lib.cm (INtHashTable).
762    5. Point 4. also goes for those uses of intmap.* in MLRISC.
763       Duplicated intmap modules thrown out.
764    6. Hunted down all duplicated SCC code and replaced it with
765       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
766    7. Rewrote Feedback module.
767    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
768       should be thrown out and equivalent modules from smlnj-lib.cm
769       should be used (IntRedBlackSet, IntListSet, ...).
770    
771    Confirmed that compiler compiles to fixpoint.
772    
773    ----------------------------------------------------------------------
774    Name: Allen Leung
775    Date: 2000/11/10 18:00:00
776    Tag: leunga-20001110-new-x86-fp
777    
778    A new x86 floating point code generator has been added.
779    By default this is turned off.  To turn this on, do:
780    
781        CM.autoload "$smlnj/compiler.cm";
782        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
783    
784    Changes:
785    
786    1.  Changed FTAN to FPTAN so that the assembly output is correct.
787    2.  Changed the extension callback for FTANGENT to generate:
788    
789              fptan
790              fstp  %st(0)
791        instead of
792              fptan
793              fstpl ftempmem
794    
795    3.  Numerous assembly fixes for x86.
796    
797    5.  Cleaned up the machine code output module x86/x86MC.sml and added
798        support for a whole bunch of instructions and addressing modes:
799    
800          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
801          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
802          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
803          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
804          fxch %st(n)
805          fld %st(n)
806          fst %st(n)
807          fst mem
808          fstp %st(n)
809          fucom %st(n)
810          fucomp %st(n)
811    
812        All these are now generated when the fast fp mode is turned on.
813    
814    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
815    
816    ----------------------------------------------------------------------
817    Name: Matthias Blume
818    Date: 2000/11/09 11:20:00 JST
819    Tag: blume-20001109-scc
820    Description:
821    
822    Eliminated some code duplication:
823    
824    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
825       (Otherwise the functor is useless.)
826    2. Used GraphSCCFn where SCCUtilFun was used previously.
827    3. Got rid of SCCUtilFun (in comp-lib.cm).
828    
829    ----------------------------------------------------------------------
830    Name: Lal George
831    Date: 2000/11/06 09:02:21 EST 2000
832    Tag: Release_110_30
833    Description:
834    
835    - Version 110.30
836    ----------------------------------------------------------------------
837    Name: Matthias Blume
838    Date: 2000/11/04 14:45:00
839    Tag: blume-20001104-mlbuild
840    Description:
841    
842    - Made ml-build faster on startup.
843    - Documentation fixes.
844    
845    ----------------------------------------------------------------------
846    Name: Matthias Blume
847    Date: 2000/11/02 17:00:00 JST
848    Tag: blume-20001102-condcomp
849    Description:
850    
851    - Small tweaks to pickler -- new BOOTFILES!
852    - Version bumped to 110.29.2.
853    - Added conditional compilation facility to init.cmi (see comment there).
854    ----------------------------------------------------------------------
855    Name: Allen Leung
856    Date: 2000/10/23 19:31:00
857    Tag: leunga-20001023-demo-ra
858    
859    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
860    2. Test programs for the graph library updated
861    3. Some new MLRISC demo programs added
862    
863    ----------------------------------------------------------------------
864    Name: Matthias Blume
865    Date: 2000/08/31 22:15:00 JST
866    Tag: blume-20001017-errmsg
867    Description:
868    
869    More error message grief: Where there used to be no messages, there
870    now were some that had bogus error regions.  Fixed.
871    
872    ----------------------------------------------------------------------
873    Name: Matthias Blume
874    Date: 2000/08/31 17:30:00 JST
875    Tag: blume-20001017-v110p29p1
876    Description:
877    
878    I made a version 110.29.1 with new bootfiles.
879    
880    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
881              CM documentation changes and a small bugfix in CM's error reporting.
882    
883    ----------------------------------------------------------------------
884    Name: Lal George
885    Date: 2000/09/27 14:42:35 EDT
886    Tag: george-20000927-nodestatus
887    Description:
888    
889    Changed the type of the nodestatus, so that:
890    
891            SPILLED(~1)             is now SPILLED
892            SPILLED(m) where m>=0   is now MEMREG(m)
893            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
894    
895    ----------------------------------------------------------------------
896    Name: Matthias Blume
897    Date: 2000/09/07 14:45:00 JST
898    Tag: blume-20000907-cmerrmsg
899    Description:
900    
901    Small tweak to CM to avoid getting ML syntax error messages twice.
902    
903    ----------------------------------------------------------------------
904    Name: Matthias Blume
905    Date: 2000/08/31 18:00:00 JST
906    Tag: blume-20000831-cvsbootfiles
907    Description:
908    
909    New URL for boot files (because the 110.29 files on the BL server do
910    now work correctly with my updated install scripts for yacc and lex).
911    
912    ----------------------------------------------------------------------
913    Name: Matthias Blume
914    Date: 2000/08/08 12:33:00 JST
915    Tag: blume-20000808-manual
916    Description:
917    
918    Tiny update to CM manual.
919    
920    ----------------------------------------------------------------------
921    Name: Allen Leung
922    Date: 2000/08/7 19:31:00
923    Tag: leunga-20000807-a-whole-bunch-of-stuff
924    
925      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
926    and Mobdy.
927    
928    1.  x86
929    
930       a.  Fixes to peephole module by John and Dan.
931       b.  Assembly fix to SETcc by Allen.
932       c.  Fix to c-call by John.
933       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
934       e.  Instruction selection optimization to SETcc as suggested by John.
