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 586, Thu Mar 30 05:08:07 2000 UTC revision 1152, Tue Mar 19 21:36:30 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: 2002/03/19 16:37:00 EST
18    Tag: blume-20020319-witnesses
19    Description:
20    
21    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
22    
23    This implements a generic mechanism for changing constness qualifiers
24    anywhere within big C types without resorting to outright "casts".
25    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
26    modify the constness at the outermost level.)
27    The implementation of "convert" is based on the idea of "witness"
28    values -- values that are not used by the operation but whose types
29    "testify" to their applicability.  On the implementation side, "convert"
30    is simply a projection (returning its second curried argument).  With
31    cross-module inlining, it should not result in any machine code being
32    generated.
33    
34    ----------------------------------------------------------------------
35    Name: Matthias Blume
36    Date: 2002/03/15 16:40:00 EST
37    Tag: blume-20020315-basis
38    Description:
39    
40    Provided (preliminary?) implementations for
41    
42      {String,Substring}.{concatWith,isSuffix,isSubstring}
43    
44    and
45    
46      Substring.full
47    
48    Those are in the Basis spec but they were missing in SML/NJ.
49    
50    ----------------------------------------------------------------------
51    Name: Matthias Blume
52    Date: 2002/03/14 21:30:00 EST
53    Tag: blume-20020314-controls
54    Description:
55    
56    Controls:
57    ---------
58    
59    1. Factored out the recently-added Controls : CONTROLS stuff and put
60       it into its own library $/controls-lib.cm.  The source tree for
61       this is under src/smlnj-lib/Controls.
62    
63    2. Changed the names of types and functions in this interface, so they
64       make a bit more "sense":
65    
66          module -> registry
67          'a registry -> 'a group
68    
69    3. The interface now deals in ref cells only.  The getter/setter interface
70       is (mostly) gone.
71    
72    4. Added a function that lets one register an already-existing ref cell.
73    
74    5. Made the corresponding modifications to the rest of the code so that
75       everything compiles again.
76    
77    6. Changed the implementation of Controls.MLRISC back to something closer
78       to the original.  In particular, this module (and therefore MLRISC)
79       does not depend on Controls.  There now is some link-time code in
80       int-sys.sml that registers the MLRISC controls with the Controls
81       module.
82    
83    CM:
84    ---
85    
86      * One can now specify the lambda-split aggressiveness in init.cmi.
87    
88    ----------------------------------------------------------------------
89    Name: Allen Leung
90    Date: 2002/03/13 17:30:00 EST
91    Tag: leunga-20020313-x86-fp-unary
92    Description:
93    
94    Bug fix for:
95    
96    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
97    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
98    > - fun f(x,(y,z)) = Real.~ y;
99    > [autoloading]
100    > [autoloading done]
101    >       fchsl   (%eax), 184(%esp)
102    > Error: MLRisc bug: X86MCEmitter.emitInstr
103    >
104    > uncaught exception Error
105    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
106    
107    The problem was that the code generator did not generate any fp registers
108    in this case, and the ra didn't know that it needed to run the X86FP phase to
109    translate the pseudo fp instruction.   This only happened with unary fp
110    operators in certain situations.
111    
112    ----------------------------------------------------------------------
113    Name: Matthias Blume
114    Date: 2002/03/13 14:00:00 EST
115    Tag: blume-20020313-overload-etc
116    Description:
117    
118    1. Added _overload as a synonym for overload for backward compatibility.
119       (Control.overloadKW must be true for either version to be accepted.)
120    
121    2. Fixed bug in install script that caused more things to be installed
122       than what was requested in config/targets.
123    
124    3. Made CM aware of the (_)overload construct so that autoloading
125       works.
126    
127    ----------------------------------------------------------------------
128    Name: Matthias Blume
129    Date: 2002/03/12 22:03:00 EST
130    Tag: blume-20020312-url
131    Description:
132    
133    Forgot to update BOOT and srcarchiveurl.
134    
135    ----------------------------------------------------------------------
136    Name: Matthias Blume
137    Date: 2002/03/12 17:30:00 EST
138    Tag: blume-20020312-version110392
139    Description:
140    
141    Yet another version number bump (because of small changes to the
142    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
143    
144    Changes:
145    
146      The new pid generation scheme described a few weeks ago was overly
147      complicated.  I implemented a new mechanism that is simpler and
148      provides a bit more "stability":  Once CM has seen a compilation
149      unit, it keeps its identity constant (as long as you do not delete
150      those crucial CM/GUID/* files).  This means that when you change
151      an interface, compile, then go back to the old interface, and
152      compile again, you arrive at the original pid.
153    
154      There now also is a mechanism that instructs CM to use the plain
155      environment hash as a module's pid (effectively making its GUID
156      the empty string).  For this, "noguid" must be specified as an
157      option to the .sml file in question within its .cm file.
