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 583, Thu Mar 23 21:52:30 2000 UTC revision 1062, Wed Feb 13 21:15:14 2002 UTC
# Line 8  Line 8 
8  The form of an entry should be:  The form of an entry should be:
9    
10  Name:  Name:
11  Date:  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15    ----------------------------------------------------------------------
16    Name: Matthias Blume
17    Date: 2002/02/13 16:15:00 EST
18    Tag: blume-20020213-various
19    Description:
20    
21    1. More tweaks to ml-nlffigen:
22    
23       - better internal datastructures (resulting in slight speedup)
24       - "-match" option requires exact match
25       - "localized" gensym counters (untagged structs/unions nested within
26         other structs/unions or within typedefs get a fresh counter; their
27         tag will be prefixed by a concatenation of their parents' tags)
28       - bug fixes (related to calculation of transitive closure of types
29         to be included in the output)
30    
31    2. Minor Basis updates:
32    
33       - added implementations for List.collate and Option.app
34    
35    ----------------------------------------------------------------------
36    Name: Matthias Blume
37    Date: 2002/02/11 15:55:00 EST
38    Tag: blume-20020211-gensym
39    Description:
40    
41    Added a "-gensym" option to command line of ml-nlffigen.  This can be
42    used to specify a "stem" -- a string that is inserted in all "gensym'd"
43    names (ML structure names that correspond to unnamed C structs, unions,
44    and enums), so that separate runs of ml-nlffigen do not clash.
45    
46    ----------------------------------------------------------------------
47    Name: Matthias Blume
48    Date: 2002/02/11 12:05:00 EST
49    Tag: blume-20020211-gensml
50    Description:
51    
52    A quick fix for a problem with GenSML (in the pgraph-util library):
53    Make generation of toplevel "local" optional.  (Strictly speaking,
54    signature definitions within "local" are not legal SML.)
55    
56    Other than that: updates to INSTALL and cm/TODO.
57    
58    ----------------------------------------------------------------------
59    Name: Matthias Blume
60    Date: 2002/02/08 15:00:00 EST
61    Tag: blume-20020208-uniquepid
62    Description:
63    
64    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
65    
66    1. The installer (config/install.sh) has gotten smarter:
67    
68         - Configuration options are a bit easier to specify now
69           (in config/targets).
70         - Bug in recognizing .tar.bz2 files fixed.
71         - Installer automatically resolves dependencies between
72           configuration options (e.g., if you ask for eXene, you will
73           also get cml -- regardless whether you asked for it or not).
74         - Installer can run in "quieter mode" by setting the environment
75           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
76           "completely silent", though.
77         - Build HashCons library as part of smlnj-lib.
78    
79    2. A new scheme for assigning persistent identifiers to compilation
80       units (and, by extension, to types etc.) has been put into place.
81       This fixes a long-standing bug where types and even dynamic values
82       can get internally confused, thereby compromising type safety
83       (abstraction) and dynamic correctness.  See
84    
85         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
86    
87       for an example of how things could go wrong until now.
88    
89       The downside of the new scheme is that pids are not quite as
90       persistent as they used to be: CM will generate a fresh pid
91       for every compilation unit that it thinks it sees for the first
92       time.  That means that if you compile starting from a clean, fresh
93       source tree at two different times, you end up with different
94       binaries.
95    
96       Cutoff recompilation, however, has not been compromised because
97       CM keeps pid information in special caches between runs.
98    
99    ----------------------------------------------------------------------
100    Name: Lal George
101    Date: 2002/02/07 15:34:13 EST 2002
102    Tag: <none>
103    Description:
104    
105    Compilers that generate assembly code may produce  global labels
106    whose value is resolved at link time. The various peephole optimization
107    modules did not take this in account.
108    
109    TODO. The Labels.addrOf function should really return an option
110    type so that clients are forced to deal with this issue, rather
111    than an exception being raised.
112    
113    ----------------------------------------------------------------------
114    Name: Lal George
115    Date: 2002/02/06 13:55:02 EST
116    Tag: george-20020206-ra-breakup
117    Description:
118    
119    1. A bug fix from Allen.
120    
121        A typo causes extra fstp %st(0)'s to be generated at compensation
122        edges, which might cause stack underflow traps at runtime.  This
123        occurs in fft where there are extraneous fstps right before the 'into'
124        trap instruction (in this case they are harmless since none of the
125        integers overflow.)
126    
127    2. Pulled out various utility modules that were embedded in the modules
128       of the register allocator. I need these modules for other purposes, but
129       they are not complete enough to put into a library (just yet).
130    ----------------------------------------------------------------------
131    Name: Matthias Blume
132    Date: 2002/01/31 16:05:00 EST
133    Tag: blume-20020131-sparc-ccalls
134    Description:
135    
136    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
137       of extra stack space by mistake.  Fixed.
138    
139    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
140    
141    ----------------------------------------------------------------------
142    Name: Allen Leung
143    Date: 2002/01/30
144    Tag: leunga-20020130-nowhere-bug-fix
145    Description:
146    
147       MLRISC bug fixes:
148       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
149       2. MachineInt.fromString was negating its value.
150    
151    ----------------------------------------------------------------------
152    Name: Matthias Blume
153    Date: 2002/01/29
154    Tag: blume-20020129-INSTALL
155    Description:
156    
157    - Added somewhat detailed installation instructions (file INSTALL).
158    - Fixed curl-detection bug in config/install.sh.
159    - It is now possible to select the URL getter using the URLGETTER
160      environment variable:
161    
162          not set / "unknown"      --> automatic detection (script tries wget,
163                                       curl, and lynx)
164          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
165                                       how to properly invoke them)
166          other                    --> use $URLGETTER directly, it must take
167                                       precisely two command-line arguments
168                                       (source URL and destination file name)
169    
170    ----------------------------------------------------------------------
171    Name: Matthias Blume
172    Date: 2002/01/28
173    Tag: blume-20020128-sparc-ccalls
174    Description:
175    
176    - Fixed problem with calculation of "used" registers in sparc-c-calls.
177    - Make use of the allocParam argument in sparc-c-calls.
178    
179    ----------------------------------------------------------------------
180    Name: Matthias Blume
181    Date: 2002/01/28
182    Tag: blume-20020128-allocParam
183    Description:
184    
185    John Reppy:  Changes c-calls API to accept client-callback for
186    allocating extra stack space.
187    me: Corresponding changes to mlriscGen (using a dummy argument that
188        does not change the current behavior).
189    
190    ----------------------------------------------------------------------
191    Name: Matthias Blume
192    Date: 2002/01/28 12:00:00
193    Tag: Release_110_38
194    Description:
195    
196    This time for real!!!
197    
198    ----------------------------------------------------------------------
199    Name: Matthias Blume
200    Date: 2002/01/28 10:56:00 EST
201    Tag: blume-20020128-retraction
202    Description:
203    
204    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
205       with blume-Release_110_38-retracted.)
206    
207    1. Fixed a problem with incorrect rounding modes in real64.sml.
208       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
209    
210    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
211       and enums fixed.  The naming of corresponding ML identifiers should
212       now be consistent again.
213    
214    ----------------------------------------------------------------------
215    Name: Allen Leung
216    Date: 2002/01/27
217    Tag: leunga-20020127-nowhere
218    Description:
219    
220       Added a target called nowhere in the configuration scripts.
221       Enabling this will build the MLRISC 'nowhere' tool (for translating
222       programs with where-clauses into legal SML code) during installation.
223    
224    ----------------------------------------------------------------------
225    Name: Matthias Blume
226    Date: 2002/01/25 21:27:00 EST
227    Tag: blume-Release_110_38-retracted
228    Description:
229    
230    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
231    
232    README will be added later.
233    
234    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
235    (Release_110_38) removed.  Reason: Last-minute bug fixes.
236    
237    ----------------------------------------------------------------------
238    Name: Matthias Blume
239    Date: 2002/01/25
240    Tag: blume-20020125-ffi
241    Description:
242    
243    A large number of tweaks and improvements to ml-nlffi-lib and
244    ml-nlffigen:
245    
246       - ML represenation types have been streamlined
247       - getter and setter functions work with concrete values, not abstract
248         ones where possible
249       - ml-nlffigen command line more flexible (see README file there)
250       - some bugs have been fixed (hopefully)
251    
252    ----------------------------------------------------------------------
253    Name: Lal George
254    Date: 2002/01/24
255    Tag: george-20020124-risc-ra-interface
256    Description:
257    
258       There is a dramatic simplification in the interface to the
259       register allocator for RISC architectures as a result of making
260       parallel copy instructions explicit.
261    
262    ----------------------------------------------------------------------
263    Name: Matthias Blume
264    Date: 2002/01/22
265    Tag: blume-20020122-x86-ccalls
266    Description:
267    
268    Bug fix for c-calls on x86 (having to do with how char- and
269    short-arguments are being handled).
270    
271    ----------------------------------------------------------------------
272    Name: Matthias Blume
273    Date: 2002/01/21
274    Tag: blume-20020121-ff
275    Description:
276    
277    Another day of fiddling with the FFI...
278    
279    1. Bug fix/workaround:  CKIT does not complain about negative array
280       dimensions, so ml-nlffigen has to guard itself against this possibility.
281       (Otherwise a negative dimension would send it into an infinite loop.)
282    
283    2. Some of the abstract types (light objects, light pointers, most "base"
284       types) in structure C are now eqtypes.
285    
286    3. Added constructors and test functions for NULL function pointers.
287    
288    ----------------------------------------------------------------------
289    Name: Matthias Blume
290    Date: 2002/01/18
291    Tag: blume-20020118-ready-for-new-release
292    Description:
293    
294    Made config/srcarchiveurl point to a new place.  (Will provide boot
295    files shortly.)
296    
297    Maybe we christen this to be 110.38?
298    
299    ----------------------------------------------------------------------
300    Name: Matthias Blume
301    Date: 2002/01/18
302    Tag: blume-20020118-more-ffifiddle
303    Description:
304    
305    Today's FFI fiddling:
306    
307      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
308        C.Set.* that use concrete (MLRep.*) arguments and results instead
309        of abstract ones.
310    
311      - Provided word-style bit operations etc. for "int" representation
312        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
313        and Long.
314    
315    ----------------------------------------------------------------------
316    Name: Matthias Blume
317    Date: 2002/01/18
318    Tag: blume-20020118-use-x86-fp
319    Description:
320    
321    Now that x86-fast-fp seems to be working, I turned it back on again
322    by default.  (Seems to work fine now, even with the FFI.)
323    
324    Other than that, I added some documentation about the FFI to
325    src/ml-nlffigen/README and updated the FFI test examples in
326    src/ml-nlffi-lib/Tests/*.
327    
328    ----------------------------------------------------------------------
329    Name: Allen Leung
330    Date: 2002/01/17
331    Tag: leunga-20020117-x86-fast-fp-call
332    Description:
333    
334       1. Fixed a problem with handling return fp values when x86's fast fp
335          mode is turned on.
336    
337       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
338          of %f32.
339    
340       3. Added a constructor INT32lit to the ast of MLRISC tools.
341    
342    ----------------------------------------------------------------------
343    Name: Matthias Blume
344    Date: 2002/01/16
345    Tag: blume-20020116-ffifiddle
346    Description:
347    
348    More fiddling with the FFI interface:
349    
350     - Make constness 'c instead of rw wherever possible.  This eliminates
351       the need for certain explicit coercions.  (However, due to ML's
352       value polymorphism, there will still be many cases where explicit
353       coercions are necessary.  Phantom types are not the whole answer
354       to modeling a subtyping relationship in ML.)
355    
356     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
357    
358     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
359       clumsy workaround for varargs problem.)
360    
361    ----------------------------------------------------------------------
362    Name: Lal George
363    Date: 2002/01/15
364    Tag: <none>
365    Description:
366    
367    1. Since COPY instructions are no longer native to the architecture,
368       a generic functor can be used to implement the expandCopies function.
369    
370    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
371       TEXT segment.
372    
373    ----------------------------------------------------------------------
374    Name: Matthias Blume
375    Date: 2002/01/15
376    Tag: blume-20020115-ffiupdates
377    Description:
378    
379    1. Fix for bug resulting in single-precision float values being returned
380       incorrectly from FFI calls.
381    
382    2. Small modifications to C FFI API:
383    
384        - memory-allocation routines return straight objects (no options)
385          and raise an exception in out-of-memory situations
386        - unsafe extensions to cast between function pointers and pointers
387          from/to ints
388        - added structure C_Debug as an alternative to structure C where
389          pointer-dereferencing (|*| and |*!) always check for null-pointers
390        - added open_lib' to DynLinkage;  open_lib' works like open_lib
391          but also takes a (possibly empty) list of existing library handles
392          that the current library depends on
393    
394    ----------------------------------------------------------------------
395    Name: Matthias Blume
396    Date: 2002/01/10
397    Tag: blume-20020110-newffigen
398    Description:
399    
400    1. Updates to portable graph code.
401    
402    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
403       more scalable now so that even huge interfaces such as the one
404       for GTK compile in finite time and space. :-)
405       See src/ml-nlffigen/README for details on what's new.
406    
407    ----------------------------------------------------------------------
408    Name: Lal George
409    Date: 2001/01/09 14:31:35 EST 2002
410    Tag: george-20011206-rm-native-copy
411    Description:
412    
413            Removed the native COPY and FCOPY instructions
414            from all the architectures and replaced it with the
415            explicit COPY instruction from the previous commit.
416    
417            It is now possible to simplify many of the optimizations
418            modules that manipulate copies. This has not been
419            done in this change.
420    
421    ----------------------------------------------------------------------
422    Name: Lal George
423    Date: 2001/12/06 16:50:13 EST 2001
424    Tag: george-20011206-mlrisc-instruction
425    Description:
426    
427    Changed the representation of instructions from being fully abstract
428    to being partially concrete. That is to say:
429    
430      from
431            type instruction
432    
433      to
434            type instr                              (* machine instruction *)
435    
436            datatype instruction =
437                LIVE of {regs: C.cellset, spilled: C.cellset}
438              | KILL of {regs: C.cellset, spilled: C.cellset}
439              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
440              | ANNOTATION of {i: instruction, a: Annotations.annotation}
441              | INSTR of instr
442    
443    This makes the handling of certain special instructions that appear on
444    all architectures easier and uniform.
445    
446    LIVE and KILL say that a list of registers are live or killed at the
447    program point where they appear. No spill code is generated when an
448    element of the 'regs' field is spilled, but the register is moved to
449    the 'spilled' (which is present, more for debugging than anything else).
450    
451    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
452    We used to generate:
453    
454            DEFFREG f1
455            f1 := f2 + f3
456            trapb
457    
458    but now generate:
459    
460            f1 := f2 + f3
461            trapb
462            LIVE {regs=[f1,f2,f3], spilled=[]}
463    
464    Furthermore, the DEFFREG (hack) required that all floating point instruction
465    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
466    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
467    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
468    this architecture).
469    
470    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
471    all the architectures. This will result in further simplification of the
472    register allocator that must be aware of them for coalescing purposes, and
473    will also simplify certain aspects of the machine description that provides
474    callbacks related to parallel copies.
475    
476    ANNOTATION should be obvious, and now INSTR represents the honest to God
477    machine instruction set!
478    
479    The <arch>/instructions/<arch>Instr.sml files define certain utility
480    functions for making porting easier -- essentially converting upper case
481    to lower case. All machine instructions (of type instr) are in upper case,
482    and the lower case form generates an MLRISC instruction. For example on
483    the alpha we have:
484    
485      datatype instr =
486         LDA of {r:cell, b:cell, d:operand}
487       | ...
488    
489      val lda : {r:cell, b:cell, d:operand} -> instruction
490        ...
491    
492    where lda is just (INSTR o LDA), etc.
493    
494    ----------------------------------------------------------------------
495    Name: Matthias Blume
496    Date: 2001/11/22 21:40:00 EST
497    Tag: Release_110_37
498    Description:
499    
500    Release 110.37.  This time for real.
501    
502    ----------------------------------------------------------------------
503    Name: Matthias Blume
504    Date: 2001/11/21 16:35:00 EST
505    Tag: blume-20011121-foot-in-mouth
506    Description:
507    
508    Removed the "Release_110_37" tag because of a serious bug.
509    This will be re-tagged once the bug is fixed.
510    
511    ----------------------------------------------------------------------
512    Name: Matthias Blume
513    Date: 2001/11/21 16:14:00 EST
514    Tag: blume-20011121-forgottenfile
515    Description:
516    
517    Forgot to add a file.  (Just a .tex-file -- part of
518    the CM manual source.)