935    
936           For example,
937    
938            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
939    
940           should generate:
941    
942            MOVL a, x
943            SUBL b, x
944            SHRL 31, x
945    
946    2.  IR stuff
947    
948         A bunch of new DJ-graph related algorithms added.  These
949         speed up SSA construction.
950    
951    3.  SSA + Scheduling
952    
953         Added code for SSA and scheduling to the repository
954    
955    ----------------------------------------------------------------------
956    Name: Lal George
957    Date: 2000/07/27 11:53:14 EDT
958    
959    Tag: lal-20000727-linux-ppc
960    Description:
961    
962     Made changes to support Linux PPC.
963     p.s. I have confirmation that the 110.29 boot files work fine.
964    
965    ----------------------------------------------------------------------
966    Name: Matthias Blume
967    Date: 2000/07/27 17:40:00 JST
968    Tag: blume-20000727-scripts
969    Description:
970    
971    !!!! WARNING !!!!
972    You must recompile the runtime system!
973    !!!! WARNING !!!!
974    
975    This is basically another round of script-enhancements:
976    
977    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
978       and undefine CM preprocessor symbols.
979    
980    2. ml-build avoids generating a new heap image if it finds that the
981       existing one is still ok.  (The condition is that no ML file had to
982       be recompiled and all ML files are found to be older that the heap
983       file.)
984    
985       To make this work smoothly, I also hacked the runtime system as
986       well as SMLofNJ.SysInfo to get access to the heap image suffix
987       (.sparc-solaris, ...) that is currently being used.
988    
989       Moreover, the signature of CM.mk_standalone has changed.  See the
990       CM manual.
991    
992    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
993       CM manual for details.)
994    
995    4. More CM manual updates:
996        - all of the above has been documented.
997        - there is now a section describing the (CM-related) command line
998          arguments that are accepted by the "sml" command
999    
1000    ----------------------------------------------------------------------
1001    Name: Matthias Blume
1002    Date: 2000/07/25 16:20:00 JST
1003    Tag: blume-20000725-makedepend
1004    Description:
1005    
1006    Added a script called ml-makedepend.  This can be used in makefiles
1007    for Unix' make in a way very similar to the "makedepend" command for
1008    C.
1009    
1010    The script internally uses function CM.sources.
1011    
1012    Synopsis:
1013    
1014        ml-makedepend [-f makefile] cmfile targetname
1015    
1016    The default for the makefile is "makefile" (or "Makefile" should
1017    "makefile" not exist).
1018    
1019    ml-makedepend adds a cmfile/targetname-specific section to this
1020    makefile (after removing the previous version of this section).  The
1021    section contains a single dependency specification with targetname on
1022    the LHS (targetname is an arbitrary name), and a list of files derived
1023    from the cmfile on the RHS.  Some of the files on the RHS are
1024    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
1025    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
1026    path names.  The makefile writer is responsible for making sure that
1027    these variables have correct at the time "make" is invoked.
1028    
1029    ----------------------------------------------------------------------
1030    Name: Matthias Blume
1031    Date: 2000/07/22 23:30:00 JST
1032    Tag: blume-20000722-urlupdate
1033    Description:
1034    
1035    Changed BOOT and config/srcarchiveurl to point to BL server:
1036    
1037        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
1038    
1039    ----------------------------------------------------------------------
1040    Name: Matthias Blume
1041    Date: 2000/07/18 18:00:00 JST
1042    Tag: blume-20000718-Version_110_29
1043    Description:
1044    
1045    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
1046    
1047    2. Updated config/version to 110.29
1048    
1049    3. Updated config/srcarchiveurl
1050    
1051    3. New boot files!
1052       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
1053    
1054    ----------------------------------------------------------------------
1055    Name: Matthias Blume
1056    Date: 2000/07/11 13:58:00 JST
1057    Tag: blume-20000711-doctypo
1058    Description:
1059    
1060    Fixed a few typos in CM manual.
1061    
1062    ----------------------------------------------------------------------
1063    Name: Allen Leung
1064    Date: 2000/06/15 00:38:00
1065    Tag: leunga-20000704-sparc-x86
1066    
1067    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
1068    2. fix to x86 RET bug [found by Dan Grossman]
1069    3. sparc assembly bug fix for ticc instructions [found by Fermin]
1070    
1071       Affects c-- and moby only
1072    
1073    ----------------------------------------------------------------------
1074    Name: Matthias Blume
1075    Date: 2000/07/04 15:26:00
1076    Tag: blume-20000704-trigger
1077    Description:
1078    
1079    1. Improvements to CM manual.
1080    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
1081       of getting a back-trace.  The function, when called, raises an
1082       internal exception which explicitly carries the full back-trace history,
1083       so it is unaffected by any intervening handle-raise pairs ("trivial"
1084       or not).  The interactive loop will print that history once it arrives
1085       at top level.
1086       Short of having all exceptions implicitly carry the full history, the
1087       recommended way of using this facility is:
1088         - compile your program with instrumentation "on"
1089         - run it, when it raises an exception, look at the history
1090         - if the history is "cut off" because of some handler, go and modify
1091           your program so that it explicitly calls BTrace.trigger
1092         - recompile (still instrumented), and rerun; look at the full history
1093    
1094    ----------------------------------------------------------------------
1095    Name: Matthias Blume
1096    Date: 2000/07/03 15:36:00 JST
1097    Tag: blume-20000702-manual
1098    Description:
1099    
1100    Small corrections and updates to CM manual.
1101    
1102    ----------------------------------------------------------------------
1103    Name: Matthias Blume
1104    Date: 2000/06/29 16:04:00 JST
1105    Tag: blume-20000629-yacctool
1106    Description:
1107    
1108    Changes:
1109    
1110    1. Class "mlyacc" now takes separate arguments to pass options to
1111       generated .sml- and .sig-files independently.