158      This is most useful for code that is being generated by tools such
159      as ml-nlffigen (because during development programmers tend to
160      erase the tool's entire output directory tree including CM's cached
161      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
162      revert to the old, broken behavior of SML/NJ, but in specific cases
163      where there is no danger of interface confusion, its use is ok
164      (I think).
165    
166      ml-nlffigen by default generates "noguid" annotations.  They can be
167      turned off by specifying -guid in its command line.
168    
169    ----------------------------------------------------------------------
170    Name: Lal George
171    Date: 2002/03/12 12 14:42:36 EST
172    Tag: george-20020312-frequency-computation
173    Description:
174    
175    Integrated jump chaining and static block frequency into the
176    compiler. More details and numbers later.
177    
178    ----------------------------------------------------------------------
179    Name: Lal George
180    Date: 2002/03/11 11 22:38:53 EST
181    Tag: george-20020311-jump-chain-elim
182    Description:
183    
184    Tested the jump chain elimination on all architectures (except the
185    hppa).  This is on by default right now and is profitable for the
186    alpha and x86, however, it may not be profitable for the sparc and ppc
187    when compiling the compiler.
188    
189    The gc test will typically jump to a label at the end of the cluster,
190    where there is another jump to an external cluster containing the actual
191    code to invoke gc. This is to allow factoring of common gc invocation
192    sequences. That is to say, we generate:
193    
194            f:
195               testgc
196               ja   L1      % jump if above to L1
197    
198            L1:
199               jmp L2
200    
201    
202    After jump chain elimination the 'ja L1' instructions is converted to
203    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
204    up being implemented in their long form (if L2 is far away) using:
205    
206            jbe     L3      % jump if below or equal to L3
207            jmp     L2
208         L3:
209            ...
210    
211    
212    For large compilation units L2  may be far away.
213    
214    
215    ----------------------------------------------------------------------
216    Name: Matthias Blume
217    Date: 2002/03/11 13:30:00 EST
218    Tag: blume-20020311-mltreeeval
219    Description:
220    
221    A functor parameter was missing.
222    
223    ----------------------------------------------------------------------
224    Name: Allen Leung
225    Date: 2002/03/11 10:30:00 EST
226    Tag: leunga-20020311-runtime-string0
227    Description:
228    
229       The representation of the empty string now points to a
230    legal null terminated C string instead of unit.  It is now possible
231    to convert an ML string into C string with InlineT.CharVector.getData.
232    This compiles into one single machine instruction.
233    
234    ----------------------------------------------------------------------
235    Name: Allen Leung
236    Date: 2002/03/10 23:55:00 EST
237    Tag: leunga-20020310-x86-call
238    Description:
239    
240       Added machine generation for CALL instruction (relative displacement mode)
241    
242    ----------------------------------------------------------------------
243    Name: Matthias Blume
244    Date: 2002/03/08 16:05:00
245    Tag: blume-20020308-entrypoints
246    Description:
247    
248    Version number bumped to 110.39.1.  NEW BOOTFILES!
249    
250    Entrypoints: non-zero offset into a code object where execution should begin.
251    
252    - Added the notion of an entrypoint to CodeObj.
253    - Added reading/writing of entrypoint info to Binfile.
254    - Made runtime system bootloader aware of entrypoints.
255    - Use the address of the label of the first function given to mlriscGen
256      as the entrypoint.  This address is currently always 0, but it will
257      not be 0 once we turn on block placement.
258    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
259      with entry points) from mlriscGen.
260    
261    ----------------------------------------------------------------------
262    Name: Allen Leung
263    Date: 2002/03/07 20:45:00 EST
264    Tag: leunga-20020307-x86-cmov
265    Description:
266    
267       Bug fixes for CMOVcc on x86.
268    
269       1. Added machine code generation for CMOVcc
270       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
271       3. CMOVcc cannot have an immediate operand as argument.
272    
273    ----------------------------------------------------------------------
274    Name: Matthias Blume
275    Date: 2002/03/07 16:15:00 EST
276    Tag: blume-20020307-controls
277    Description:
278    
279    This is a very large but mostly boring patch which makes (almost)
280    every tuneable compiler knob (i.e., pretty much everything under
281    Control.* plus a few other things) configurable via both the command
282    line and environment variables in the style CM did its configuration
283    until now.
284    
285    Try starting sml with '-h' (or, if you are brave, '-H')
286    
287    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
288    implements the underlying generic mechanism.
289    
290    The interface to some of the existing such facilities has changed somewhat.
291    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
292    (The getFoo interface is still there for backward-compatibility, but its
293    use is deprecated.)
294    
295    The ml-build script passes -Cxxx=yyy command-line arguments through so
296    that one can now twiddle the compiler settings when using this "batch"
297    compiler.