519    
520    ----------------------------------------------------------------------
521    Name: Matthias Blume
522    Date: 2001/11/21 16:10:00 EST
523    Tag: blume-20011121-invalid_110_37
524    Description:
525    
526    Note: I removed the original tag "Release_110_37" from this commit
527          because we found a serious bug in all non-x86 backends.
528          - Matthias
529    
530    1. Modifications to the SML/NJ code generator and to the runtime system
531       so that code object name strings are directly inserted into code
532       objects at code generation time.  The only business the runtime system
533       has with this is now to read the name strings on occasions.
534       (The encoding of the name string has also changed somewhat.)
535    
536    2. CM now implements a simple "set calculus" for specifying export lists.
537       In particular, it is now possible to refer to the export lists of
538       other libraries/groups/sources and form unions as well as differences.
539       See the latest CM manual for details.
540    
541    3. An separate notion of "proxy" libraries has again be eliminated from
542       CM's model.  (Proxy libraries are now simply a special case of using
543       the export list calculus.)
544    
545    4. Some of the existing libraries now take advantage of the new set
546       calculus.
547       (Notice that not all libraries have been converted because some
548       of the existing .cm-files are supposed to be backward compatible
549       with 110.0.x.)
550    
551    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
552       "General.exnMessage"!  The former relies on a certain hook to be
553       initialized, and that often does not happen in the stand-alone case.)
554    
555    ----------------------------------------------------------------------
556    Name: Lal George
557    Date: 2001/11/21  13:56:18 EST
558    Tag: george-2001121-pseudo-ops
559    Description:
560    
561      Implemented a complete redesign of MLRISC pseudo-ops. Now there
562      ought to never be any question of incompatabilities with
563      pseudo-op syntax expected by host assemblers.
564    
565      For now, only modules supporting GAS syntax are implemented
566      but more should follow, such as MASM, and vendor assembler
567      syntax, e.g. IBM as, Sun as, etc.
568    
569    ----------------------------------------------------------------------
570    Name: Matthias Blume
571    Date: 2001/11/14 11:52:00 EST
572    Tag: blume-20011114-srcname
573    Description:
574    
575    1. Routed the name of the current source file to mlriscgen where it
576       should be directly emitted into the code object.  (This last part
577       is yet to be done.)
578    
579    2. Some cleanup of the pgraph code to make it match the proposal that
580       I put out the other day.  (The proposal notwithstanding, things are
581       still in flux here.)
582    
583    ----------------------------------------------------------------------
584    Name: Lal George
585    Date: 2001/11/14 09:44:04 EST
586    Tag:
587    Description:
588    
589      Fix for a backpatching bug reported by Allen.
590    
591      Because the boundary between short and long span-dependent
592      instructions is +/- 128, there are an astounding number of
593      span-dependent instructions whose size is over estimated.
594    
595      Allen came up with the idea of letting the size of span
596      dependent instructions be non-monotonic, for a maxIter
597      number of times, after which the size must be monotonically
598      increasing.
599    
600      This table shows the number of span-dependent instructions
601      whose size was over-estimated as a function of maxIter, for the
602      file Parse/parse/ml.grm.sml:
603    
604         maxIter            # of instructions:
605            10                      687
606            20                      438
607            30                      198
608            40                        0
609    
610      In compiling the compiler, there is no significant difference in
611      compilation speed between maxIter=10 and maxIter=40. Actually,
612      my measurements showed that maxIter=40 was a tad faster than
613      maxIter=10! Also 96% of the  files in the compiler reach a fix
614      point within 13 iterations, so fixing maxIter at 40, while high,
615      is okay.
616    
617    ----------------------------------------------------------------------
618    Name: Matthias Blume
619    Date: 2001/10/31 15:25:00 EST
620    Tag: blume-20011031-pgraph
621    Description:
622    
623    CKIT:
624    * Changed the "Function" constructor of type Ast.ctype to carry optional
625      argument identifiers.
626    * Changed the return type of TypeUtil.getFunction accordingly.
627    * Type equality ignores the argument names.
628    * TypeUtil.composite tries to preserve argument names but gives up quickly
629      if there is a mismatch.
630    
631    installation script:
632    * attempts to use "curl" if available (unless "wget" is available as well)
633    
634    CM:
635    * has an experimental implementation of "portable graphs" which I will
636      soon propose as an implementation-independent library format
637    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
638    
639    NLFFI-LIB:
640    * some cleanup (all cosmetic)
641    
642    NLFFIGEN:
643    * temporarily disabled the mechanism that suppresses ML output for
644      C definitions whose identifiers start with an underscore character
645    * generate val bindings for enum constants
646    * user can request that only one style (light or heavy) is being used;
647      default is to use both (command-line arguments: -heavy and -light)
648    * fixed bug in handling of function types involving incomplete pointers
649    * generate ML entry points that take record arguments (i.e., using
650      named arguments) for C functions that have a prototype with named
651      arguments
652      (see changes to CKIT)
653    
654    ----------------------------------------------------------------------
655    Name: Allen Leung
656    Date: 2001/10/27 20:34:00 EDT
657    Tag: leunga-20011027-x86-fast-fp-call
658    Description:
659    
660       Fixed the bug described in blume-20010920-slowfp.
661    
662       The fix involves
663          1. generating FCOPYs in FSTP in ia32-svid
664          2. marking a CALL with the appropriate annotation
665    
666    ----------------------------------------------------------------------
667    Name: Matthias Blume
668    Date: 2001/10/16 11:32:00 EDT
669    Tag: blume-20011016-netbsd
670    Description:
671    
672    Underscore patch from Chris Richards (fixing problem with compiling
673    runtime system under recent NetBSD).
674    
675    ----------------------------------------------------------------------
676    Name: Allen Leung
677    Date: 2001/10/12 17:18:32 EDT 2001
678    Tag: leung-20011012-x86-printflowgraph
679    Description:
680    
681    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
682    
683    ----------------------------------------------------------------------
684    Name: Lal George
685    Date: 2001/10/11 23:51:34 EDT
686    Tag: george-20011011-too-many-instrs
687    Description:
688    
689    The representation of a program point never expected to see more
690    than 65536 instructions in a basic block!
691    
692    ----------------------------------------------------------------------
693    Name: Lal George
694    Date: 2001/10/09 09:41:37 EDT
695    Tag: george-20011008-mlrisc-labels
696    Description:
697    
698    Changed the machine description files to support printing of
699    local and global labels in assembly code, based on host assembler
700    conventions.
701    
702    ----------------------------------------------------------------------
703    Name: Matthias Blume
704    Date: 2001/09/25 15:25:00 EDT
705    Tag: blume-20010925-exninfo
706    Description:
707    
708    I provided a non-hook implementation of exnName (at the toplevel) and
709    made the "dummy" implementation of exnMessage (at the toplevel) more
710    useful: if nothing gets "hooked in", then at least you are going to
711    see the exception name and a message indicating why you don't see more.
712    
713    [For the time being, programs that need exnMessage and want to use
714    ml-build should either use General.exnMessage (strongly recommended) or
715    refer to structure General at some other point so that CM sees a
716    static dependency.]
717    
718    [Similar remarks go for "print" and "use":  If you want to use their
719    functionality in stand-alone programs generated by ml-build, then use
720    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
721    
722    ----------------------------------------------------------------------
723    Name: Matthias Blume
724    Date: 2001/09/20 17:28:00 EDT
725    Tag: blume-20010920-slowfp
726    Description:
727    
728    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
729    
730    ----------------------------------------------------------------------
731    Name: Matthias Blume
732    Date: 2001/09/20 17:20:00 EDT
733    Tag: blume-20010920-canonicalpaths
734    Description:
735    
736    0. Updated the BOOT file (something that I forgot to do earlier).
737    
738    1. Small internal change to CM so that it avoids "/../" in filenames
739       as much as possible (but only where it is safe).
740    
741    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
742       that arguments that contain delimiters are passed through correctly.
743       This change also means that all "special" arguments of the form
744       @SMLxxx... must come first.
745    
746    3. Changed install script to put relative anchor names for tool commands
747       into pathconfig.
748    
749    ----------------------------------------------------------------------
750    Name: Matthias Blume
751    Date: 2001/09/18 15:35:00 EDT
752    Tag: blume-20010918-readme11036
753    Description:
754    
755    Added README files.
756    
757    ----------------------------------------------------------------------
758    Name: Matthias Blume
759    Date: 2001/09/18 11:45:00 EDT
760    Tag: Release_110_36 (retag)
761    Description:
762    
763    Fixed mistake in config/preloads. Retagged as 110.36.
764    
765    ----------------------------------------------------------------------
766    Name: Matthias Blume
767    Date: 2001/09/18 09:40:00 EDT
768    Tag: Release_110_36_orig (tag changed)
769    Description:
770    
771    New version (110.36).  New bootfiles.
772    
773    ----------------------------------------------------------------------
774    Name: Matthias Blume
775    Date: 2001/09/14 16:15:00 EDT
776    Tag: blume-20010914-x86fastfp
777    Description:
778    
779    John committed some changes that Allen made, in particular a (hopefully)
780    correctly working version of the x86-fp module.
781    
782    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
783    flag to "true".  Everything seems to compile to a fixpoint ok, and
784    "mandelbrot" speeds up by about 15%.
785    
786    ----------------------------------------------------------------------
787    Name: Matthias Blume
788    Date: 2001/09/13 11:20:00 EDT
789    Tag: blume-20010913-minimal
790    Description:
791    
792    1. Stefan Monnier's patch to fix a miscompilation problem that
793       was brought to light by John Reppy's work on Moby.
794    
795    2. Implemented a minimal "structure Compiler" that contains just
796       "version" and "architecture".  The minimal version will be
797       available when the full version is not.  This is for backward-
798       compatibility with code that wants to test Compiler.version.
799    
800    ----------------------------------------------------------------------
801    Name: Matthias Blume
802    Date: 2001/08/28 14:03:00 EDT
803    Tag: blume-20010828-ml-lex
804    Description:
805    
806    Fix for bug 1581, received from Neophytos Michael.
807    
808    ----------------------------------------------------------------------
809    Name: Matthias Blume
810    Date: 2001/08/27 11:20:00 EDT
811    Tag: blume-20010827-readme11035
812    Description:
813    
814    Fleshed out the README file for 110.35.
815    
816    ----------------------------------------------------------------------
817    Name: Matthias Blume
818    Date: 2001/08/24 17:10:00 EDT
819    Tag: Release_110_35
820    Description:
821    
822    New version number (110.35).  New bootfiles.
823    
824    ----------------------------------------------------------------------
825    Name: Lal George
826    Date: 2001/08/24 13:47:18 EDT 2001
827    Tag: george-20010824-MLRISC-graphs
828    Description:
829    
830     removed clusters from MLRISC completely and replaced with graphs.
831    
832    ----------------------------------------------------------------------
833    Name: Matthias Blume
834    Date: 2001/08/23 17:50:00 EDT
835    Tag: blume-20010823-toplevel
836    Description:
837    
838    - some reorganization of the code that implements various kinds of
839      environments in the compiler (static, dynamic, symbolic, combined)
840    - re-implemented the EnvRef module so that evalStream works properly
841      (if the stream contains references to "use", "CM.make", etc.)
842    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
843    
844    ----------------------------------------------------------------------
845    Name: Matthias Blume
846    Date: 2001/08/20 15:50 EDT
847    Tag: blume20010820-slipup
848    Description:
849    
850    I forgot to commit a few files.  Here they are...
851    
852    ----------------------------------------------------------------------
853    Name: Matthias Blume
854    Date: 2001/08/20 15:35:00 EDT
855    Tag: blume-20010820-debugprof
856    Description:
857    
858    !!!! NEW BOOTFILES !!!!
859    
860    This is another round of reorganizing the compiler sources.  This
861    time the main goal was to factor out all the "instrumentation"
862    passes (for profiling and backtracing) into their own library.
863    The difficulty was to do it in such a way that it does not depend
864    on elaborate.cm but only on elabdata.cm.
865    
866    Therefore there have been further changes to both elaborate.cm and
867    elabdata.cm -- more "generic" things have been moved from the former
868    to the latter.  As a result, I was forced to split the assignment
869    of numbers indicating "primtyc"s into two portions: SML-generic and
870    SML/NJ-specific.  Since it would have been awkward to maintain,
871    I bit the bullet and actually _changed_ the mapping between these
872    numbers and primtycs.  The bottom line of this is that you need
873    a new set of bin- and bootfiles.
874    
875    I have built new bootfiles for all architectures, so doing a fresh
876    checkout and config/install.sh should be all you need.
877    
878    The newly created library's name is
879    
880        $smlnj/viscomp/debugprof.cm
881    
882    and its sources live under
883    
884        src/compiler/DebugProf
885    
886    ----------------------------------------------------------------------
887    Name: Matthias Blume
888    Date: 2001/08/15 17:15:00 EDT
889    Tag: blume-20010815-compreorg
890    Description:
891    
892    This is a first cut at reorganizing the CM libraries that make up the
893    core of the compiler.  The idea is to separate out pieces that could
894    be used independently by tools, e.g., the parser, the typechecker, etc.
895    
896    The current status is a step in this direction, but it is not quite
897    satisfactory yet.  Expect more changes in the future.
898    
899    Here is the current (new) organization...
900    
901        What used to be $smlnj/viscomp/core.cm is now divided into
902        six CM libraries:
903    
904             $smlnj/viscomp/basics.cm
905                           /parser.cm
906                           /elabdata.cm
907                           /elaborate.cm
908                           /execute.cm
909                           /core.cm
910    
911        The CM files for these libraries live under src/system/smlnj/viscomp.
912        All these libraries are proxy libraries that contain precisely
913        one CM library component.  Here are the locations of the components
914        (all within the src/compiler tree):
915    
916             Basics/basics.cm
917             Parse/parser.cm
918             ElabData/elabdata.cm
919             Elaborator/elaborate.cm
920             Execution/execute.cm
921             core.cm
922    
923         [This organization is the same that has been used already
924         for a while for the architecture-specific parts of the visible
925         compiler and for the old version of core.cm.]
926    
927         As you will notice, many source files have been moved from their
928         respective original locations to a new home in one of the above
929         subtrees.
930    
931         The division of labor between the new libraries is the following:
932    
933             basics.cm:
934                - Simple, basic definitions that pertain to many (or all) of
935                  the other libraries.
936             parser.cm:
937                - The SML parser, producing output of type Ast.dec.
938                - The type family for Ast is also defined and exported here.
939             elabdata.cm:
940                - The datatypes that describe input and output of the elaborator.
941                  This includes types, absyn, and static environments.
942             elaborator.cm:
943                - The SML/NJ type checker and elaborator.
944                  This maps an Ast.dec (with a given static environment) to
945                  an Absyn.dec (with a new static environment).
946                - This libraries implements certain modules that used to be
947                  structures as functors (to remove dependencies on FLINT).
948             execute.cm:
949                - Everything having to do with executing binary code objects.
950                - Dynamic environments.
951             core.cm:
952                - SML/NJ-specific instantiations of the elaborator and MLRISC.
953                - Top-level modules.
954                - FLINT (this should eventually become its own library)
955    
956    Notes:
957    
958    I am not 100% happy with the way I separated the elaborator (and its
959    data structures) from FLINT.  Two instances of the same problem:
960    
961        1. Data structures contain certain fields that carry FLINT-specific
962           information.  I hacked around this using exn and the property list
963           module from smlnj-lib.  But the fact that there are middle-end
964           specific fields around at all is a bit annoying.
965    
966        2. The elaborator calculates certain FLINT-related information.  I tried
967           to make this as abstract as I could using functorization, but, again,
968           the fact that the elaborator has to perform calculations on behalf
969           of the middle-end at all is not nice.
970    
971        3. Having to used exn and property lists is unfortunate because it
972           weakens type checking.  The other alternative (parameterizing
973           nearly *everything*) is not appealing, though.