1112    2. Corresponding CM manual updates.
1113    3. BTrace module now also reports call sites.  (However, for loop clusters
1114       it only shows from where the cluster was entered.)  There are associated
1115       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1116    
1117    ----------------------------------------------------------------------
1118    Name: Matthias Blume
1119    Date: 2000/06/27 16:51:00 JST
1120    Tag: blume-20000627-noweb
1121    Description:
1122    
1123    Changes:
1124    
1125     1. Implemented "subdir" and "witness" options for noweb tool.
1126        This caused some slight internal changes in CM's tool implementation.
1127     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
1128        some remaining issues from earlier path anchor changes.
1129     3. Updated CM manual accordingly.
1130    
1131     4. Changed implementation of back-tracing so that I now consider it
1132        ready for prime-time.
1133    
1134        In particular, you don't have to explicitly trigger the back-trace
1135        anymore.  Instead, if you are running BTrace-instrumented code and
1136        there is an uncaught exception (regardless of whether or not it was
1137        raised in instrumented code), the top-level evalloop will print
1138        the back-trace.
1139    
1140        Features:
1141    
1142          - Instrumented and uninstrumented code work together seemlessly.
1143            (Of course, uninstrumented code is never mentioned in actual
1144             back-traces.)
1145    
1146          - Asymptotic time- and space-complexity of instrumented code is
1147            equal to that of uninstrumented code.  (This means that
1148            tail-recursion is preserved by the instrumentation phase.)
1149    
1150          - Modules whose code has been instrumented in different sessions
1151            work together without problem.
1152    
1153          - There is no penalty whatsoever on uninstrumented code.
1154    
1155          - There is no penalty on "raise" expressions, even in
1156            instrumented code.
1157    
1158        A potential bug (or perhaps it is a feature, too):
1159    
1160          A back-trace reaches no further than the outermost instrumented
1161          non-trivial "raise".  Here, a "trivial" raise is one that is the
1162          sole RHS of a "handle" rule.  Thus, back-traces reach trough
1163    
1164               <exp> handle e => raise e
1165    
1166          and even
1167    
1168               <exp> handle Foo => raise Bar
1169    
1170          and, of course, through
1171    
1172               <exp> handle Foo => ...
1173    
1174         if the exception was not Foo.
1175    
1176         Back-traces always reach right through any un-instrumented code
1177         including any of its "handle" expressions, trivial or not.
1178    
1179       To try this out, do the following:
1180    
1181         - Erase all existing binfiles for your program.
1182           (You may keep binfiles for those modules where you think you
1183            definitely don't need back-tracing.)
1184         - Turn on back-trace instrumentation:
1185              SMLofNJ.Internals.BTrace.mode (SOME true);
1186         - Recompile your program.  (I.e., run "CM.make" or "use".)
1187         - You may now turn instrumentation off again (if you want):
1188              SMLofNJ.Internals.BTrace.mode (SOME false);
1189         - Run your program as usual.  If it raises an exception that
1190           reaches the interactive toplevel, then a back-trace will
1191           automatically be printed.  After that, the toplevel loop
1192           will print the exception history as usual.
1193    
1194    ----------------------------------------------------------------------
1195    Name: Matthias Blume
1196    Date: 2000/06/26 09:56:46 JST
1197    Tag: blume-20000626-setup
1198    Description:
1199    
1200    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1201          ML code before and after compiling a file (e.g., to set compiler
1202          flags)
1203    
1204    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1205              - associated changes to btrace.sml (BTrace instrumentation pass)
1206              - cleaner implementation of btimp.sml (BTrace tracing and report
1207                module)
1208    
1209    CM manual: * new path encoding documented
1210               * description of setup-parameter to "sml" added
1211    
1212    The biggest user-visible change to back-tracing is that it is no
1213    longer necessary to compile all traced modules within the same
1214    session.  (This was a real limitation.)
1215    
1216    ----------------------------------------------------------------------
1217    Name: Matthias Blume
1218    Date: 2000/06/24 12:40:00 JST
1219    Tag: blume-20000624-startup
1220    Description:
1221    
1222    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1223    bit too often -- to put it mildly. :)
1224    
1225    ----------------------------------------------------------------------
1226    Name: Matthias Blume
1227    Date: 2000/06/23 18:20:00 JST
1228    Tag: blume-20000623-btrace
1229    Description:
1230    
1231    This updates adds a backtrace facility to aid programmers in debugging
1232    their programs.  This involves the following changes:
1233    
1234    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1235       keeping track of the current call stack.  When programs are compiled
1236       in a special mode, the compiler will insert calls to these hooks
1237       into the user program.
1238       "Hook" means that it is possible for different implementations of
1239       back-tracing to register themselves (at different times).
1240    
1241    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1242       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1243       It can be turned on using this call:
1244         SMLofNJ.Internals.BTrace.mode (SOME true);
1245       Turning it off again:
1246         SMLofNJ.Internals.BTrace.mode (SOME false);
1247       Querying the current status:
1248         SMLofNJ.Internals.BTrace.mode NONE;
1249       Annotated programs are about twice as big as normal ones, and they
1250       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1251       where all hooks do nothing).  The slowdown with a plugin that is
1252       actually useful (such as the one supplied by default) is even greater,
1253       but in the case of the default plugin it is still only an constant
1254       factor (amortized).
1255    
1256    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1257       with a sub-structure BTrace for controlling back-tracing.  In particular,
1258       the above-mentioned function "mode" controls whether the annotation
1259       phase is invoked by the compiler.  Another important function is
1260       "trigger": when called it aborts the current execution and causes
1261       the top-level loop to print a full back-trace.