974    
975    I removed the "rebinding =" warning hack because due to the new organization
976    it was awkward to maintain it.  As a result, the compiler now issues some of
977    these warnings when compiling init.cmi during bootstrap compilation. On
978    the plus side, you also get a warning when you do, for example:
979       val op = = Int32.+
980    which was not the case up to now.
981    
982    I placed "assign" and "deref" into the _Core structure so that the
983    code that deals with the "lazy" keyword can find them there.  This
984    removes the need for having access to the primitive environment
985    during elaboration.
986    
987    ----------------------------------------------------------------------
988    Name: Matthias Blume
989    Date: 2001/08/13
990    Tag: blume-20010813-closures
991    Description:
992    
993    This fix was sent to us by Zhong Shao.  It is supposed to improve the
994    performance of certain loops by avoiding needless closure allocation.
995    
996    ----------------------------------------------------------------------
997    Name: Lal George
998    Date: 2001/07/31 10:03:23 EDT 2001
999    Tag: george-20010731-x86-fmalloc
1000    Description: Fixed bug in x86 calls
1001    
1002        There was a bug where call instructions would mysteriously
1003        vanish. The call instruction had to be one that returned
1004        a floating point value.
1005    
1006    ----------------------------------------------------------------------
1007    Name: Lal George
1008    Date: 2001/07/19 16:36:29 EDT 2001
1009    Tag: george-20010719-simple-cells
1010    Description:
1011    
1012    I have dramatically simplified the interface for CELLS in MLRISC.
1013    
1014    In summary, the cells interface is broken up into three parts:
1015    
1016      1. CellsBasis : CELLS_BASIS
1017    
1018            CellsBasis is a top level structure and common for all
1019            architectures.  it contains the definitions of basic datatypes
1020            and utility  functions over these types.
1021    
1022      2. functor Cells() : CELLS
1023    
1024            Cells generates an interface for CELLS that incorporates the
1025            specific resources on the target architecture, such as the
1026            presence of special register classes, their number and size,
1027            and various useful substructures.
1028    
1029      3. <ARCH>CELLS
1030    
1031            e.g. SparcCells: SPARCCELLS
1032    
1033            <ARCH>CELLS usually contains additional bindings for special
1034            registers  on the architecture, such as:
1035    
1036                    val r0 : cell           (* register zero *)
1037                    val y : cell            (* Y register *)
1038                    val psr : cell          (* processor status register *)
1039                    ...
1040    
1041            The structure returned by applying the Cells functor is opened
1042            in this interface.
1043    
1044    The main implication of all this is that the datatypes for cells is
1045    split between CellsBasis and CELLS -- a fairly simple change for user
1046    code.
1047    
1048    In the old scheme the CELLS interface had a definitional binding of
1049    the form:
1050    
1051            signature CELLS = sig
1052    
1053               structure CellsBasis = CellsBasis
1054    
1055               ...
1056    
1057            end
1058    
1059    With all the sharing constraints that goes on in MLRISC, this old
1060    design  quickly leads to errors such as:
1061    
1062            "structure definition spec inside of sharing ... "
1063    
1064    
1065    and appears to require an unacceptable amount of sharing and where
1066    constraint hackery.
1067    
1068    I think this error message (the interaction of definitional specs and
1069    sharing) requires more explanation on our web page.
1070    
1071    ----------------------------------------------------------------------
1072    Name: Matthias Blume
1073    Date: 2001/07/19 15:00:00 EDT
1074    Tag: blume-20010719-libreorg
1075    Description:
1076    
1077    This update puts together a fairly extensive but straightforward change
1078    to the way the libraries that implement the interactive system are
1079    organized:
1080    
1081       The biggest change is the elimination of structure Compiler.  As a
1082       replacement for this structure, there is now a CM library
1083       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
1084       that exports all the substructures of the original structure Compiler
1085       directly.  So instead of saying Compiler.Foo.bar one now simply
1086       says Foo.bar.  (The CM libraries actually export a collection of
1087       structures that is richer than the collection of substructures of
1088       structure Compiler.)
1089    
1090       To make the transition smooth, there is a separate library called
1091       $smlnj/compiler/compiler.cm which puts together and exports the
1092       original structure Compiler (or at least something very close to it).
1093    
1094       There are five members of the original structure Compiler
1095       that are not exported directly but which instead became members
1096       of a new structure Backend (described by signature BACKEND).  These are:
1097       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
1098       Interact (: INTERACT), structure Machine (: MACHINE), and val
1099       architecture (: string).
1100    
1101       Structure Compiler.Version has become structure CompilerVersion.
1102    
1103       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
1104       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
1105       or x86, respectively.
1106       Each of these exports the same frontend structures that
1107       $smlnj/compiler.cm exports.  But they do not have a structure Backend
1108       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
1109       Hppa, PPC, Sparc, or X86, respectively.
1110    
1111       Library $smlnj/compiler/all.cm exports the union of the exports of
1112       $smlnj/compiler/<arch>.cm
1113    
1114       There are no structures <Arch>Compiler anymore, use
1115       $smlnj/compiler/<arch>.cm instead.
1116    
1117       Library host-compiler-0.cm is gone.  Instead, the internal library
1118       that instantiates CM is now called cm0.cm.  Selection of the host
1119       compiler (backend) is no longer done here but. (Responsibility for it
1120       now lies with $smlnj/compiler/current.cm.  This seems to be more
1121       logical.)
1122    
1123       Many individual files have been moved or renamed.  Some files have
1124       been split into multiple files, and some "dead" files have been deleted.
1125    
1126    Aside from these changes to library organization, there are also changes
1127    to the way the code itself is organized:
1128    
1129       Structure Binfile has been re-implemented in such a way that it no
1130       longer needs any knowledge of the compiler.  It exclusively deals
1131       with the details of binfile layout.  It no longer invokes the
1132       compiler (for the purpose of creating new prospective binfile
1133       content), and it no longer has any knowledge of how to interpret
1134       pickles.
1135    
1136       Structure Compile (: COMPILE) has been stripped down to the bare
1137       essentials of compilation.  It no longer deals with linking/execution.
1138       The interface has been cleaned up considerably.
1139    
1140       Utility routines for dealing with linking and execution have been
1141       moved into their own substructures.
1142    
1143       (The ultimate goal of these changes is to provide a light-weight
1144       binfile loader/linker (at least for, e.g., stable libraries) that
1145       does not require CM or the compiler to be present.)
1146    
1147    CM documentation has been updated to reflect the changes to library
1148    organization.
1149    
1150    ----------------------------------------------------------------------
1151    Name: Matthias Blume
1152    Date: 2001/07/10 17:30:00 EDT
1153    Tag: Release_110_34
1154    Description:
1155    
1156    Minor tweak to 110.34 (re-tagged):
1157    
1158      - README.html file added to CVS repository
1159      - runtime compiles properly under FreeBSD 3.X and 4.X
1160    
1161    ----------------------------------------------------------------------
1162    Name: Matthias Blume
1163    Date: 2001/07/10 17:30:00 EDT
1164    Tag: Release_110_34
1165    Description:
1166    
1167    New version number (110.34). New bootfiles.
1168    
1169    ----------------------------------------------------------------------
1170    Name: Matthias Blume
1171    Date: 2001/07/09 16:00:00 EDT
1172    Tag: blume-20010709-more-varargs
1173    Description:
1174    
1175    I changed the handling of varargs in ml-nlffigen again:
1176    The ellipsis ... will now simply be ignored (with an accompanying warning).
1177    
1178    The immediate effect is that you can actually call a varargs function
1179    from ML -- but you can't actually supply any arguments beyond the ones
1180    specified explicitly.  (For example, you can call printf with its format
1181    string, but you cannot pass additional arguments.)
1182    
1183    This behavior is only marginally more useful than the one before, but
1184    it has the advantage that a function or, more importantly, a function
1185    type never gets dropped on the floor, thus avoiding follow-up problems with
1186    other types that refer to the offending one.
1187    
1188    ----------------------------------------------------------------------
1189    Name: Matthias Blume
1190    Date: 2001/07/09 11:25:00 EDT
1191    Tag: blume-20010709-varargs
1192    Description:
1193    
1194    1. ckit-lib.cm now exports structure Error
1195    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1196       with a warning accompanied by a source location.  Moreover, it
1197       merely skips the offending function or type and proceeds with the
1198       rest of its work.u  As a result, one can safely feed C code containing
1199       "..." to ml-nlffigen.
1200    3. There are some internal improvements to CM, providing slightly
1201       more general string substitutions in the tools subsystem.
1202    
1203    ----------------------------------------------------------------------
1204    Name: Matthias Blume
1205    Date: 2001/06/27 15:10:00 EDT
1206    Tag: blume-20010627-concur
1207    Description:
1208    
1209    Fixed a small bug in CM's handling of parallel compilation.
1210    (You could observe the bug by Control-C-interrupting an ordinary
1211    CMB.make or CM.stabilize and then attaching some compile servers.
1212    The result was that all of a sudden the previously interrupted
1213    compilation would continue on its own.  This was because of
1214    an over-optimization: CM did not bother to clean out certain queues
1215    when no servers were attached "anyway", resulting in the contents
1216    of these queues to grab control when new servers did get attached.)
1217    
1218    There is also another minor update to the CM manual.
1219    
1220    ----------------------------------------------------------------------
1221    Name: Matthias Blume
1222    Date: 2001/06/26 16:15:00 EDT
1223    Tag: blume-20010626-cmdoc
1224    Description:
1225    
1226    Minor typo fixed in CM manual (syntax diagram for libraries).
1227    
1228    ----------------------------------------------------------------------
1229    Name: Matthias Blume
1230    Date: 2001/06/25 22:55:00 EDT
1231    Tag: blume-20010625-x86pc
1232    Description:
1233    
1234    Fixed a nasty bug in the X86 assembly code that caused signal
1235    handlers to fail (crash) randomly.
1236    
1237    ----------------------------------------------------------------------
1238    Name: Matthias Blume
1239    Date: 2001/06/25 12:05:00 EDT
1240    Tag: blume-20010625-nlffigen
1241    Description:
1242    
1243    This update fixes a number of minor bugs in ml-nlffigen as reported by
1244    Nick Carter <nbc@andrew.cmu.edu>.
1245    
1246      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
1247      2. Default names for generated files are now derived from the name of
1248         the C file *without its directory*.  In particular, this causes generated
1249         files to be placed locally even if the C file is in some system directory.
1250      3. Default names for generated signatures and structures are also derived
1251         from the C file name without its directory.  This avoids silly things
1252         like "structure GL/GL".
1253         (Other silly names are still possible because ml-nlffigen does not do
1254          a thorough check of whether generated names are legal ML identifiers.
1255          When in doubt, use command line arguments to force particular names.)
1256    
1257    ----------------------------------------------------------------------
1258    Name: Matthias Blume
1259    Date: 2001/06/21 12:25:00 EDT
1260    Tag: blume-20010621-eXene
1261    Description:
1262    
1263    eXene now compiles and (sort of) works again.
1264    
1265    The library name (for version > 110.33) is $/eXene.cm.
1266    
1267    I also added an new example in src/eXene/examples/nbody.  See the
1268    README file there for details.
1269    
1270    ----------------------------------------------------------------------
1271    Name: Matthias Blume
1272    Date: 2001/06/20 16:40:00 EDT
1273    Tag: blume-20010620-cml
1274    Description:
1275    
1276    CML now compiles and works again.
1277    
1278    Libraries (for version > 110.33):
1279    
1280      $cml/cml.cm            Main CML library.
1281      $cml/basis.cm          CML's version of $/basis.cm.
1282      $cml/cml-internal.cm   Internal helper library.
1283      $cml/core-cml.cm       Internal helper library.
1284      $cml-lib/trace-cml.cm  Tracing facility.
1285      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
1286    
1287    The installer (config/install.sh) has been taught how to properly
1288    install this stuff.
1289    
1290    ----------------------------------------------------------------------
1291    Name: Matthias Blume
1292    Date: 2001/06/19 17:55:00 EDT
1293    Tag: blume-20010619-instantiate
1294    Description:
1295    
1296    This un-breaks the fix for bug 1432.
1297    (The bug was originally fixed in 110.9 but I broke it again some
1298    time after that.)
1299    
1300    ----------------------------------------------------------------------
1301    Name: Matthias Blume
1302    Date: 2001/06/19 17:25:00 EDT
1303    Tag: blume-20010619-signals
1304    Description:
1305    
1306    This should (hopefully) fix the long-standing signal handling bug.
1307    (The runtime system was constructing a continuation record with an
1308    incorrect descriptor which would cause the GC to drop data on the floor...)
1309    
1310    ----------------------------------------------------------------------
1311    Name: Matthias Blume
1312    Date: 2001/06/15 15:05:00 EDT
1313    Tag: blume-20010615-moresparc
1314    Description:
1315    
1316    Here is a short late-hour update related to Sparc c-calls:
1317    
1318     -- made handling of double-word arguments a bit smarter
1319    
1320     -- instruction selection phase tries to collapse certain clumsily
1321        constructed ML-Trees; typical example:
1322    
1323            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
1324    
1325        This currently has no further impact on SML/NJ since mlriscGen does
1326        not seem to generate such patterns in the first place, and c-calls
1327        (which did generate them in the beginning) has meanwhile been fixed
1328        so as to avoid them as well.
1329    
1330    ----------------------------------------------------------------------
1331    Name: Matthias Blume
1332    Date: 2001/06/15 15:05:00 EDT
1333    Tag: blume-20010615-sparc
1334    Description:
1335    
1336    The purpose of this update is to provide an implementation of NLFFI
1337    on Sparc machines.
1338    
1339    Here are the changes in detail:
1340    
1341       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
1342       the Sparc implementation of the c-calls API.
1343       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
1344       for accessing the ML frame.  Thus, we have a real frame pointer and
1345       can freely modify %sp without need for an omit-frame-ptr phase.
1346       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
1347       this case.
1348       * ml-nlffigen has been taught to produce code for different architectures
1349       and calling conventions.
1350       * In a way similar to what was done in the x86 case, the Sparc
1351       backend uses its own specific extension to mltree.  (For example,
1352       it needs to be able to generate UNIMP instructions which are part
1353       of the calling convention.)
1354       * ml-nlffi-lib was reorganized to make it more modular (in particular,
1355       to make it easier to plug in new machine- and os-dependent parts).
1356    
1357    There are some other fairly unrelated bug fixes and cleanups as well:
1358    
1359       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
1360       that they properly share their libraries with existing SML/NJ libraries.
1361       * I fixed a minor cosmetic bug in CM, supressing certain spurious
1362       follow-up error messages.
1363       * Updates to CM/CMB documentation.
1364    
1365    TODO items:
1366    
1367       * MLRISC should use a different register as its asmTemp on the Sparc.
1368         (The current %o2 is a really bad choice because it is part of the
1369         calling conventions, so things might interfere in unexpected ways.)
1370    
1371    ----------------------------------------------------------------------
1372    Name: Matthias Blume
1373    Date: 2001/06/07
1374    Tag: blume-20010607-calls
1375    Description:
1376    
1377    A number of internal changes related to C calls and calling conventions:
1378    
1379    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
1380       number of bytes popped implicitly (by the callee).  In most cases
1381       this field is 0 but on x86/win32 it is some non-zero value.  This
1382       is information provided for the benefit of the "omit-frameptr" pass.
1383    2. The CALL instruction on the x86 carries a similar "pops" field.
1384       The instruction selection phase copies its value from the ML-Tree
1385       CALL statement.
1386    3. On all other architectures, the instruction selection phase checks
1387       whether "pops=0" and complains if not.
1388    4. The c-calls implementation for x86 now accepts two calling conventions:
1389       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
1390       up after the call and pops is set to 0.  For "stdcall", the caller
1391       does nothing, leaving the cleanup to the callee; pops is set to
1392       the number of bytes that were pushed onto the stack.
1393    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
1394       distinguish between "ccall" and "stdcall".
1395    6. The UNIMP instruction has been added to the supported Sparc instruction
1396       set. (This is needed for implementing the official C calling convention
1397       on this architecture.)
1398    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
1399       work with the latest CM.
1400    
1401    ----------------------------------------------------------------------
1402    Name: Matthias Blume
1403    Date: 2001/06/05 15:10:00 EDT
1404    Tag: blume-20010605-cm-index
1405    Description:
1406    
1407    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
1408    
1409    1. CM can now generate "index files".  These are human-readable files
1410       that list on a per-.cm-file basis each toplevel symbol defined or
1411       imported.  The location of the index file for
1412       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
1413       To enable index-file generation, set CM.Control.generate_index to true
1414       or export an environment-symbol: export CM_GENERATE_INDEX=true.
1415    
1416       The CM manual has been updated accordingly.
1417    
1418    2. I made some slight modifications to the c-calls API in MLRISC.
1419    
1420         a) There is now a callback to support saving/restoring of
1421            dedicated but caller-save registers around the actual call
1422            instruction.
1423         b) One can optionally specify a comment-annotation for the
1424            call instruction.
1425    
1426    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
1427       (For example, the comment annotation shows the C prototype of
1428        the function being called.)
1429    
1430    ----------------------------------------------------------------------
1431    Name: Matthias Blume
1432    Date: 2001/06/01 13:30:00 EDT
1433    Tag: blume-20010601-nlffi-cleanup
1434    Description:
1435    
1436    This is mostly a cleanup of MLFFI stuff:
1437    
1438       - some signature files have been put into a more exposed place
1439       - the ugly 'f type parameter is gone (simplifies types tremendously!)
1440       - ml-nlffigen changed accordingly
1441       - tutorial updated
1442    
1443    Other changes:
1444    
1445       - author's affiliation in CM manual(s) updated
1446       - some more recognized keywords added to Allen's sml.sty
1447    
1448    ----------------------------------------------------------------------
1449    Name: Matthias Blume
1450    Date: 2001/05/25 15:30:00 EDT
1451    Tag: blume-20010525-iptr
1452    Description:
1453    
1454      - put the official 110.33-README (as it appears on the ftp server) under
1455        CVS
1456      - fixed a small bug related to incomplete pointer types in
1457        ml-nlffigen
1458      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
1459        (it does not need the 'f type parameter)
1460    
1461    ----------------------------------------------------------------------
1462    Name: Matthias Blume
1463    Date: 2001/05/23 14:30:00 EDT
1464    Tag: Release_110_33
1465    Description:
1466    
1467    New version number (110.33).  New bootfiles.
1468    
1469    ----------------------------------------------------------------------
1470    Name: Matthias Blume
1471    Date: 2001/05/22 18:06:00 EDT
1472    Tag: blume-20010522-targets
1473    Description:
1474    
1475    Made install.sh use file config/targets.customized if it exists, falling
1476    back to config/targets if it doesn't.  This way one can have a customized
1477    version of the targets file without touching the "real thing", thus
1478    eliminating the constant fear of accidentally checking something bogus
1479    back into the CVS repository...  (File config/targets.customized must
1480    not be added to the repository!)
1481    
1482    ----------------------------------------------------------------------
1483    Name: Matthias Blume
1484    Date: 2001/05/22 16:30:00 EDT
1485    Tag: blume-20010522-minitut
1486    Description:
1487    
1488    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
1489       struct returns.
1490    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
1491       incomplete, preliminary documentation for NLFFI.
1492    
1493    ----------------------------------------------------------------------
1494    Name: Matthias Blume
1495    Date: 2001/05/14 11:30:00 EDT
1496    Tag: blume-20010514-script
1497    Description:
1498    
1499    Some bugs in install script fixed.
1500    
1501    In addition to that I also made a slight change to the NLFFI API:
1502    Functors generated by ml-nlffigen now take the dynamic library as a
1503    straight functor argument, not as a suspended one.  (The original
1504    functor code used to force the suspension right away anyway, so there
1505    was nothing gained by this complication of the interface.)
1506    
1507    ----------------------------------------------------------------------
1508    Name: Matthias Blume
1509    Date: 2001/05/11 14:35:00 EDT
1510    Tag: blume-20010511-ml-nlffi
1511    Description:
1512    
1513    I finally took the plunge and added my new FFI code to the main
1514    repository.  For x86-linux it is now ready for prime-time.
1515    
1516    There are two new subdirectories of "src":
1517    
1518      - ml-nlffi-lib:
1519           The utility library for programs using the FFI interface.
1520           Here is the implementation of $/c.cm and its associated low-level
1521           partners $/c-int.cm and $/memory.cm.
1522      - ml-nlffigen:
1523           A stand-alone program for generating ML glue code from C source
1524           code.
1525    
1526    Building ml-nlffigen requires $/ckit-lib.cm.
1527    
1528    The config/install.sh script has been updates to do the Right Thing
1529    (hopefully).
1530    
1531    Notice that the source tree for the C-Kit will not be put under "src"
1532    but directly under the installation root directory.  (This is the
1533    structure that currently exists on the CVS server when you check out
1534    module "sml".)  Fortunately, config/install.sh knows about this oddity.
1535    
1536    Bugs: No documentation yet.
1537    
1538    ----------------------------------------------------------------------
1539    Name: Matthias Blume
1540    Date: 2001/05/09 16:35:00 EDT
1541    Tag: blume-20010509-cpscontract
1542    Description:
1543    
1544    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
1545    wrapper/unwrapper elimination did not decrement usage counts and some
1546    dead variables got overlooked by the dead-up logic.)
1547    
1548    ----------------------------------------------------------------------
1549    Name: Lal George
1550    Date: 2001/05/08  17:26:09 EDT
1551    Tag: george-20010508-omit-frameptr
1552    Description:
1553    
1554    Changes to implement the omit-frame-pointer optimization to support
1555    raw C calls. For now, there is only support on the Intel x86, but
1556    other architectures will follow as more experience is gained with this.
1557    
1558    
1559    ----------------------------------------------------------------------
1560    Name: Matthias Blume
1561    Date: 2001/05/07 14:40:00 EDT
1562    Tag: blume-20010507-proxies
1563    Description:
1564    
1565    I made into "proxy libraries" all libraries that qualify for such a
1566    change.  (A qualifying library is a library that has another library or
1567    groups as its sole member and repeats that member's export list
1568    verbatim.  A proxy library avoids this repetition by omitting its export
1569    list, effectively inheriting the list that its (only) member exports.
1570    See the CM manual for more explanation.)
1571    The main effect is that explicit export lists for these libraries
1572    do not have to be kepts in sync, making maintenance a bit easier.
1573    
1574    I also added copyright notices to many .cm-files.
1575    
1576    Last but not least, I made a new set of bootfiles.
1577    
1578    ----------------------------------------------------------------------
1579    Name: Matthias Blume
1580    Date: 2001/05/04 17:00:00 EDT
1581    Tag: blume-20010504-cm-lsplit
1582    Description:
1583    
1584    0. John merged pending changes to $/smlnj-lib.cm
1585    
1586    1. Allen's previous change accidentally backed out of one of Lal's
1587       earlier changes.  I undid this mistake (re-introducing Lal's change).
1588    
1589    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
1590       within the compiler where applicable.  There is some code simplification
1591       because of that.
1592    
1593    3. The "split" phase (in FLINT) is now part of the default list of phases.
1594       Compiler.Control.LambdaSplitting.* can be used to globally control the
1595       lambda-splitting (cross-module-inlining) engine.  In addition to that,
1596       it can now also be controlled on a per-source basis: CM has been taught
1597       a new tool parameter applicable to ML source files.
1598    
1599       - To turn lambda-splitting off completely:
1600            local open Compiler.Control.LambdaSplitting in
1601                val _ = set Off
1602            end
1603       - To make "no lambda-splitting" the global default (but allow per-source
1604         overriding); this is the initial setting:
1605            local open Compiler.Control.LambdaSplitting in
1606                val _ = set (Default NONE)
1607            end
1608       - To make "lambda-splitting with aggressiveness a" the global default
1609         (and allow per-source overriding):
1610            local open Compiler.Control.LambdaSplitting in
1611                val _ = set (Default (SOME a))
1612            end
1613    
1614       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
1615         write (in the respective .cm-file):
1616            a.sml (lambdasplitting:off)
1617       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
1618            a.sml (lambdasplitting:on)
1619       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
1620         <a> is a decimal non-negative integer):
1621            a.sml (lambdasplitting:<a>)
1622       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
1623            a.sml (lambdasplitting:infinity)
1624       - To use the global default for a.sml:
1625            a.sml (lambdasplitting:default)
1626         or simply
1627            a.sml
1628    
1629    ----------------------------------------------------------------------
1630    Name: Allen Leung
1631    Date: 2001/05/04 01:57:00 EDT
1632    Tag: leunga-20010504-sync
1633    Description:
1634    
1635      MLRISC features.
1636    
1637      1. Fix to CMPXCHG instructions.
1638      2. Changed RA interface to allow annotations in callbacks.
1639      3. Added a new method to the stream interface to allow annotations updates.
1640    
1641    ----------------------------------------------------------------------
1642    Name: Matthias Blume
1643    Date: 2001/05/01 11:45:00 EDT
1644    Tag: blume-20010501-pcedittmp
1645    Description:
1646    
1647    Changed install.sh to use the current working directory instead of
1648    /usr/tmp for a temporary file (pcedittmp).  The previous choice
1649    of /usr/tmp caused trouble with MacOS X because of file premission
1650    problems.
1651    
1652    ----------------------------------------------------------------------
1653    Name: Matthias Blume
1654    Date: 2001/04/20 11:10:00 EDT
1655    Tag: blume-20010420-inMLflag
1656    Description:
1657    
1658     - added vp_limitPtrMask to vproc-state.h
1659       (for use by the raw-C-calls mechanism to implement proper interrupt
1660        handling)
1661     - made the ML compiler aware of various data-structure offsets so it
1662       can generate code for accessing the vp_inML flag and vp_limitPtrMask
1663     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
1664       raw C-calls
1665    
1666    ----------------------------------------------------------------------
1667    Name: Lal George
1668    Date: 2001/04/20 09:15:28 EDT
1669    Tag: george-20010420-macosX
1670    Description:
1671    
1672     - Changes to port to Mac OS X; Darwin.
1673    
1674     - In the process I found that sqrt was broken on the PPC, because the
1675       fsqrt instruction is not implemented.
1676    
1677    ----------------------------------------------------------------------
1678    Name: Matthias Blume
1679    Date: 2001/04/18 12:45:00 EDT
1680    Tag: blume-20010418-ccalls
1681    Description:
1682    
1683     - fixed two off-by-4 errors in the x86-specific c-calls implementation
1684       (this bug prevented structure arguments containing pointers from being
1685        passed correctly)
1686     - changed the raw-C-call code in mlriscGen.sml in such a way that
1687       structure arguments are represented as a pointer to the beginning
1688       of the structure (instead of having a series of synthesized arguments,
1689       one for each structure member)
1690    
1691     - made makeml script's verbosity level configurable via environment
1692       variable (MAKEML_VERBOSITY)
1693    
1694     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
1695       in rawmem-x86.sml; we are now using the real thing
1696    
1697    ----------------------------------------------------------------------
1698    Name: Matthias Blume
1699    Date: 2001/03/22 16:25:00 EST
1700    Tag: blume-20010322-bootfiles
1701    Description:
1702    
1703    Created a new set of bootfiles (for your automatic installation convenience).
1704    
1705    ----------------------------------------------------------------------
1706    Name: Matthias Blume
1707    Date: 2001/03/22 15:10:00 EST
1708    Tag: blume-20010322-rawmem-parcm
1709    Description:
1710    
1711    1. All "raw memory access" primitives for the new FFI are implemented now
1712       (at least on the x86).
1713    2. Some further cleanup of CM's parallel make mechanism.
1714    
1715    ----------------------------------------------------------------------
1716    Name: Matthias Blume
1717    Date: 2001/03/19 17:53:00 EST
1718    Tag: blume-20010319-parallel
1719    Description:
1720    
1721    Parallel make (using compile servers) now works again.
1722    
1723    To this end, CM.stabilize and CMB.make have been modified to work in
1724    two passes when compile servers are attached:
1725       1. Compile everything, do not perform stabilization; this pass
1726          uses compile servers
1727       2. Stabilize everything; this pass does not use compile servers
1728    If there are no compile servers, the two passes are combined into one
1729    (as before).  Splitting the passes increases the inherent parallelism
1730    in the dependency graph because the entire graph including all
1731    libraries is available at the same time.  This, in turn, improves
1732    server utilization.  The downside is that the master process will
1733    have to do some extra work after compilation is done (because for
1734    technical reasons it must re-read all the binfiles during stabilization).
1735    
1736    ----------------------------------------------------------------------
1737    Name: Matthias Blume
1738    Date: 2001/03/16 12:22:00 EST
1739    Tag: blume-20010316-bootfiles
1740    Description:
1741    
1742    Created a new set of bootfiles (for your automatic installation convenience).
1743    
1744    ----------------------------------------------------------------------
1745    Name: Matthias Blume
1746    Date: 2001/03/16 11:00:00 EST
1747    Tag: blume-20010316-MLTREE-fixup
1748    Description:
1749    
1750    This is a minor fixup for an (untagged) earlier commit by Allen.
1751    (A file was missing).
1752    
1753    ----------------------------------------------------------------------
1754    Name: Allen Leung
1755    Date: Mon Mar  5 18:54:57 EST 2001
1756    Tag: leunga-20010305-cut-support
1757    
1758    1. New support for alternative control-flow in MLTREE.
1759       Currently we support
1760    
1761          FLOW_TO(CALL ...., [k1,...,kn])
1762    
1763       This is needed for 'cuts to' in C-- and try/handle-like constructs
1764       in Moby
1765    
1766       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
1767    
1768    2. Register Allocator
1769    
1770       Changes in interface [from Fermin, John]
1771    
1772    3. Alpha 8-bit SLL support [Fermin]
1773    
1774    4. All architectures
1775    
1776       A new module (ClusterExpandCopies) for expanding parallel copies.
1777    
1778    ----------------------------------------------------------------------
1779    Name: Allen Leung
1780    Date: 2001/02/27 23:07:00 EST
1781    Tag: leunga-20010227-minor-stuff
1782    
1783    1. Alpha bug fix for CMOVNE
1784    2. Handle mltree COND(..,FCMP ...,...)
1785    3. Bug fix in simplifier
1786    
1787    ----------------------------------------------------------------------
1788    Name: Matthias Blume
1789    Date: 2001/01/30 17:50:00 EST
1790    Tag: blume-20010130-sync
1791    Description:
1792    
1793    This is just a minor update to sync my devel branch with the main brach.
1794    The only visible change is the addition of some README files.
1795    
1796    ----------------------------------------------------------------------
1797    Name: Matthias Blume
1798    Date: 2001/01/12 23:30:00 JST
1799    Tag: blume-20010112-bootfiles
1800    Description:
1801    
1802    Made a new set of bootfiles that goes with the current state of the
1803    repository.
1804    
1805    ----------------------------------------------------------------------
1806    Name: Matthias Blume
1807    Date: 2001/01/12 21:20:00 JST
1808    Tag: blume-20010112-sync
1809    Description:
1810    
1811    I am just flushing out some minor changes that had accumulated in
1812    my private branch in order to sync with the main tree.  (This is
1813    mainly because I had CVS trouble when trying to merge _into_ my
1814    private branch.)
1815    
1816    Most people should be completely unaffected by this.
1817    
1818    ----------------------------------------------------------------------
1819    Name: Allen Leung
1820    Date: Thu Jan 11 21:03:00 EST 2001
1821    Tag: leunga-20010111-labexp=mltree
1822    Description:
1823    
1824    1.  Removed the type LabelExp and replace it by MLTree.
1825    2.  Rewritten mltree-simplify with the pattern matcher tool.
1826    3.  There were some bugs in alpha code generator which would break
1827        64-bit code generation.
1828    4.  Redo the tools to generate code with the
1829    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
1830        are now generated by perl scripts.