1262    
1263    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1264       for back-tracing.  It keeps track of the dynamic call stack and in
1265       addition to that it keeps a partial history at each "level" of that
1266       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1267       d and b (at separate times, dynamically), then the report will show:
1268    
1269       GOTO   d
1270             /c
1271       GOTO  \b
1272       CALL   a
1273    
1274       This shows that there was an initial non-tail call of a, then a
1275       tail-call to b or c, looping behavior in a cluster of functions that
1276       consist of b and c, and then a goto from that cluster (i.e., either from
1277       b or from c) to d.
1278    
1279       Note that (depending on the user program) the amount of information
1280       that the back-trace module has to keep track of at each level is bounded
1281       by a constant.  Thus, the whole implementation has the same asymptotical
1282       complexity as the original program (both in space and in time).
1283    
1284    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1285       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1286       which is raised by the "trigger" function mentioned above.
1287    
1288    Notes on usage:
1289    
1290    - Annotated code works well together with unannotated code:
1291    Unannotated calls simply do not show up at all in the backtrace.
1292    
1293    - It is not a good idea to let modules that were annotated during
1294    different sessions run at the same time.  This is because the compiler
1295    chooses small integers to identify individual functions, and there
1296    will be clashes if different modules were compiled in separate sessions.
1297    (Nothing will crash, and you will even be told about the clashes, but
1298    back-trace information will in general not be useful.)
1299    
1300    - Back-tracing can be confused by callcc and capture.
1301    
1302    - The only way of getting a back-trace right now is to explicitly
1303    invoke the "trigger" function from your user program.  Eventually, we
1304    should make every exception carry back-trace information (if
1305    available).  But since this creates more overhead at "raise"-time
1306    (similar to the current exnHistory overhead), I have not yet
1307    implemented this.  (The implementation will be rather easy.)  With
1308    exceptions carrying back-trace information, this facility will be even
1309    more useful because users don't need to modify their programs...
1310    
1311    - While it is possible to compile the compiler with back-trace
1312    annotations turned on (I did it to get some confidence in
1313    correctness), you must make absolutely sure that core.sml and
1314    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1315    be compiled with annotation because there is no core access yet, but
1316    if you compile btimp.sml with annotation, then the system will go into
1317    an infinite recursion and crash.)
1318    Since CM currently does not know about BTrace, the only way to turn
1319    annotations on and off for different modules of the compiler is to
1320    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1321    this is awkward and clumsy.
1322    
1323    Sample sessions:
1324    
1325    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1326    - SMLofNJ.Internals.BTrace.mode (SOME true);
1327    [autoloading]
1328    [autoloading done]
1329    val it = false : bool
1330    - structure X = struct
1331    -     fun main n = let
1332    -         fun a (x, 0) = d x
1333    -           | a (x, n) = b (x, n - 1)
1334    -         and b (x, n) = c (x, n)
1335    -         and c (x, n) = a (x, n)
1336    -         and d x = e (x, 3)
1337    -         and e (x, 0) = f x
1338    -           | e (x, n) = e (x, n - 1)
1339    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1340    -           | f n = n * g (n - 1)
1341    -         and g n = a (n, 3)
1342    -     in
1343    -         f n
1344    -     end
1345    - end;
1346    structure X : sig val main : int -> int end
1347    - X.main 3;
1348    *** BACK-TRACE ***
1349    GOTO   stdIn:4.2-13.20: X.main[2].f
1350    GOTO-( stdIn:4.2-13.20: X.main[2].e
1351    GOTO   stdIn:4.2-13.20: X.main[2].d
1352         / stdIn:4.2-13.20: X.main[2].a
1353         | stdIn:4.2-13.20: X.main[2].b
1354    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1355    CALL   stdIn:4.2-13.20: X.main[2].g
1356    GOTO   stdIn:4.2-13.20: X.main[2].f
1357    GOTO-( stdIn:4.2-13.20: X.main[2].e
1358    GOTO   stdIn:4.2-13.20: X.main[2].d
1359         / stdIn:4.2-13.20: X.main[2].a
1360         | stdIn:4.2-13.20: X.main[2].b
1361    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1362    CALL   stdIn:4.2-13.20: X.main[2].g
1363    GOTO   stdIn:4.2-13.20: X.main[2].f
1364    GOTO-( stdIn:4.2-13.20: X.main[2].e
1365    GOTO   stdIn:4.2-13.20: X.main[2].d
1366         / stdIn:4.2-13.20: X.main[2].a
1367         | stdIn:4.2-13.20: X.main[2].b
1368    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1369    CALL   stdIn:4.2-13.20: X.main[2].g
1370    GOTO   stdIn:4.2-13.20: X.main[2].f
1371    CALL   stdIn:2.15-17.4: X.main[2]
1372    -
1373    
1374    (Note that because of a FLINt bug the above code currently does not
1375    compile without BTrace turned on.)
1376    
1377    Here is another example, using my modified Tiger compiler:
1378    
1379    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1380    - SMLofNJ.Internals.BTrace.mode (SOME true);
1381    [autoloading]
1382    [autoloading done]
1383    val it = false : bool
1384    - CM.make "sources.cm";
1385    [autoloading]
1386    ...
1387    [autoloading done]
1388    [scanning sources.cm]
1389    [parsing (sources.cm):parse.sml]
1390    [creating directory CM/SKEL ...]
1391    [parsing (sources.cm):tiger.lex.sml]
1392    ...