1831    
1832    ----------------------------------------------------------------------
1833    Name: Matthias Blume
1834    Date: 2001/01/10 21:55:00 JST
1835    Tag: blume-20010110-rcc
1836    Description:
1837    
1838    The RCC stuff now seems to work (but only on the x86).
1839    This required hacking of the c-calls interface (and -implementation) in
1840    MLRISC.
1841    
1842    Normal compiler users should be unaffected.
1843    
1844    ----------------------------------------------------------------------
1845    Name: Matthias Blume
1846    Date: 2001/01/09 01:20:00 JST
1847    Tag: blume-20010109-rcc
1848    Description:
1849    
1850    This is a fairly big patch, flushing out a large number of pending
1851    changes that I made to my development copy over the last couple of days.
1852    
1853    Of practical relevance at this moment is a workaround for a pickling
1854    bug that Allen ran into the other day.  The cause of the bug itself is
1855    still unknown and it might be hard to fix it properly, but the
1856    workaround has some merits of its own (namely somewhat reducing pickling
1857    overhead for certain libraries).  Therefore, I think this solution should
1858    be satisfactory at this time.
1859    
1860    The rest of the changes (i.e., the vast majority) has to do with my
1861    ongoing efforts of providing direct support for C function calls from
1862    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
1863    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
1864    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
1865    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
1866    handle RCC.
1867    
1868    The last part (the changes to mlriscGen.sml) are still known to be
1869    wrong on the x86 and not implemented on all other architectures.  But
1870    the infrastructure is in place. I had to change a few functor
1871    signatures in the backend to be able to route the CCalls interface
1872    from MLRISC there, and I had to specialize the mltree type (on the
1873    x86) to include the necessary extensions. (The extensions themselves
1874    were already there and redy to go in MLRISC/x86).
1875    
1876    Everything should be very happy as soon as someone helps me with
1877    mlriscGen.sml...
1878    
1879    In any case, nothing of this should matter to anyone as long as the
1880    new primop is not being used (which is going to be the case unless you
1881    find it where I hid it :). The rest of the compiler is completely
1882    unaffected.
1883    
1884    ----------------------------------------------------------------------
1885    Name: Matthias Blume
1886    Date: 2001/01/05 00:30:00 JST
1887    Tag: blume-20010105-primops
1888    Description:
1889    
1890    Added some experimental support for work that I am doing right now.
1891    These changes mostly concern added primops, but there is also a new
1892    experimental C library in the runtime system (but currently not enabled
1893    anywhere except on Linux/X86).
1894    
1895    In the course of adding primops (and playing with them), I discovered that
1896    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
1897    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
1898    To recover, I made type information in INL_PRIM mandatory and changed
1899    prim.sml as well as built-in.sml accordingly.  The InLine structure now
1900    has complete, correct type information (i.e., no bottom types).
1901    
1902    Since all these changes mean that we need new binfiles, I also bumped the
1903    version number to 110.32.1.
1904    
1905    ----------------------------------------------------------------------
1906    Name: Matthias Blume
1907    Date: 2000/12/30 22:10:00 JST
1908    Tag: blume-20001230-various
1909    Description:
1910    
1911    Added proxy libraries for MLRISC and let MLRISC libraries refer
1912    to each other using path anchors.  (See CM manual for explanation.)
1913    
1914    Updated CM documentation.
1915    
1916    Fixed some bugs in CM.
1917    
1918    Implemented "proxy" libraries (= syntactic sugar for CM).
1919    
1920    Added "-quiet" option to makeml and changed runtime system accordingly.
1921    
1922    Added cleanup handler for exportML to reset timers and compiler stats.
1923    
1924    ----------------------------------------------------------------------
1925    Name: Lal George
1926    Date: 2000/12/22 22:22:58 EST 2000
1927    Tag: Release_110_32
1928    Description:
1929    
1930            Infinite precision used throughout MLRISC.
1931            see MLRISC/mltree/machine-int.sig
1932    
1933    ----------------------------------------------------------------------
1934    Name: Matthias Blume
1935    Date: 2000/12/22 23:16:00 JST
1936    Tag: blume-20001222-warn
1937    Description:
1938    
1939    Corrected wording and formatting of some CM warning message which I
1940    broke in my previous patch.
1941    
1942    ----------------------------------------------------------------------
1943    Name: Matthias Blume
1944    Date: 2000/12/22 21:20:00 JST
1945    Tag: blume-20001222-anchorenv
1946    Description:
1947    
1948    Fixed CM's handling of anchor environments in connection with CMB.make.
1949    
1950    ----------------------------------------------------------------------
1951    Name: Matthias Blume
1952    Date: 2000/12/22 13:15:00 JST
1953    Tag: blume-20001222-cleanup
1954    Description:
1955    
1956    Removed src/cm/ffi which does not (and did not) belong here.
1957    
1958    ----------------------------------------------------------------------
1959    Name: Matthias Blume
1960    Date: 2000/12/21 23:55:00 JST
1961    Tag: blume-20001221-exn
1962    Description:
1963    
1964    Probably most important: CM no longer silently swallows all exceptions
1965    in the compiler.
1966    Plus: some other minor CM changes.  For example, CM now reports some
1967    sizes for generated binfiles (code, data, envpickle, lambdapickle).
1968    
1969    ----------------------------------------------------------------------
1970    Name: Matthias Blume
1971    Date: 2000/12/15 00:01:05 JST
1972    Tag: blume-20001215-dirtool
1973    Description:
1974    
1975    - "dir" tool added.
1976    - improvements and cleanup to Tools structure
1977    - documentation updates
1978    
1979    ----------------------------------------------------------------------
1980    Name: Allen Leung
1981    Date: Thu Dec 14 03:45:24 EST 2000
1982    Description:
1983    Tag:  leunga-20001214-int-inf
1984    Description:
1985    
1986       In IntInf, added these standard functions, which are missing from our
1987    implementation:
1988    
1989        andb : int * int -> int
1990        xorb : int * int -> int
1991        orb  : int * int -> int
1992        notb : int -> int
1993         <<   : int * word -> int
1994        ~>>  : int * word -> int
1995    
1996       Not tested, I hope they are correct.
1997    
1998    ----------------------------------------------------------------------
1999    Name: Allen Leung
2000    Date: Fri Dec  8 19:23:26 EST 2000
2001    Description:
2002    Tag:  leunga-20001208-nowhere
2003    Description:
2004    
2005      Slight improvements to the 'nowhere' tool to handle OR-patterns,
2006    to generate better error messages etc.  Plus a brief manual.
2007    
2008    ----------------------------------------------------------------------
2009    Name: Lal George
2010    Date: 2000/12/08 09:54:02 EST 2000
2011    Tag: Release_110_31
2012    Description:
2013    
2014    - Version 110.31
2015    ----------------------------------------------------------------------
2016    Name: Allen Leung
2017    Date: Thu Dec  7 22:01:04 EST 2000
2018    Tag:  leunga-20001207-cell-monster-hack
2019    Description:
2020    
2021    Major MLRISC internal changes.  Affect all clients.
2022    Summary:
2023    
2024    1.  Type CELLS.cell = int is now replaced by a datatype.
2025        As a result, the old regmap is now gone.  Almost all interfaces
2026        in MLRISC change as a consequence.
2027    
2028    2.  A new brand version of machine description tool (v3.0) that generates
2029        modules expecting the new interface.  The old version is removed.
2030    
2031    3.  The RA interface has been further abstracted into two new functors.
2032        RISC_RA and X86RA.  These functors have much simpler interfaces.
2033        [See also directory MLRISC/demo.]
2034    
2035    4.  Some other new source->source code generation tools are available:
2036    
2037        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
2038        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
2039           I use this tool to generate the peephole optimizers---with the new
2040           cell type changes, peephole rules are becoming difficult to write
2041           without conditional pattern matching.
2042    
2043    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
2044        cover the entire MLRISC source tree so many things broke.
2045    
2046    6.  CM files have been moved to the subdirectory MLRISC/cm.
2047        They are moved because there are a lot of them and they clutter up the
2048        root dir.
2049    
2050    7.  More detailed documentation to come...
2051    
2052        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
2053        a makeml -rebuild first.  This is because of other other
2054        changes that Matthias has made (see below).
2055    
2056    
2057    ----------------------------------------------------------------------
2058    Name: Matthias Blume
2059    Date: 2000/11/30 23:12:00 JST
2060    Tag: blume-20001130-filereorg
2061    Description:
2062    
2063    Some manual updates and some file reorganizations in CM.
2064    
2065    ----------------------------------------------------------------------
2066    Name: Matthias Blume
2067    Date: 2000/11/24 17:45:00 JST
2068    Tag: blume-20001124-link
2069    Description:
2070    
2071    Drastically improved link traversal code for the case that the dynamic
2072    value was already loaded at bootstrap time.  As a result, CM and CMB
2073    now both load blazingly fast -- even on a very slow machine.  Also,
2074    memory consumption has been further reduced by this.
2075    
2076    Warning: The format of the PIDMAP file has changed.  THerefore, to
2077    bootstrap you have to do this:
2078    
2079    1. Run CMB.make
2080    2. Make a symbolic link for the boot directory:
2081         ln -s sml.boot.ARCH-OS xxx
2082    3. "Rebuild" the boot directory:
2083         ./makeml -boot xxx -rebuild sml ; rm xxx
2084    4. Boot normally:
2085          ./makeml
2086    
2087    ----------------------------------------------------------------------
2088    Name: Matthias Blume
2089    Date: 2000/11/21 21:20:00 JST
2090    Tag: blume-20001121-tools
2091    Description:
2092    
2093    Continued hacking on autoloading problem -- with success this time.
2094    Also changed tool-plugin mechanism.  See new CM manual.
2095    
2096    ----------------------------------------------------------------------
2097    Name: Matthias Blume
2098    Date: 2000/11/19 14:30:00 JST
2099    Tag:  blume-20001119-autoload
2100    Description:
2101    
2102    Some hacking to make autoloading faster.  Success for CMB, no success
2103    so far for CM.  There is a reduced structure CM' that autoloads faster.
2104    (This is a temporary, non-documented hack to be eliminated again when
2105    the general problem is solved.)
2106    
2107    ----------------------------------------------------------------------
2108    Name: Matthias Blume
2109    Date: 2000/11/17 14:10:00 JST
2110    Tag: blume-20001117-pickle-lib
2111    Description:
2112    
2113    1. Eliminated comp-lib.cm
2114    2. Made pickle-lib.cm
2115    3. Eliminated all uses of intset.sml (from comp-lib.cm)
2116    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
2117       equivalent constructs from smlnj-lib.cm (INtHashTable).
2118    5. Point 4. also goes for those uses of intmap.* in MLRISC.
2119       Duplicated intmap modules thrown out.
2120    6. Hunted down all duplicated SCC code and replaced it with
2121       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
2122    7. Rewrote Feedback module.
2123    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
2124       should be thrown out and equivalent modules from smlnj-lib.cm
2125       should be used (IntRedBlackSet, IntListSet, ...).
2126    
2127    Confirmed that compiler compiles to fixpoint.
2128    
2129    ----------------------------------------------------------------------
2130    Name: Allen Leung
2131    Date: 2000/11/10 18:00:00
2132    Tag: leunga-20001110-new-x86-fp
2133    
2134    A new x86 floating point code generator has been added.
2135    By default this is turned off.  To turn this on, do:
2136    
2137        CM.autoload "$smlnj/compiler.cm";
2138        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
2139    
2140    Changes:
2141    
2142    1.  Changed FTAN to FPTAN so that the assembly output is correct.
2143    2.  Changed the extension callback for FTANGENT to generate:
2144    
2145              fptan
2146              fstp  %st(0)
2147        instead of
2148              fptan
2149              fstpl ftempmem
2150    
2151    3.  Numerous assembly fixes for x86.
2152    
2153    5.  Cleaned up the machine code output module x86/x86MC.sml and added
2154        support for a whole bunch of instructions and addressing modes:
2155    
2156          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
2157          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
2158          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
2159          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
2160          fxch %st(n)
2161          fld %st(n)
2162          fst %st(n)
2163          fst mem
2164          fstp %st(n)
2165          fucom %st(n)
2166          fucomp %st(n)
2167    
2168        All these are now generated when the fast fp mode is turned on.
2169    
2170    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
2171    
2172    ----------------------------------------------------------------------
2173    Name: Matthias Blume
2174    Date: 2000/11/09 11:20:00 JST
2175    Tag: blume-20001109-scc
2176    Description:
2177    
2178    Eliminated some code duplication:
2179    
2180    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
2181       (Otherwise the functor is useless.)
2182    2. Used GraphSCCFn where SCCUtilFun was used previously.
2183    3. Got rid of SCCUtilFun (in comp-lib.cm).
2184    
2185    ----------------------------------------------------------------------
2186    Name: Lal George
2187    Date: 2000/11/06 09:02:21 EST 2000
2188    Tag: Release_110_30
2189    Description:
2190    
2191    - Version 110.30
2192    ----------------------------------------------------------------------
2193    Name: Matthias Blume
2194    Date: 2000/11/04 14:45:00
2195    Tag: blume-20001104-mlbuild
2196    Description:
2197    
2198    - Made ml-build faster on startup.
2199    - Documentation fixes.
2200    
2201    ----------------------------------------------------------------------
2202    Name: Matthias Blume
2203    Date: 2000/11/02 17:00:00 JST
2204    Tag: blume-20001102-condcomp
2205    Description:
2206    
2207    - Small tweaks to pickler -- new BOOTFILES!
2208    - Version bumped to 110.29.2.
2209    - Added conditional compilation facility to init.cmi (see comment there).
2210    ----------------------------------------------------------------------
2211    Name: Allen Leung
2212    Date: 2000/10/23 19:31:00
2213    Tag: leunga-20001023-demo-ra
2214    
2215    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
2216    2. Test programs for the graph library updated
2217    3. Some new MLRISC demo programs added
2218    
2219    ----------------------------------------------------------------------
2220    Name: Matthias Blume
2221    Date: 2000/08/31 22:15:00 JST
2222    Tag: blume-20001017-errmsg
2223    Description:
2224    
2225    More error message grief: Where there used to be no messages, there
2226    now were some that had bogus error regions.  Fixed.
2227    
2228    ----------------------------------------------------------------------
2229    Name: Matthias Blume
2230    Date: 2000/08/31 17:30:00 JST
2231    Tag: blume-20001017-v110p29p1
2232    Description:
2233    
2234    I made a version 110.29.1 with new bootfiles.
2235    
2236    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
2237              CM documentation changes and a small bugfix in CM's error reporting.
2238    
2239    ----------------------------------------------------------------------
2240    Name: Lal George
2241    Date: 2000/09/27 14:42:35 EDT
2242    Tag: george-20000927-nodestatus
2243    Description:
2244    
2245    Changed the type of the nodestatus, so that:
2246    
2247            SPILLED(~1)             is now SPILLED
2248            SPILLED(m) where m>=0   is now MEMREG(m)
2249            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
2250    
2251    ----------------------------------------------------------------------
2252    Name: Matthias Blume
2253    Date: 2000/09/07 14:45:00 JST
2254    Tag: blume-20000907-cmerrmsg
2255    Description:
2256    
2257    Small tweak to CM to avoid getting ML syntax error messages twice.
2258    
2259    ----------------------------------------------------------------------
2260    Name: Matthias Blume
2261    Date: 2000/08/31 18:00:00 JST
2262    Tag: blume-20000831-cvsbootfiles
2263    Description:
2264    
2265    New URL for boot files (because the 110.29 files on the BL server do
2266    now work correctly with my updated install scripts for yacc and lex).
2267    
2268    ----------------------------------------------------------------------
2269    Name: Matthias Blume
2270    Date: 2000/08/08 12:33:00 JST
2271    Tag: blume-20000808-manual
2272    Description:
2273    
2274    Tiny update to CM manual.
2275    
2276    ----------------------------------------------------------------------
2277    Name: Allen Leung
2278    Date: 2000/08/7 19:31:00
2279    Tag: leunga-20000807-a-whole-bunch-of-stuff
2280    
2281      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
2282    and Mobdy.
2283    
2284    1.  x86
2285    
2286       a.  Fixes to peephole module by John and Dan.
2287       b.  Assembly fix to SETcc by Allen.
2288       c.  Fix to c-call by John.
2289       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
2290       e.  Instruction selection optimization to SETcc as suggested by John.