1393    [wrote CM/sparc-unix/semant.sml]
1394    [compiling (sources.cm):main.sml]
1395    [wrote CM/sparc-unix/main.sml]
1396    [New bindings added.]
1397    val it = true : bool
1398    - Main.compile ("../testcases/merge.tig", "foo.out");
1399    *** BACK-TRACE ***
1400    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
1401    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1402    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1403    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1404    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1405    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1406    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
1407         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
1408    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
1409    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1410    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
1411    CALL   main.sml:1.18-118.4: Main.compile[2]
1412    -
1413    
1414    ----------------------------------------------------------------------
1415    Name: Matthias Blumen
1416    Date: 2000/06/21 18:00:00 JST
1417    Tag: blume-20000621-manual
1418    Description:
1419    
1420    CM manual update: Path environments documented.
1421    
1422    ----------------------------------------------------------------------
1423    Name: Matthias Blume
1424    Date: 2000/06/19 13:40:00
1425    Tag: blume-20000619-manual
1426    Description:
1427    
1428    CM manual and system/README update.  This only covers the fact that
1429    there are no more implicit anchors.  (Path environments and the "bind"
1430    option to "cm" have yet to be documented.)
1431    
1432    ----------------------------------------------------------------------
1433    Name: Matthias Blume
1434    Date: 2000/06/19 11:05:00 JST
1435    Tag: blume-20000619-chdir-bugfix
1436    Description:
1437    
1438    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1439    
1440    ----------------------------------------------------------------------
1441    Name: Matthias Blume
1442    Date: 2000/06/18 22:00:10 JST
1443    Tag: blume-20000618-implicit-anchors-really-gone
1444    Description:
1445    
1446    I updates the previous HISTORY entry where I forgot to mention that
1447    implicit anchors are no longer with us.
1448    
1449    The current update also gets rid of the (now useless) controller
1450    CM.Control.implicit_anchors.
1451    
1452    ----------------------------------------------------------------------
1453    Name: Matthias Blume
1454    Date: 2000/06/16 17:30:00 JST
1455    Tag: blume-20000616-anchorenv
1456    Description:
1457    
1458    This patch implements the long anticipated (just kidding :) "anchor
1459    environment" mechanism.  In the course of doing this, I also
1460    re-implemented CM's internal "SrcPath" module from scratch.  The new
1461    one should be more robust in certain boundary cases.  In any case, it
1462    is a lot cleaner than its predecessor (IMHO).
1463    
1464    This time, although there is yet another boot file format change, I
1465    kept the unpickler backward-compatible.  As a result, no new bootfiles
1466    are necessary and bootstrapping is straightforward.  (You cannot read
1467    new bootfiles into an old system, but the other way around is no
1468    problem.)
1469    
1470    Visible changes:
1471    
1472    ** 0. Implicit path anchors (without the leading $-symbol) are no
1473    longer recognized at all. This means that such path names are not
1474    illegal either.  For example, the name basis.cm simply refers to a
1475    local file called "basis.cm" (i.e, the name is an ordinary path
1476    relative to .cm-files directory).  Or, to put it differently, only
1477    names that start with $ are anchored paths.
1478    
1479    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1480    vanished.
1481    
1482    John (Reppy) had critizised this as soon as I originally proposed and
1483    implemented it, but at that time I did not really deeply believe
1484    him. :) Now I came full-circle because I need the $<singlearc> syntax
1485    in another place where it cannot be seen as an abbreviation for
1486    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1487    seems to mean (i.e., it "expands" into the corresponding anchor
1488    value).
1489    
1490    However, when paths are used as members in CM description files, it
1491    continues to be true that there must be at least another arc after the
1492    anchor.  This is now enforced separately during semantic analysis
1493    (i.e., from a lexical/syntactical point of view, the notation is ok.)
1494    
1495    ** 2. The "cm" class now accepts an option "bind".  The option's value
1496    is a sub-option list of precisely two items -- one labeled "anchor"
1497    and the other one labeled "value".  As you might expect, "anchor" is
1498    used to specify an anchor name to be bound, and "value" specifies what
1499    the anchor is being bound to.
1500    
1501    The value must be a directory name and can be given in either standard
1502    syntax (including the possibility that it is itself an anchored path)
1503    or native syntax.
1504    
1505    Examples:
1506    
1507       foo.cm (bind:(anchor:bar value:$mystuff/bar))
1508       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1509    
1510    and so on.
1511    
1512    The meaning of this is that the .cm-file will be processed with an
1513    augmented anchor environment where the given anchor(s) is/are bound to
1514    the given values(s).
1515    
1516    The rationale for having this feature is this: Suppose you are trying
1517    to use two different (already stable) libraries a.cm and b.cm (that
1518    you perhaps didn't write yourself).  Further, suppose each of these
1519    two libraries internally uses its own auxiliary library $aux/lib.cm.
1520    Normally you would now have a problem because the anchor "lib" can not
1521    be bound to more than one value globally.  Therefore, the project that
1522    uses both a.cm and b.cm must locally redirect the anchor to some other
1523    place:
1524    
1525       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1526       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1527    
1528    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1529    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
1530    
1531    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1532    you when you do so at the time of CM.stabilize).  Therefore, you can
1533    also use an anchored path as the value:
1534    
1535      a.cm (bind:(anchor:lib value:$a-lib))
1536      b.cm (bind:(anchor:lib value:$b-lib))
1537    
1538    Now you can globally configure (using the usual CM.Anchor.anchor or
1539    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1540    itself is always locally bound, setting it globally is no longer
1541    meaningful or necessary (but it does not hurt either).  In fact, "lib"
1542    can still be used as a global anchor for separate purposes.  As a
1543    matter of fact, one can locally define "lib" in terms of a global
1544    "lib":
1545    
1546      a.cm (bind:(anchor:lib value:$lib/a))
1547      b.cm (bind:(anchor:lib value:$lib/b))
1548    
1549    ** 3: The encoding of path names has changed.  This affects the way
1550    path names are shown in CM's progress report and also the internal
1551    protocol encoding used for parallel make.
1552    
1553    The encoding now uses one or more ':'-separated segments.  Each
1554    segments corresponds to a file that has been specified relative to the
1555    file given by its preceding segment.  The first segment is either
1556    relative to the CWD, absolute, or anchored.  Each segment itself is
1557    basically a Unix pathname; all segments but the first are relative.
1558    
1559    Example:
1560    
1561       $foo/bar/baz.cm:a/b/c.sml
1562    
1563    This path denotes the file bar/a/b/c.sml relative to the directory
1564    denoted by anchor "foo".  Notice that the encoding also includes
1565    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1566    paths are resolved relative to the .cm-files directory, so baz.cm must
1567    be ignored to get the "real" pathname.
1568    
1569    To make this fact more obvious, CM puts the names of such "virtual
1570    arcs" into parentheses when they appear in progress reports. (No
1571    parentheses will appear in the internal protocol encoding.)  Thus,
1572    what you really see is:
1573    
1574      $foo/bar/(baz.cm):a/b/c.sml
1575    
1576    I find this notation to be much more informative than before.
1577    
1578    Another new feature of the encoding is that special characters
1579    including parentheses, colons, (back)slashes, and white space are
1580    written as \ddd (where ddd is the decimal encoding of the character).
1581    
1582    *** The CM manual still needs to be updated.
1583    
1584    ----------------------------------------------------------------------
1585    Name: Allen Leung
1586    Date: 2000/06/15 00:38:00
1587    Tag: leunga-20000615-x86-peephole
1588    
1589    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1590    
1591    ----------------------------------------------------------------------
1592    Name: Matthias Blume
1593    Date: 2000/06/12 11:40:00
1594    Tag: blume-20000612-parmakefix
1595    Description:
1596    
1597    More cleanup after changing the file naming scheme: This time I
1598    repaired the parallel make mechanism for CMB.make which I broke earlier.
1599    
1600    ----------------------------------------------------------------------
1601    Name: Allen Leung
1602    Date: 2000/06/09 01:25:00
1603    Tag: leunga-20000609-various
1604    
1605    None of these things should affect normal SML/NJ operations
1606    
1607    1. Peephole improvements provided by Fermin (c--)
1608    2. New annotation DEFUSE for adding extra dependence (moby)
1609    3. New X86 LOCK instructions (moby)
1610    4. New machine description language for reservation tables (scheduling)
1611    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1612       trees etc.)
1613    6. I've changed the CM files so that they can work with versions
1614       110.0.6, 110.25 and 110.28
1615    
1616    ----------------------------------------------------------------------
1617    Name: Matthias Blume
1618    Date: 2000/06/09 12:40:00
1619    Tag: blume-20000609-log
1620    Description:
1621    
1622    - Removed all(?) remaining RCS Log entries from sources.
1623    
1624    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1625      anchored paths).
1626    
1627    ----------------------------------------------------------------------
1628    Name: Matthias Blume
1629    Date: 2000/06/07 17:00:00 JST
1630    Tag: blume-20000607-no-implicit-anchors
1631    Description:
1632    
1633    1. This update changes the default setting for
1634    CM.Control.implicit_anchors from true to false.  This means that
1635    implicit anchors are no longer permitted by default.  I also tried to
1636    make sure that nothing else still relies on implicit anchors.
1637    (This is the next step on the schedule towards a CM that does not even
1638    have the notion of implicit anchors anymore.)
1639    
1640    2. More CM manual updates.
1641    
1642    3. I managed to track down and fix the pickling bug I mentioned last
1643    time.  Because of the previously existing workaround, this entails no
1644    immediate practical changes.
1645    
1646    ----------------------------------------------------------------------
1647    Name: Matthias Blume
1648    Date: 2000/06/06 11:15:00 JST
1649    Tag: blume-20000606-lazierpickle
1650    Description:
1651    
1652    !!!! NEW BOOT FILES !!!!
1653    
1654    * The main purpose of this update is to make library pickles lazier in
1655    order to reduce the initial space penalty for autoloading a library.
1656    As a result, it is now possible to have $smlnj/compiler.cm
1657    pre-registered.  This should take care of the many complaints or
1658    inquiries about missing structure Compiler.  This required changes to
1659    CM's internal data structures and small tweaks to some algorithms.
1660    
1661    As a neat additional effect, it is no longer necessary (for the sake
1662    of lean heap image files) to distinguish between a "minimal" CM and a
1663    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1664    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1665    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1666    
1667    To make the life of compiler-hackers easier, "makeml" now also
1668    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1669    words, after you bootstrap a new sml for the first time, you will not
1670    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1671    around you will still have to do it, though.)
1672    
1673    * A second change consists of major updates to the CM manual.  There
1674    are now several appendices with summary information and also a full
1675    specification of the CM description file syntax.
1676    
1677    * In directory src/system I added the script "allcross".  This script
1678    invokes sml and cross-compiles the compiler for all supported
1679    architectures.  (Useful when providing a new set of boot files.)
1680    
1681    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1682    added a small tweak to pickle-util.sml to work around this problem,
1683    but it is not a proper fix yet.  I will investigate further.  (The
1684    effect of the bug was an inflation of library pickle size.)
1685    
1686    * Version number increased to 110.28.1 (to avoid compatibility problems).