2291    
2292           For example,
2293    
2294            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
2295    
2296           should generate:
2297    
2298            MOVL a, x
2299            SUBL b, x
2300            SHRL 31, x
2301    
2302    2.  IR stuff
2303    
2304         A bunch of new DJ-graph related algorithms added.  These
2305         speed up SSA construction.
2306    
2307    3.  SSA + Scheduling
2308    
2309         Added code for SSA and scheduling to the repository
2310    
2311    ----------------------------------------------------------------------
2312    Name: Lal George
2313    Date: 2000/07/27 11:53:14 EDT
2314    
2315    Tag: lal-20000727-linux-ppc
2316    Description:
2317    
2318     Made changes to support Linux PPC.
2319     p.s. I have confirmation that the 110.29 boot files work fine.
2320    
2321    ----------------------------------------------------------------------
2322    Name: Matthias Blume
2323    Date: 2000/07/27 17:40:00 JST
2324    Tag: blume-20000727-scripts
2325    Description:
2326    
2327    !!!! WARNING !!!!
2328    You must recompile the runtime system!
2329    !!!! WARNING !!!!
2330    
2331    This is basically another round of script-enhancements:
2332    
2333    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
2334       and undefine CM preprocessor symbols.
2335    
2336    2. ml-build avoids generating a new heap image if it finds that the
2337       existing one is still ok.  (The condition is that no ML file had to
2338       be recompiled and all ML files are found to be older that the heap
2339       file.)
2340    
2341       To make this work smoothly, I also hacked the runtime system as
2342       well as SMLofNJ.SysInfo to get access to the heap image suffix
2343       (.sparc-solaris, ...) that is currently being used.
2344    
2345       Moreover, the signature of CM.mk_standalone has changed.  See the
2346       CM manual.
2347    
2348    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
2349       CM manual for details.)
2350    
2351    4. More CM manual updates:
2352        - all of the above has been documented.
2353        - there is now a section describing the (CM-related) command line
2354          arguments that are accepted by the "sml" command
2355    
2356    ----------------------------------------------------------------------
2357    Name: Matthias Blume
2358    Date: 2000/07/25 16:20:00 JST
2359    Tag: blume-20000725-makedepend
2360    Description:
2361    
2362    Added a script called ml-makedepend.  This can be used in makefiles
2363    for Unix' make in a way very similar to the "makedepend" command for
2364    C.
2365    
2366    The script internally uses function CM.sources.
2367    
2368    Synopsis:
2369    
2370        ml-makedepend [-f makefile] cmfile targetname
2371    
2372    The default for the makefile is "makefile" (or "Makefile" should
2373    "makefile" not exist).
2374    
2375    ml-makedepend adds a cmfile/targetname-specific section to this
2376    makefile (after removing the previous version of this section).  The
2377    section contains a single dependency specification with targetname on
2378    the LHS (targetname is an arbitrary name), and a list of files derived
2379    from the cmfile on the RHS.  Some of the files on the RHS are
2380    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
2381    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
2382    path names.  The makefile writer is responsible for making sure that
2383    these variables have correct at the time "make" is invoked.
2384    
2385    ----------------------------------------------------------------------
2386    Name: Matthias Blume
2387    Date: 2000/07/22 23:30:00 JST
2388    Tag: blume-20000722-urlupdate
2389    Description:
2390    
2391    Changed BOOT and config/srcarchiveurl to point to BL server:
2392    
2393        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
2394    
2395    ----------------------------------------------------------------------
2396    Name: Matthias Blume
2397    Date: 2000/07/18 18:00:00 JST
2398    Tag: blume-20000718-Version_110_29
2399    Description:
2400    
2401    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
2402    
2403    2. Updated config/version to 110.29
2404    
2405    3. Updated config/srcarchiveurl
2406    
2407    3. New boot files!
2408       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
2409    
2410    ----------------------------------------------------------------------
2411    Name: Matthias Blume
2412    Date: 2000/07/11 13:58:00 JST
2413    Tag: blume-20000711-doctypo
2414    Description:
2415    
2416    Fixed a few typos in CM manual.
2417    
2418    ----------------------------------------------------------------------
2419    Name: Allen Leung
2420    Date: 2000/06/15 00:38:00
2421    Tag: leunga-20000704-sparc-x86
2422    
2423    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
2424    2. fix to x86 RET bug [found by Dan Grossman]
2425    3. sparc assembly bug fix for ticc instructions [found by Fermin]
2426    
2427       Affects c-- and moby only
2428    
2429    ----------------------------------------------------------------------
2430    Name: Matthias Blume
2431    Date: 2000/07/04 15:26:00
2432    Tag: blume-20000704-trigger
2433    Description:
2434    
2435    1. Improvements to CM manual.
2436    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
2437       of getting a back-trace.  The function, when called, raises an
2438       internal exception which explicitly carries the full back-trace history,
2439       so it is unaffected by any intervening handle-raise pairs ("trivial"
2440       or not).  The interactive loop will print that history once it arrives
2441       at top level.
2442       Short of having all exceptions implicitly carry the full history, the
2443       recommended way of using this facility is:
2444         - compile your program with instrumentation "on"
2445         - run it, when it raises an exception, look at the history
2446         - if the history is "cut off" because of some handler, go and modify
2447           your program so that it explicitly calls BTrace.trigger
2448         - recompile (still instrumented), and rerun; look at the full history
2449    
2450    ----------------------------------------------------------------------
2451    Name: Matthias Blume
2452    Date: 2000/07/03 15:36:00 JST
2453    Tag: blume-20000702-manual
2454    Description:
2455    
2456    Small corrections and updates to CM manual.
2457    
2458    ----------------------------------------------------------------------
2459    Name: Matthias Blume
2460    Date: 2000/06/29 16:04:00 JST
2461    Tag: blume-20000629-yacctool
2462    Description:
2463    
2464    Changes:
2465    
2466    1. Class "mlyacc" now takes separate arguments to pass options to
2467       generated .sml- and .sig-files independently.
2468    2. Corresponding CM manual updates.
2469    3. BTrace module now also reports call sites.  (However, for loop clusters
2470       it only shows from where the cluster was entered.)  There are associated
2471       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
2472    
2473    ----------------------------------------------------------------------
2474    Name: Matthias Blume
2475    Date: 2000/06/27 16:51:00 JST
2476    Tag: blume-20000627-noweb
2477    Description:
2478    
2479    Changes:
2480    
2481     1. Implemented "subdir" and "witness" options for noweb tool.
2482        This caused some slight internal changes in CM's tool implementation.
2483     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
2484        some remaining issues from earlier path anchor changes.
2485     3. Updated CM manual accordingly.
2486    
2487     4. Changed implementation of back-tracing so that I now consider it
2488        ready for prime-time.
2489    
2490        In particular, you don't have to explicitly trigger the back-trace
2491        anymore.  Instead, if you are running BTrace-instrumented code and
2492        there is an uncaught exception (regardless of whether or not it was
2493        raised in instrumented code), the top-level evalloop will print
2494        the back-trace.
2495    
2496        Features:
2497    
2498          - Instrumented and uninstrumented code work together seemlessly.
2499            (Of course, uninstrumented code is never mentioned in actual
2500             back-traces.)
2501    
2502          - Asymptotic time- and space-complexity of instrumented code is
2503            equal to that of uninstrumented code.  (This means that
2504            tail-recursion is preserved by the instrumentation phase.)
2505    
2506          - Modules whose code has been instrumented in different sessions
2507            work together without problem.
2508    
2509          - There is no penalty whatsoever on uninstrumented code.
2510    
2511          - There is no penalty on "raise" expressions, even in
2512            instrumented code.
2513    
2514        A potential bug (or perhaps it is a feature, too):
2515    
2516          A back-trace reaches no further than the outermost instrumented
2517          non-trivial "raise".  Here, a "trivial" raise is one that is the
2518          sole RHS of a "handle" rule.  Thus, back-traces reach trough
2519    
2520               <exp> handle e => raise e
2521    
2522          and even
2523    
2524               <exp> handle Foo => raise Bar
2525    
2526          and, of course, through
2527    
2528               <exp> handle Foo => ...
2529    
2530         if the exception was not Foo.
2531    
2532         Back-traces always reach right through any un-instrumented code
2533         including any of its "handle" expressions, trivial or not.
2534    
2535       To try this out, do the following:
2536    
2537         - Erase all existing binfiles for your program.
2538           (You may keep binfiles for those modules where you think you
2539            definitely don't need back-tracing.)
2540         - Turn on back-trace instrumentation:
2541              SMLofNJ.Internals.BTrace.mode (SOME true);
2542         - Recompile your program.  (I.e., run "CM.make" or "use".)
2543         - You may now turn instrumentation off again (if you want):
2544              SMLofNJ.Internals.BTrace.mode (SOME false);
2545         - Run your program as usual.  If it raises an exception that
2546           reaches the interactive toplevel, then a back-trace will
2547           automatically be printed.  After that, the toplevel loop
2548           will print the exception history as usual.
2549    
2550    ----------------------------------------------------------------------
2551    Name: Matthias Blume
2552    Date: 2000/06/26 09:56:46 JST
2553    Tag: blume-20000626-setup
2554    Description:
2555    
2556    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
2557          ML code before and after compiling a file (e.g., to set compiler
2558          flags)
2559    
2560    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
2561              - associated changes to btrace.sml (BTrace instrumentation pass)
2562              - cleaner implementation of btimp.sml (BTrace tracing and report
2563                module)
2564    
2565    CM manual: * new path encoding documented
2566               * description of setup-parameter to "sml" added
2567    
2568    The biggest user-visible change to back-tracing is that it is no
2569    longer necessary to compile all traced modules within the same
2570    session.  (This was a real limitation.)
2571    
2572    ----------------------------------------------------------------------
2573    Name: Matthias Blume
2574    Date: 2000/06/24 12:40:00 JST
2575    Tag: blume-20000624-startup
2576    Description:
2577    
2578    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
2579    bit too often -- to put it mildly. :)
2580    
2581    ----------------------------------------------------------------------
2582    Name: Matthias Blume
2583    Date: 2000/06/23 18:20:00 JST
2584    Tag: blume-20000623-btrace
2585    Description:
2586    
2587    This updates adds a backtrace facility to aid programmers in debugging
2588    their programs.  This involves the following changes:
2589    
2590    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
2591       keeping track of the current call stack.  When programs are compiled
2592       in a special mode, the compiler will insert calls to these hooks
2593       into the user program.
2594       "Hook" means that it is possible for different implementations of
2595       back-tracing to register themselves (at different times).
2596    
2597    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
2598       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
2599       It can be turned on using this call:
2600         SMLofNJ.Internals.BTrace.mode (SOME true);
2601       Turning it off again:
2602         SMLofNJ.Internals.BTrace.mode (SOME false);
2603       Querying the current status:
2604         SMLofNJ.Internals.BTrace.mode NONE;
2605       Annotated programs are about twice as big as normal ones, and they
2606       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
2607       where all hooks do nothing).  The slowdown with a plugin that is
2608       actually useful (such as the one supplied by default) is even greater,
2609       but in the case of the default plugin it is still only an constant
2610       factor (amortized).
2611    
2612    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
2613       with a sub-structure BTrace for controlling back-tracing.  In particular,
2614       the above-mentioned function "mode" controls whether the annotation
2615       phase is invoked by the compiler.  Another important function is
2616       "trigger": when called it aborts the current execution and causes
2617       the top-level loop to print a full back-trace.
2618    
2619    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
2620       for back-tracing.  It keeps track of the dynamic call stack and in
2621       addition to that it keeps a partial history at each "level" of that
2622       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
2623       d and b (at separate times, dynamically), then the report will show:
2624    
2625       GOTO   d
2626             /c
2627       GOTO  \b
2628       CALL   a
2629    
2630       This shows that there was an initial non-tail call of a, then a
2631       tail-call to b or c, looping behavior in a cluster of functions that
2632       consist of b and c, and then a goto from that cluster (i.e., either from
2633       b or from c) to d.
2634    
2635       Note that (depending on the user program) the amount of information
2636       that the back-trace module has to keep track of at each level is bounded
2637       by a constant.  Thus, the whole implementation has the same asymptotical
2638       complexity as the original program (both in space and in time).
2639    
2640    5. compiler/TopLevel/interact/evalloop.sml has been modified to
2641       handle the special exception SMLofNJ.Internals.BTrace.BTrace
2642       which is raised by the "trigger" function mentioned above.
2643    
2644    Notes on usage:
2645    
2646    - Annotated code works well together with unannotated code:
2647    Unannotated calls simply do not show up at all in the backtrace.
2648    
2649    - It is not a good idea to let modules that were annotated during
2650    different sessions run at the same time.  This is because the compiler
2651    chooses small integers to identify individual functions, and there
2652    will be clashes if different modules were compiled in separate sessions.
2653    (Nothing will crash, and you will even be told about the clashes, but
2654    back-trace information will in general not be useful.)
2655    
2656    - Back-tracing can be confused by callcc and capture.
2657    
2658    - The only way of getting a back-trace right now is to explicitly
2659    invoke the "trigger" function from your user program.  Eventually, we
2660    should make every exception carry back-trace information (if
2661    available).  But since this creates more overhead at "raise"-time
2662    (similar to the current exnHistory overhead), I have not yet
2663    implemented this.  (The implementation will be rather easy.)  With
2664    exceptions carrying back-trace information, this facility will be even
2665    more useful because users don't need to modify their programs...
2666    
2667    - While it is possible to compile the compiler with back-trace
2668    annotations turned on (I did it to get some confidence in
2669    correctness), you must make absolutely sure that core.sml and
2670    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
2671    be compiled with annotation because there is no core access yet, but
2672    if you compile btimp.sml with annotation, then the system will go into
2673    an infinite recursion and crash.)
2674    Since CM currently does not know about BTrace, the only way to turn
2675    annotations on and off for different modules of the compiler is to
2676    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
2677    this is awkward and clumsy.
2678    
2679    Sample sessions:
2680    
2681    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
2682    - SMLofNJ.Internals.BTrace.mode (SOME true);
2683    [autoloading]
2684    [autoloading done]
2685    val it = false : bool
2686    - structure X = struct
2687    -     fun main n = let
2688    -         fun a (x, 0) = d x
2689    -           | a (x, n) = b (x, n - 1)
2690    -         and b (x, n) = c (x, n)
2691    -         and c (x, n) = a (x, n)
2692    -         and d x = e (x, 3)
2693    -         and e (x, 0) = f x
2694    -           | e (x, n) = e (x, n - 1)
2695    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
2696    -           | f n = n * g (n - 1)
2697    -         and g n = a (n, 3)
2698    -     in
2699    -         f n
2700    -     end
2701    - end;
2702    structure X : sig val main : int -> int end
2703    - X.main 3;
2704    *** BACK-TRACE ***
2705    GOTO   stdIn:4.2-13.20: X.main[2].f
2706    GOTO-( stdIn:4.2-13.20: X.main[2].e
2707    GOTO   stdIn:4.2-13.20: X.main[2].d
2708         / stdIn:4.2-13.20: X.main[2].a
2709         | stdIn:4.2-13.20: X.main[2].b
2710    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2711    CALL   stdIn:4.2-13.20: X.main[2].g
2712    GOTO   stdIn:4.2-13.20: X.main[2].f
2713    GOTO-( stdIn:4.2-13.20: X.main[2].e
2714    GOTO   stdIn:4.2-13.20: X.main[2].d
2715         / stdIn:4.2-13.20: X.main[2].a
2716         | stdIn:4.2-13.20: X.main[2].b
2717    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2718    CALL   stdIn:4.2-13.20: X.main[2].g
2719    GOTO   stdIn:4.2-13.20: X.main[2].f
2720    GOTO-( stdIn:4.2-13.20: X.main[2].e
2721    GOTO   stdIn:4.2-13.20: X.main[2].d
2722         / stdIn:4.2-13.20: X.main[2].a
2723         | stdIn:4.2-13.20: X.main[2].b
2724    GOTO-\ stdIn:4.2-13.20: X.main[2].c
2725    CALL   stdIn:4.2-13.20: X.main[2].g
2726    GOTO   stdIn:4.2-13.20: X.main[2].f
2727    CALL   stdIn:2.15-17.4: X.main[2]
2728    -
2729    
2730    (Note that because of a FLINt bug the above code currently does not
2731    compile without BTrace turned on.)