1687    
1688    ----------------------------------------------------------------------
1689    Name: Allen Leung
1690    Date: 2000/05/25 17:28 EDT
1691    Tag: leunga-20000525-ra
1692    Description:
1693    
1694      Fixed a bug in freezing phase of the register allocator.
1695    
1696    ----------------------------------------------------------------------
1697    Name: Allen Leung
1698    Date: 2000/05/15 22:53 EDT
1699    Tag: leunga-20000515-alpha-x86-ra
1700    Description:
1701    
1702      1. Alpha
1703    
1704          Slight cleanup.  Removed the instruction SGNXL
1705    
1706      2. X86
1707    
1708          Added the following instructions to the instruction set:
1709    
1710            ROLx, RORx,
1711            BTx, BTSx, BTLx, BTRx,
1712            XCHGx, and variants with the LOCK prefix
1713    
1714      3. Register Allocation
1715    
1716          The module ra-rewrite-with-renaming has been improved.
1717    
1718      These have no effect on SML/NJ.
1719    
1720    ----------------------------------------------------------------------
1721    Name: Matthias Blume
1722    Date: 2000/05/15 16:20:00 JST
1723    Tag: blume-20000515-lightrebuild
1724    Description:
1725    
1726    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1727       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1728       defined.  In effect, the command will not build any cross-compiler
1729       backends and therefore finish more quickly.
1730    
1731       The "fixpt" script also takes a "-light" switch to be able to use
1732       this new facility while compiling for a fixpoint.
1733    
1734    2. I replaced all mentions of anchored paths in group owner specifications
1735       with simple relative paths (usually starting with "..").
1736       The rationale is that a library's internal workings should not be
1737       compromised by the lack of some anchor.  (An anchor is necessary
1738       for someone who wants to refer to the library by an anchored path,
1739       but it should not be necessary to build the same library in the first
1740       place.)
1741    
1742    3. I changed the way CM's tool mechanism determines the shell command
1743       string used for things like ml-yacc etc. so that it does not break
1744       when CM.Control.implicit_anchors is turned off.
1745    
1746    ----------------------------------------------------------------------
1747    Name: Matthias Blume
1748    Date: 2000/05/12 18:20:00 JST
1749    Tag: blume-20000512-ml-build
1750    Description:
1751    
1752    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1753    getting installed properly (by config/install.sh).
1754    
1755    ----------------------------------------------------------------------
1756    Name: Matthias Blume
1757    Date: 2000/05/12 17:30:00 JST
1758    Tag: blume-20000512-anchors
1759    Description:
1760    
1761    !!! NEW BOOT FILES !!!
1762    
1763    This change is in preparation of fading out support for "implicitly
1764    anchored path names".  I went through all sources and used the
1765    explicit (and relatively new) $-notation.  See system/README and the
1766    CM manual for more info on this.
1767    
1768    I also modified the anchoring scheme for some things such as "smlnj",
1769    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1770    anchors are more expressive: anchor name and first arc do not have to
1771    coincide.  This entails the following user-visible change:
1772    
1773    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1774    particular, when you fire up sml with a command-line argument, say,
1775    e.g.:
1776    
1777       sml '$smlnj/cmb.cm'
1778    
1779    At the ML toplevel prompt:
1780    
1781       CM.autoload "$smlnj/cmb.cm";
1782    
1783    There is also a new controller in CM.Control that can be used to turn
1784    off all remaining support for implicit anchors by saying:
1785    
1786        CM.autoload "$smlnj/
1787        #set CM.Control.implicit_anchors false;
1788    
1789    This causes CM to reject implicitly anchored paths.  This is (for the
1790    time being) less permissive than the "final" version where there will
1791    be no more such implicit anchors and relative paths will be just that:
1792    relative.
1793    
1794    The next step (version after next version?) will be to make the
1795    default for CM.Control.implicit_anchors false.  After the dust has
1796    settled, I can then produce the "final" version of this...
1797    
1798    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1799    
1800    ----------------------------------------------------------------------
1801    Name: Matthias Blume
1802    Date: 2000/05/11 16:30:00 JST
1803    Tag: blume-20000511-sources
1804    Description:
1805    
1806    The main change is that I added function CM.sources as a generalized
1807    version of the earlier CM.makedepend.  This entails the following
1808    additional changes:
1809    
1810      - CM.makedepend has been dropped.
1811    
1812      - CM manual has been updated.
1813    
1814      - TOOLS signature and API have been changed.
1815    
1816    ----------------------------------------------------------------------
1817    Name: Allen Leung
1818    Date: 2000/05/10 21:17 EDT
1819    Tag: leunga-20000510-moby-c--ssa
1820    Description:
1821    
1822      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1823    None of these affect SML/NJ.
1824    
1825    1. Register Allocation
1826    
1827        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1828           This module tries to remove local (i.e. basic block level) redundancies
1829           during spilling.
1830    
1831        b. A new framework for performing region based register allocation.
1832           Not yet entirely functional.
1833    
1834    2. X86
1835    
1836       a. DefUse for POP was missing the stack pointer [found by Lal]
1837       b. Reload for CALL was incorrect in X86Spill [found by John]
1838       c. Various fixes in X86Spill so that it can be used correctly for
1839          the new spilling module.
1840    
1841    3. SSA/IR
1842    
1843       a. New module ir/dj-dataflow.sml implements elimination based
1844          data flow analysis.
1845    
1846    4. MLRiscGen
1847    
1848       a. Fix for gc type annotation
1849    
1850    5. MDGen
1851    
1852       Various fixes for machine description -> ml code translation.  For ssa
1853       only.