2732    
2733    Here is another example, using my modified Tiger compiler:
2734    
2735    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
2736    - SMLofNJ.Internals.BTrace.mode (SOME true);
2737    [autoloading]
2738    [autoloading done]
2739    val it = false : bool
2740    - CM.make "sources.cm";
2741    [autoloading]
2742    ...
2743    [autoloading done]
2744    [scanning sources.cm]
2745    [parsing (sources.cm):parse.sml]
2746    [creating directory CM/SKEL ...]
2747    [parsing (sources.cm):tiger.lex.sml]
2748    ...
2749    [wrote CM/sparc-unix/semant.sml]
2750    [compiling (sources.cm):main.sml]
2751    [wrote CM/sparc-unix/main.sml]
2752    [New bindings added.]
2753    val it = true : bool
2754    - Main.compile ("../testcases/merge.tig", "foo.out");
2755    *** BACK-TRACE ***
2756    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
2757    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2758    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2759    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2760    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2761    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2762    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
2763         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
2764    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
2765    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2766    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
2767    CALL   main.sml:1.18-118.4: Main.compile[2]
2768    -
2769    
2770    ----------------------------------------------------------------------
2771    Name: Matthias Blumen
2772    Date: 2000/06/21 18:00:00 JST
2773    Tag: blume-20000621-manual
2774    Description:
2775    
2776    CM manual update: Path environments documented.
2777    
2778    ----------------------------------------------------------------------
2779    Name: Matthias Blume
2780    Date: 2000/06/19 13:40:00
2781    Tag: blume-20000619-manual
2782    Description:
2783    
2784    CM manual and system/README update.  This only covers the fact that
2785    there are no more implicit anchors.  (Path environments and the "bind"
2786    option to "cm" have yet to be documented.)
2787    
2788    ----------------------------------------------------------------------
2789    Name: Matthias Blume
2790    Date: 2000/06/19 11:05:00 JST
2791    Tag: blume-20000619-chdir-bugfix
2792    Description:
2793    
2794    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
2795    
2796    ----------------------------------------------------------------------
2797    Name: Matthias Blume
2798    Date: 2000/06/18 22:00:10 JST
2799    Tag: blume-20000618-implicit-anchors-really-gone
2800    Description:
2801    
2802    I updates the previous HISTORY entry where I forgot to mention that
2803    implicit anchors are no longer with us.
2804    
2805    The current update also gets rid of the (now useless) controller
2806    CM.Control.implicit_anchors.
2807    
2808    ----------------------------------------------------------------------
2809    Name: Matthias Blume
2810    Date: 2000/06/16 17:30:00 JST
2811    Tag: blume-20000616-anchorenv
2812    Description:
2813    
2814    This patch implements the long anticipated (just kidding :) "anchor
2815    environment" mechanism.  In the course of doing this, I also
2816    re-implemented CM's internal "SrcPath" module from scratch.  The new
2817    one should be more robust in certain boundary cases.  In any case, it
2818    is a lot cleaner than its predecessor (IMHO).
2819    
2820    This time, although there is yet another boot file format change, I
2821    kept the unpickler backward-compatible.  As a result, no new bootfiles
2822    are necessary and bootstrapping is straightforward.  (You cannot read
2823    new bootfiles into an old system, but the other way around is no
2824    problem.)
2825    
2826    Visible changes:
2827    
2828    ** 0. Implicit path anchors (without the leading $-symbol) are no
2829    longer recognized at all. This means that such path names are not
2830    illegal either.  For example, the name basis.cm simply refers to a
2831    local file called "basis.cm" (i.e, the name is an ordinary path
2832    relative to .cm-files directory).  Or, to put it differently, only
2833    names that start with $ are anchored paths.
2834    
2835    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
2836    vanished.
2837    
2838    John (Reppy) had critizised this as soon as I originally proposed and
2839    implemented it, but at that time I did not really deeply believe
2840    him. :) Now I came full-circle because I need the $<singlearc> syntax
2841    in another place where it cannot be seen as an abbreviation for
2842    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
2843    seems to mean (i.e., it "expands" into the corresponding anchor
2844    value).
2845    
2846    However, when paths are used as members in CM description files, it
2847    continues to be true that there must be at least another arc after the
2848    anchor.  This is now enforced separately during semantic analysis
2849    (i.e., from a lexical/syntactical point of view, the notation is ok.)
2850    
2851    ** 2. The "cm" class now accepts an option "bind".  The option's value
2852    is a sub-option list of precisely two items -- one labeled "anchor"
2853    and the other one labeled "value".  As you might expect, "anchor" is
2854    used to specify an anchor name to be bound, and "value" specifies what
2855    the anchor is being bound to.
2856    
2857    The value must be a directory name and can be given in either standard
2858    syntax (including the possibility that it is itself an anchored path)
2859    or native syntax.
2860    
2861    Examples:
2862    
2863       foo.cm (bind:(anchor:bar value:$mystuff/bar))
2864       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
2865    
2866    and so on.
2867    
2868    The meaning of this is that the .cm-file will be processed with an
2869    augmented anchor environment where the given anchor(s) is/are bound to
2870    the given values(s).
2871    
2872    The rationale for having this feature is this: Suppose you are trying
2873    to use two different (already stable) libraries a.cm and b.cm (that
2874    you perhaps didn't write yourself).  Further, suppose each of these
2875    two libraries internally uses its own auxiliary library $aux/lib.cm.
2876    Normally you would now have a problem because the anchor "lib" can not
2877    be bound to more than one value globally.  Therefore, the project that
2878    uses both a.cm and b.cm must locally redirect the anchor to some other
2879    place:
2880    
2881       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
2882       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
2883    
2884    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
2885    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
2886    
2887    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
2888    you when you do so at the time of CM.stabilize).  Therefore, you can
2889    also use an anchored path as the value:
2890    
2891      a.cm (bind:(anchor:lib value:$a-lib))
2892      b.cm (bind:(anchor:lib value:$b-lib))
2893    
2894    Now you can globally configure (using the usual CM.Anchor.anchor or
2895    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
2896    itself is always locally bound, setting it globally is no longer
2897    meaningful or necessary (but it does not hurt either).  In fact, "lib"
2898    can still be used as a global anchor for separate purposes.  As a
2899    matter of fact, one can locally define "lib" in terms of a global
2900    "lib":
2901    
2902      a.cm (bind:(anchor:lib value:$lib/a))
2903      b.cm (bind:(anchor:lib value:$lib/b))
2904    
2905    ** 3: The encoding of path names has changed.  This affects the way
2906    path names are shown in CM's progress report and also the internal
2907    protocol encoding used for parallel make.
2908    
2909    The encoding now uses one or more ':'-separated segments.  Each
2910    segments corresponds to a file that has been specified relative to the
2911    file given by its preceding segment.  The first segment is either
2912    relative to the CWD, absolute, or anchored.  Each segment itself is
2913    basically a Unix pathname; all segments but the first are relative.
2914    
2915    Example:
2916    
2917       $foo/bar/baz.cm:a/b/c.sml
2918    
2919    This path denotes the file bar/a/b/c.sml relative to the directory
2920    denoted by anchor "foo".  Notice that the encoding also includes
2921    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
2922    paths are resolved relative to the .cm-files directory, so baz.cm must
2923    be ignored to get the "real" pathname.
2924    
2925    To make this fact more obvious, CM puts the names of such "virtual
2926    arcs" into parentheses when they appear in progress reports. (No
2927    parentheses will appear in the internal protocol encoding.)  Thus,
2928    what you really see is:
2929    
2930      $foo/bar/(baz.cm):a/b/c.sml
2931    
2932    I find this notation to be much more informative than before.
2933    
2934    Another new feature of the encoding is that special characters
2935    including parentheses, colons, (back)slashes, and white space are
2936    written as \ddd (where ddd is the decimal encoding of the character).
2937    
2938    *** The CM manual still needs to be updated.
2939    
2940    ----------------------------------------------------------------------
2941    Name: Allen Leung
2942    Date: 2000/06/15 00:38:00
2943    Tag: leunga-20000615-x86-peephole
2944    
2945    x86 Peephole fix by Fermin.  Affects c-- and moby only.
2946    
2947    ----------------------------------------------------------------------
2948    Name: Matthias Blume
2949    Date: 2000/06/12 11:40:00
2950    Tag: blume-20000612-parmakefix
2951    Description:
2952    
2953    More cleanup after changing the file naming scheme: This time I
2954    repaired the parallel make mechanism for CMB.make which I broke earlier.
2955    
2956    ----------------------------------------------------------------------
2957    Name: Allen Leung
2958    Date: 2000/06/09 01:25:00
2959    Tag: leunga-20000609-various
2960    
2961    None of these things should affect normal SML/NJ operations
2962    
2963    1. Peephole improvements provided by Fermin (c--)
2964    2. New annotation DEFUSE for adding extra dependence (moby)
2965    3. New X86 LOCK instructions (moby)
2966    4. New machine description language for reservation tables (scheduling)
2967    5. Fixes to various optimization/analysis modules (branch chaining, dominator
2968       trees etc.)
2969    6. I've changed the CM files so that they can work with versions
2970       110.0.6, 110.25 and 110.28
2971    
2972    ----------------------------------------------------------------------
2973    Name: Matthias Blume
2974    Date: 2000/06/09 12:40:00
2975    Tag: blume-20000609-log
2976    Description:
2977    
2978    - Removed all(?) remaining RCS Log entries from sources.
2979    
2980    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
2981      anchored paths).
2982    
2983    ----------------------------------------------------------------------
2984    Name: Matthias Blume
2985    Date: 2000/06/07 17:00:00 JST
2986    Tag: blume-20000607-no-implicit-anchors
2987    Description:
2988    
2989    1. This update changes the default setting for
2990    CM.Control.implicit_anchors from true to false.  This means that
2991    implicit anchors are no longer permitted by default.  I also tried to
2992    make sure that nothing else still relies on implicit anchors.
2993    (This is the next step on the schedule towards a CM that does not even
2994    have the notion of implicit anchors anymore.)
2995    
2996    2. More CM manual updates.
2997    
2998    3. I managed to track down and fix the pickling bug I mentioned last
2999    time.  Because of the previously existing workaround, this entails no
3000    immediate practical changes.
3001    
3002    ----------------------------------------------------------------------
3003    Name: Matthias Blume
3004    Date: 2000/06/06 11:15:00 JST
3005    Tag: blume-20000606-lazierpickle
3006    Description:
3007    
3008    !!!! NEW BOOT FILES !!!!
3009    
3010    * The main purpose of this update is to make library pickles lazier in
3011    order to reduce the initial space penalty for autoloading a library.
3012    As a result, it is now possible to have $smlnj/compiler.cm
3013    pre-registered.  This should take care of the many complaints or
3014    inquiries about missing structure Compiler.  This required changes to
3015    CM's internal data structures and small tweaks to some algorithms.
3016    
3017    As a neat additional effect, it is no longer necessary (for the sake
3018    of lean heap image files) to distinguish between a "minimal" CM and a
3019    "full" CM.  Now, there is only one CM (i.e., the "full" version:
3020    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
3021    interactive top level. ($smlnj/cm/minimal.cm is gone.)
3022    
3023    To make the life of compiler-hackers easier, "makeml" now also
3024    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
3025    words, after you bootstrap a new sml for the first time, you will not
3026    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
3027    around you will still have to do it, though.)
3028    
3029    * A second change consists of major updates to the CM manual.  There
3030    are now several appendices with summary information and also a full
3031    specification of the CM description file syntax.
3032    
3033    * In directory src/system I added the script "allcross".  This script
3034    invokes sml and cross-compiles the compiler for all supported
3035    architectures.  (Useful when providing a new set of boot files.)
3036    
3037    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
3038    added a small tweak to pickle-util.sml to work around this problem,
3039    but it is not a proper fix yet.  I will investigate further.  (The
3040    effect of the bug was an inflation of library pickle size.)
3041    
3042    * Version number increased to 110.28.1 (to avoid compatibility problems).
3043    
3044    ----------------------------------------------------------------------
3045    Name: Allen Leung
3046    Date: 2000/05/25 17:28 EDT
3047    Tag: leunga-20000525-ra
3048    Description:
3049    
3050      Fixed a bug in freezing phase of the register allocator.
3051    
3052    ----------------------------------------------------------------------
3053    Name: Allen Leung
3054    Date: 2000/05/15 22:53 EDT
3055    Tag: leunga-20000515-alpha-x86-ra
3056    Description:
3057    
3058      1. Alpha
3059    
3060          Slight cleanup.  Removed the instruction SGNXL
3061    
3062      2. X86
3063    
3064          Added the following instructions to the instruction set:
3065    
3066            ROLx, RORx,
3067            BTx, BTSx, BTLx, BTRx,
3068            XCHGx, and variants with the LOCK prefix
3069    
3070      3. Register Allocation
3071    
3072          The module ra-rewrite-with-renaming has been improved.
3073    
3074      These have no effect on SML/NJ.
3075    
3076    ----------------------------------------------------------------------
3077    Name: Matthias Blume
3078    Date: 2000/05/15 16:20:00 JST
3079    Tag: blume-20000515-lightrebuild
3080    Description:
3081    
3082    1. I added an alternative to "-rebuild" to "makeml".  The difference is
3083       that prior to calling CMB.make' the CM-variable "LIGHT" will be
3084       defined.  In effect, the command will not build any cross-compiler
3085       backends and therefore finish more quickly.
3086    
3087       The "fixpt" script also takes a "-light" switch to be able to use
3088       this new facility while compiling for a fixpoint.
3089    
3090    2. I replaced all mentions of anchored paths in group owner specifications
3091       with simple relative paths (usually starting with "..").
3092       The rationale is that a library's internal workings should not be
3093       compromised by the lack of some anchor.  (An anchor is necessary
3094       for someone who wants to refer to the library by an anchored path,
3095       but it should not be necessary to build the same library in the first
3096       place.)
3097    
3098    3. I changed the way CM's tool mechanism determines the shell command
3099       string used for things like ml-yacc etc. so that it does not break
3100       when CM.Control.implicit_anchors is turned off.
3101    
3102    ----------------------------------------------------------------------
3103    Name: Matthias Blume
3104    Date: 2000/05/12 18:20:00 JST
3105    Tag: blume-20000512-ml-build
3106    Description:
3107    
3108    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
3109    getting installed properly (by config/install.sh).
3110    
3111    ----------------------------------------------------------------------
3112    Name: Matthias Blume
3113    Date: 2000/05/12 17:30:00 JST
3114    Tag: blume-20000512-anchors
3115    Description:
3116    
3117    !!! NEW BOOT FILES !!!
3118    
3119    This change is in preparation of fading out support for "implicitly
3120    anchored path names".  I went through all sources and used the
3121    explicit (and relatively new) $-notation.  See system/README and the
3122    CM manual for more info on this.
3123    
3124    I also modified the anchoring scheme for some things such as "smlnj",
3125    "MLRISC", "cm", etc. to take advantage of the fact that explicit
3126    anchors are more expressive: anchor name and first arc do not have to
3127    coincide.  This entails the following user-visible change:
3128    
3129    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
3130    particular, when you fire up sml with a command-line argument, say,
3131    e.g.:
3132    
3133       sml '$smlnj/cmb.cm'
3134    
3135    At the ML toplevel prompt:
3136    
3137       CM.autoload "$smlnj/cmb.cm";
3138    
3139    There is also a new controller in CM.Control that can be used to turn
3140    off all remaining support for implicit anchors by saying:
3141    
3142        CM.autoload "$smlnj/
3143        #set CM.Control.implicit_anchors false;
3144    
3145    This causes CM to reject implicitly anchored paths.  This is (for the
3146    time being) less permissive than the "final" version where there will
3147    be no more such implicit anchors and relative paths will be just that:
3148    relative.
3149    
3150    The next step (version after next version?) will be to make the
3151    default for CM.Control.implicit_anchors false.  After the dust has
3152    settled, I can then produce the "final" version of this...