1854    
1855    ----------------------------------------------------------------------
1856    Name: Allen Leung
1857    Date: 2000/05/08 22:17 EDT
1858    Tag: leunga-20000508-labexp
1859    Description:
1860    
1861      Fermin has found a few assembly problems with constant expressions
1862      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1863      which choke on dumb assemblers.  This is his fix.
1864    
1865    ----------------------------------------------------------------------
1866    Name: Dave MacQueen
1867    Date: 2000/04/09 14:00 EDT
1868    Tag: dbm-20000502-Version_110_28
1869    Description:
1870    
1871    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1872    
1873    2. Updated config/version to 110.28
1874    
1875    3. Updated config/srcarchiveurl
1876    
1877    3. New boot files!
1878       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1879    
1880    ----------------------------------------------------------------------
1881    Name: Matthias Blume
1882    Date: 2000/05/01 19:05:00 JST
1883    Tag: blume-20000501-noweb
1884    Description:
1885    
1886    A new noweb tool has been added.  The existing system is entirely
1887    unaffected by this, but some CM users have asked for renewed noweb
1888    support.  Everything is documented in the CM manual.
1889    
1890    New (plugin) libraries:
1891    
1892       noweb-tool.cm
1893       nw-ext.cm
1894    
1895    ----------------------------------------------------------------------
1896    Name: Dave MacQueen
1897    Date: 2000/04/30 12:40PM EDT
1898    Tag: dbm-20000430-bug_fixes
1899    Description:
1900    
1901    1. Fix for bug 1498
1902       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1903       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1904         added toRealArray function
1905       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1906         added check for tag Obj.RealArray to array printing case in ppObj
1907    
1908    2. Fix for bug 1510
1909       smlnj/src/compiler/Semant/types/typesutil.sml
1910         fixed definition of dummyargs (used by equalTycon) so that
1911         dummy args are distinct types
1912    
1913    ----------------------------------------------------------------------
1914    Name: Matthias Blume
1915    Date: 2000/04/30 01:00:00 JST
1916    Tag: blume-20000430-versions
1917    Description:
1918    
1919    1. CM version numbering added.  This is an implementation of Lal's
1920       proposal for adding version numbers and version checking to .cm
1921       files.  Lal said that his proposal was just that -- a proposal.
1922       For the time being I went ahead and implemented it so that people
1923       can comment on it.  Everything is completely backward-compatible
1924       (except for the stable library format, i.e., new bootfiles!).
1925    
1926       As usual, see the CM manual for details.
1927    
1928    2. An alternative syntax for anchored paths has been implemented.
1929       Dave has recently voiced the same concerns that I had when I did
1930       this, so there should be some support.  My take is that eventually
1931       I will let support for the current syntax (where anchors are
1932       "implicit") fade out in favor of the new, explicit syntax.
1933       In order to be backward-compatible, both old and new syntax are
1934       currently supported.
1935    
1936       Again, see the CM manual for details.
1937    
1938    3. Parallel make is trying to be slightly smarter:  When the master
1939       process finds a "bottleneck", i.e., when there is only one
1940       compilation unit that can be compiled and everybody else is
1941       waiting on it, then it will simply compile it directly instead
1942       of clumsily telling one of the slaves to do it.
1943    
1944    4. Support for "unsharing" added.  This is necessary in order to be
1945       able to have two different versions of the same library running
1946       at the same time (e.g., for trying out a new MLRISC while still
1947       having the old MLRISC linked into the current compiler, etc.)
1948       See the CM manual.
1949    
1950    5. Simple "makedepend" functionality added for generating Makefile
1951       dependency information.  (This is rather crude at the moment.
1952       Expect some changes here in the future.)
1953    
1954    6. ".fun" added as a recognized suffix for ML files. Also documented
1955       explicitly in the manual that the fallback behavior (unknown suffix
1956       -> ML file) is not an official feature!
1957    
1958    7. Small changes to the pickler for stable libraries.
1959    
1960    8. Several internal changes to CM (for cleanup/improvement).
1961    
1962    
1963    !!!! NEW BINFILES !!!!
1964    
1965  ----------------------------------------------------------------------  ----------------------------------------------------------------------
1966  Name: Matthias Blume  Name: Matthias Blume
1967  Date: 2000/04/28 17:30:00 JST  Date: 2000/04/28 17:30:00 JST
# Line 225  Line 2176 
2176    
2177  3.  Assembly  3.  Assembly
2178    
2179      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
2180      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
2181      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
2182    
# Line 248  Line 2199 
2199    
2200  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
2201  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
2202  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
2203  legal surface-syntax symbol).  legal surface-syntax symbol).
2204    
2205  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 258  Line 2209 
2209  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
2210    
2211  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2212  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
2213  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
2214  more details.  more details.
2215    
# Line 373  Line 2324 
2324     (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
2325     used).     used).
2326    
2327     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
2328     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
2329     Makefiles or Shell scripts.     Makefiles or Shell scripts.
2330    
2331     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 386  Line 2337 
2337    
2338     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
2339    
2340     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
2341     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
2342    
2343  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 520  Line 2471 
2471    
2472       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
2473       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
2474       testl is perferable to andl.       testl is preferable to andl.
2475    
2476  7. RA (x86 only)  7. RA (x86 only)
2477    
# Line 702  Line 2653 
2653    
2654  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2655  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2656  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
2657  explanation.  explanation.
2658    
2659  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2660  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2661  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2662  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 749  Line 2700 
2700  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
2701  <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,
2702  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
2703  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2704  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2705  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2706    
# Line 799  Line 2750 
2750    it from that remote directory.    it from that remote directory.
2751    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2752    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2753    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2754    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2755    
2756    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.631  
changed lines
  Added in v.839

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