3153    
3154    Note: Since bootstrapping is a bit tricky, I provided new boot files.
3155    
3156    ----------------------------------------------------------------------
3157    Name: Matthias Blume
3158    Date: 2000/05/11 16:30:00 JST
3159    Tag: blume-20000511-sources
3160    Description:
3161    
3162    The main change is that I added function CM.sources as a generalized
3163    version of the earlier CM.makedepend.  This entails the following
3164    additional changes:
3165    
3166      - CM.makedepend has been dropped.
3167    
3168      - CM manual has been updated.
3169    
3170      - TOOLS signature and API have been changed.
3171    
3172    ----------------------------------------------------------------------
3173    Name: Allen Leung
3174    Date: 2000/05/10 21:17 EDT
3175    Tag: leunga-20000510-moby-c--ssa
3176    Description:
3177    
3178      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
3179    None of these affect SML/NJ.
3180    
3181    1. Register Allocation
3182    
3183        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
3184           This module tries to remove local (i.e. basic block level) redundancies
3185           during spilling.
3186    
3187        b. A new framework for performing region based register allocation.
3188           Not yet entirely functional.
3189    
3190    2. X86
3191    
3192       a. DefUse for POP was missing the stack pointer [found by Lal]
3193       b. Reload for CALL was incorrect in X86Spill [found by John]
3194       c. Various fixes in X86Spill so that it can be used correctly for
3195          the new spilling module.
3196    
3197    3. SSA/IR
3198    
3199       a. New module ir/dj-dataflow.sml implements elimination based
3200          data flow analysis.
3201    
3202    4. MLRiscGen
3203    
3204       a. Fix for gc type annotation
3205    
3206    5. MDGen
3207    
3208       Various fixes for machine description -> ml code translation.  For ssa
3209       only.
3210    
3211    ----------------------------------------------------------------------
3212    Name: Allen Leung
3213    Date: 2000/05/08 22:17 EDT
3214    Tag: leunga-20000508-labexp
3215    Description:
3216    
3217      Fermin has found a few assembly problems with constant expressions
3218      generated in LabelExp.  Mostly, the problems involve extra parentheses,
3219      which choke on dumb assemblers.  This is his fix.
3220    
3221    ----------------------------------------------------------------------
3222    Name: Dave MacQueen
3223    Date: 2000/04/09 14:00 EDT
3224    Tag: dbm-20000502-Version_110_28
3225    Description:
3226    
3227    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
3228    
3229    2. Updated config/version to 110.28
3230    
3231    3. Updated config/srcarchiveurl
3232    
3233    3. New boot files!
3234       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
3235    
3236    ----------------------------------------------------------------------
3237    Name: Matthias Blume
3238    Date: 2000/05/01 19:05:00 JST
3239    Tag: blume-20000501-noweb
3240    Description:
3241    
3242    A new noweb tool has been added.  The existing system is entirely
3243    unaffected by this, but some CM users have asked for renewed noweb
3244    support.  Everything is documented in the CM manual.
3245    
3246    New (plugin) libraries:
3247    
3248       noweb-tool.cm
3249       nw-ext.cm
3250    
3251    ----------------------------------------------------------------------
3252    Name: Dave MacQueen
3253    Date: 2000/04/30 12:40PM EDT
3254    Tag: dbm-20000430-bug_fixes
3255    Description:
3256    
3257    1. Fix for bug 1498
3258       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
3259       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
3260         added toRealArray function
3261       smlnj/src/compiler/MiscUtil/print/ppobj.sml
3262         added check for tag Obj.RealArray to array printing case in ppObj
3263    
3264    2. Fix for bug 1510
3265       smlnj/src/compiler/Semant/types/typesutil.sml
3266         fixed definition of dummyargs (used by equalTycon) so that
3267         dummy args are distinct types
3268    
3269    ----------------------------------------------------------------------
3270    Name: Matthias Blume
3271    Date: 2000/04/30 01:00:00 JST
3272    Tag: blume-20000430-versions
3273    Description:
3274    
3275    1. CM version numbering added.  This is an implementation of Lal's
3276       proposal for adding version numbers and version checking to .cm
3277       files.  Lal said that his proposal was just that -- a proposal.
3278       For the time being I went ahead and implemented it so that people
3279       can comment on it.  Everything is completely backward-compatible
3280       (except for the stable library format, i.e., new bootfiles!).
3281    
3282       As usual, see the CM manual for details.
3283    
3284    2. An alternative syntax for anchored paths has been implemented.
3285       Dave has recently voiced the same concerns that I had when I did
3286       this, so there should be some support.  My take is that eventually
3287       I will let support for the current syntax (where anchors are
3288       "implicit") fade out in favor of the new, explicit syntax.
3289       In order to be backward-compatible, both old and new syntax are
3290       currently supported.
3291    
3292       Again, see the CM manual for details.
3293    
3294    3. Parallel make is trying to be slightly smarter:  When the master
3295       process finds a "bottleneck", i.e., when there is only one
3296       compilation unit that can be compiled and everybody else is
3297       waiting on it, then it will simply compile it directly instead
3298       of clumsily telling one of the slaves to do it.
3299    
3300    4. Support for "unsharing" added.  This is necessary in order to be
3301       able to have two different versions of the same library running
3302       at the same time (e.g., for trying out a new MLRISC while still
3303       having the old MLRISC linked into the current compiler, etc.)
3304       See the CM manual.
3305    
3306    5. Simple "makedepend" functionality added for generating Makefile
3307       dependency information.  (This is rather crude at the moment.
3308       Expect some changes here in the future.)
3309    
3310    6. ".fun" added as a recognized suffix for ML files. Also documented
3311       explicitly in the manual that the fallback behavior (unknown suffix
3312       -> ML file) is not an official feature!
3313    
3314    7. Small changes to the pickler for stable libraries.
3315    
3316    8. Several internal changes to CM (for cleanup/improvement).
3317    
3318    
3319    !!!! NEW BINFILES !!!!
3320    
3321    ----------------------------------------------------------------------
3322    Name: Matthias Blume
3323    Date: 2000/04/28 17:30:00 JST
3324    Tag: blume-20000428-pathconfig
3325    Description:
3326    
3327    1. I changed config/install.sh to remove duplicate entries from the
3328       lib/pathconfig file at the end.  Moreover, the final version of
3329       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
3330       in src/system/installml.
3331    
3332    2. The config/install.sh script now consistently uses relative
3333       pathnames in lib/pathconfig whenever the anchor is in the lib
3334       directory.  (So far this was true for the libraries that come
3335       pre-compiled and bundled as part of the bootfiles but not for
3336       libraries that are compiled by the script itself.)
3337    
3338    ----------------------------------------------------------------------
3339    Name: Matthias Blume
3340    Date: 2000/04/26 13:10:00 JST
3341    Tag: blume-20000426-fun_suffix
3342    Description:
3343    
3344    Added ".fun" as a recognized file name suffix (for ML code).
3345    
3346    ----------------------------------------------------------------------
3347    Name: Allen Leung
3348    Date: 2000/04/25 17:00:00 EST
3349    Tag: leunga-20000425-alpha-ra
3350    Description:
3351    
3352    1. Alpha
3353    
3354        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
3355    in C--.
3356    
3357    2. RA
3358    
3359       Added a flag "ra-dump-size" to print out the size of the flowgraph
3360       and the interference graph.
3361    
3362    ----------------------------------------------------------------------
3363    Name: Dave MacQueen
3364    Date: 2000/04/25/
3365    Tag: dbm-20000425-mlyacc_doc_examples
3366    Description:
3367      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
3368      Updated all three examples in src/ml-yacc/examples to run
3369      under 110.* using CM.make.
3370    
3371    ----------------------------------------------------------------------
3372    Name: Allen Leung
3373    Date: 2000/04/20 23:04:00 EST
3374    Tag: leunga-20000420-ssa-c---stuff
3375    Description:
3376    
3377      This update synchronizes my repository with Yale's.  Most of these
3378    changes, however, do not affect SML/NJ at all (the RA is an exception).
3379    
3380    1. Register Allocator
3381    
3382       a. An improvement in the interference graph construction:
3383          Given a copy
3384    
3385                s <- t
3386    
3387          no interference edge between s and t is added for this definition of s.
3388    
3389       b. I've added two new spill heuristic modules that Fermin and I developed
3390          (in the new library RA.cm). These are unused in SML/NJ but maybe
3391          useful for others (Moby?)
3392    
3393    2. X86
3394    
3395       a. Various fixes in the backend provided by Fermin [C--] and Lal.
3396    
3397    3. Alpha
3398    
3399       a. Added the BSR instruction and code generation that goes with it [C--]
3400       b. Other fixes too numerous to recount provided by Fermin [C--]
3401    
3402    4. Regmaps
3403    
3404       a. The regmaps are not initialized with the identity physical bindings
3405          at creation time.  This is unneeded.
3406    
3407    5. MLRISC Optimizations
3408    
3409       a. The DJ-Graph module can now compute the iterated dominance frontiers
3410          intersects with liveness incrementally in linear time! Woohoo!
3411          This is now used in my new SSA construction algorithm.
3412    
3413       b. THe branch reorganization module is now smarter about linear chains of
3414          basic blocks.
3415    
3416    
3417    ----------------------------------------------------------------------
3418    Name: Matthias Blume
3419    Date: 2000/04/12 13:52:00 JST
3420    Tag: blume_main_v110p27_1
3421    Description:
3422    
3423    Changed install.sh script to handle archive files without version number
3424    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
3425    name of the boot file archive.
3426    
3427    ----------------------------------------------------------------------
3428    Name: Dave MacQueen
3429    Date: 2000/04/09 14:00 EDT
3430    Tag: dbm-20000410-Version_110_27
3431    Description:
3432    
3433    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
3434    
3435    2. Updated src/config/version to 110.27
3436    
3437    3. New boot files!
3438    
3439    ----------------------------------------------------------------------
3440    Name: Allen Leung
3441    Date: 2000/04/09 19:09:00 EST
3442    Tag: leunga-20000409-misc
3443    Description:
3444    
3445    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
3446    
3447    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
3448    
3449    ----------------------------------------------------------------------
3450    Name: Stefan
3451    Date: 2000/04/07 10:00:00 EDT
3452    Tag: monnier-20000406-branch-handling
3453    Description:
3454    
3455    Improved handling of branches (mostly those generated from
3456    polymorphic equality), removed switchoff and changed the
3457    default optimization settings (more cpsopt and less flintopt).
3458    
3459    ----------------------------------------------------------------------
3460    Name: Allen Leung
3461    Date: 2000/04/06 01:30:00 EST
3462    Tag: leunga-20000406-peephole-x86-SSA-2
3463    Description:
3464    
3465       Forgot a few files.
3466    
3467    ----------------------------------------------------------------------
3468    Name: Allen Leung
3469    Date: 2000/04/06 00:36:00 EST
3470    Tag: leunga-20000406-peephole-x86-SSA
3471    Description:
3472    
3473    1.  New Peephole code
3474    
3475    2.  Minor improvement to X86 instruction selection
3476    
3477    3.  Various fixes to SSA and machine description -> code translator
3478    
3479    ----------------------------------------------------------------------
3480    Name: Matthias Blume
3481    Date: 2000/04/05 12:30:00 JST
3482    Tag: blume_main_v110p26p2_3
3483    Description:
3484    
3485    This update just merges three minor cosmetic updates to CM's sources
3486    to get ready for the 110.27 code freeze on Friday.  No functionality
3487    has changed.
3488    
3489    ----------------------------------------------------------------------
3490    Name: Allen Leung
3491    Date: 2000/04/04 19:39:00 EST
3492    Tag: leunga-20000404-x86-asm
3493    Description:
3494    
3495    1.  Fixed a problem in X86 assembly.
3496    
3497        Things like
3498    
3499           jmp %eax
3500           jmp (%eax)
3501    
3502        should be output as
3503    
3504           jmp *%eax
3505           jmp *(%eax)
3506    
3507    2.  Assembly output
3508    
3509          Added a new flag
3510    
3511              "asm-indent-copies" (default to false)
3512    
3513          When this flag is on, parallel copies will be indented an extra level.
3514    
3515    ----------------------------------------------------------------------
3516    Name: Allen Leung
3517    Date: 2000/04/04 03:18:00 EST
3518    Tag: leunga-20000404-C--Moby
3519    Description:
3520    
3521        All of these fixes are related to C--, Moby, and my own optimization
3522        stuff; so they shouldn't affect SML/NJ.
3523    
3524    1.  X86
3525    
3526        Various fixes related floating point, and extensions.
3527    
3528    2.  Alpha
3529    
3530        Some extra patterns related to loads with signed/zero extension
3531        provided by Fermin.
3532    
3533    3.  Assembly
3534    
3535        When generating assembly, resolve the value of client defined constants,
3536        instead of generating symbolic values.  This is controlled by the
3537        new flag "asm-resolve-constants", which is default to true.
3538    
3539    4.  Machine Descriptions
3540    
3541        a. The precedence parser was slightly broken when parsing infixr symbols.
3542        b. The type generalizing code had the bound variables reversed, resulting
3543           in a problem during arity raising.
3544        c. Various fixes in machine descriptions.
3545    
3546    ----------------------------------------------------------------------
3547    Name: Matthias Blume
3548    Date: 2000/04/03 16:05:00 JST
3549    Tag: blume_main_v110p26p2_2
3550    Description:
3551    
3552    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
3553    now done via the ordinary static environment that is context to each
3554    compilation unit.
3555    
3556    To this end, I arranged that instead of "structure Core" as "structure
3557    _Core" is bound in the pervasive environment.  Core access is done via
3558    _Core (which can never be accidentally rebound because _Core is not a
3559    legal surface-syntax symbol).
3560    
3561    The current solution is much cleaner because the core environment is
3562    now simply part of the pervasive environment which is part of every
3563    compilation unit's context anyway.  In particular, this eliminates all
3564    special-case handling that was necessary until now in order to deal
3565    with dynamic and symbolic parts of the core environment.
3566    
3567    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
3568    compilation manager's bootstrap compiler (actually: in the "init group"
3569    handling).  See the comments in src/system/smlnj/init/init.cmi for
3570    more details.
3571    
3572    I also tried to track down all mentions of "Core" (as string argument
3573    to Symbol.strSymbol) in the compiler and replaced them with a
3574    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
3575    name appears in one place only.
3576    
3577    Binfile and bootfile format have not changed, but the switchover from
3578    the old "init.cmi" to the new one is a bit tricky, so I supplied new
3579    bootfiles anyway.
3580    
3581    ----------------------------------------------------------------------
3582    Name: Allen Leung
3583    Date: 2000/04/02 21:17:00 EST
3584    Tag: leunga-20000402-mltree
3585    Description:
3586    
3587       1. Renamed the constructor CALL in MLTREE by popular demand.
3588       2. Added a bunch of files from my repository.  These are currently
3589          used by other non-SMLNJ backends.
3590    
3591    ----------------------------------------------------------------------
3592    Name: Allen Leung
3593    Date: 2000/03/31 21:15:00 EST
3594    Tag: leunga-20000331-aliasing
3595    Description:
3596    
3597    This update contains a rewritten (and hopefully more correct) module
3598    for extracting aliasing information from CPS.
3599    
3600       To turn on this feature:
3601    
3602            Compiler.Control.CG.memDisambiguate := true
3603    
3604       To pretty print the region information with assembly
3605    
3606           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
3607    
3608       To control how many levels of aliasing information are printed, use:
3609    
3610           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
3611    
3612       The default of n is 3.
3613    
3614    ----------------------------------------------------------------------
3615    Name: David MacQueen
3616    Date: 2000/03/31 11:15:00 EST
3617    Tag: dbm-20000331-runtime_fix
3618    Description:
3619    
3620    This update contains:
3621    
3622    1. runtime/c-lib/c-libraries.c
3623       includes added in revision 1.2 caused compilation errors on hppa-hpux
3624    
3625    2. fix for bug 1556
3626       system/Basis/Implementation/NJ/internal-signals.sml
3627    
3628    ----------------------------------------------------------------------
3629    Name: Matthias Blume
3630    Date: 2000/03/31 18:00:00 JST
3631    Tag: blume_main_v110p26p2_1
3632    Description:
3633    
3634    This update contains:
3635    
3636    1. A small change to CM's handling of stable libraries:
3637       CM now maintains one "global" modmap that is used for all stable
3638       libr