298    
299    TODO items:
300    
301    We should go through and throw out all controls that are no longer
302    connected to anything.  Moreover, we should go through and provide
303    meaningful (and correct!) documentation strings for those controls
304    that still are connected.
305    
306    Currently, multiple calls to Controls.new are accepted (only the first
307    has any effect).  Eventually we should make sure that every control
308    is being made (via Controls.new) exactly once.  Future access can then
309    be done using Controls.acc.
310    
311    Finally, it would probably be a good idea to use the getter-setter
312    interface to controls rather than ref cells.  For the time being, both
313    styles are provided by the Controls module, but getter-setter pairs are
314    better if thread-safety is of any concern because they can be wrapped.
315    
316    *****************************************
317    
318    One bug fix: The function blockPlacement in three of the MLRISC
319    backpatch files used to be hard-wired to one of two possibilities at
320    link time (according to the value of the placementFlag).  But (I
321    think) it should rather sense the flag every time.
322    
323    *****************************************
324    
325    Other assorted changes (by other people who did not supply a HISTORY entry):
326    
327    1. the cross-module inliner now works much better (Monnier)
328    2. representation of weights, frequencies, and probabilities in MLRISC
329       changed in preparation of using those for weighted block placement
330       (Reppy, George)
331    
332    ----------------------------------------------------------------------
333    Name: Lal George
334    Date: 2002/03/07 14:44:24 EST 2002
335    Tag: george-20020307-weighted-block-placement
336    
337    Tested the weighted block placement optimization on all architectures
338    (except the hppa) using AMPL to generate the block and edge frequencies.
339    Changes were required in the machine properties to correctly
340    categorize trap instructions. There is an MLRISC flag
341    "weighted-block-placement" that can be used to enable weighted block
342    placement, but this will be ineffective without block/edge
343    frequencies (coming soon).
344    
345    
346    ----------------------------------------------------------------------
347    Name: Lal George
348    Date: 2002/03/05 17:24:48 EST
349    Tag: george-20020305-linkage-cluster
350    
351    In order to support the block placement optimization, a new cluster
352    is generated as the very first cluster (called the linkage cluster).
353    It contains a single jump to the 'real' entry point for the compilation
354    unit. Block placement has no effect on the linkage cluster itself, but
355    all the other clusters  have full freedom in the manner in which they
356    reorder blocks or functions.
357    
358    On the x86 the typical linkage code that is generated is:
359       ----------------------
360            .align 2
361       L0:
362            addl    $L1-L0, 72(%esp)
363            jmp     L1
364    
365    
366            .align  2
367       L1:
368       ----------------------
369    
370    72(%esp) is the memory location for the stdlink register. This
371    must contain the address of the CPS function being called. In the
372    above example, it contains the address of  L0; before
373    calling L1 (the real entry point for the compilation unit), it
374    must contain the address for L1, and hence
375    
376            addl $L1-L0, 72(%esp)
377    
378    I have tested this on all architectures except the hppa.The increase
379    in code size is of course negligible
380    
381    ----------------------------------------------------------------------
382    Name: Allen Leung
383    Date: 2002/03/03 13:20:00 EST
384    Tag: leunga-20020303-mlrisc-tools
385    
386      Added #[ ... ] expressions to mlrisc tools
387    
388    ----------------------------------------------------------------------
389    Name: Matthias Blume
390    Date: 2002/02/27 12:29:00 EST
391    Tag: blume-20020227-cdebug
392    Description:
393    
394    - made types in structure C and C_Debug to be equal
395    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
396    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
397    
398    ----------------------------------------------------------------------
399    Name: Matthias Blume
400    Date: 2002/02/26 12:00:00 EST
401    Tag: blume-20020226-ffi
402    Description:
403    
404    1. Fixed a minor bug in CM's "noweb" tool:
405       If numbering is turned off, then truly don't number (i.e., do not
406       supply the -L option to noweb).  The previous behavior was to supply
407       -L'' -- which caused noweb to use the "default" line numbering scheme.
408       Thanks to Chris Richards for pointing this out (and supplying the fix).
409    
410    2. Once again, I reworked some aspects of the FFI:
411    
412       A. The incomplete/complete type business:
413    
414       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
415         gone!
416       - ML types representing an incomplete type are now *equal* to
417         ML types representing their corresponding complete types (just like
418         in C).  This is still safe because ml-nlffigen will not generate
419         RTTI for incomplete types, nor will it generate functions that
420         require access to such RTTI.   But when ML code generated from both
421         incomplete and complete versions of the C type meet, the ML types
422         are trivially interoperable.
423    
424         NOTE:  These changes restore the full generality of the translation
425         (which was previously lost when I eliminated functorization)!
426    
427       B. Enum types:
428    
429       - Structure C now has a type constructor "enum" that is similar to
430         how the "su" constructor works.  However, "enum" is not a phantom
431         type because each "T enum" has values (and is isomorphic to
432         MLRep.Signed.int).
433       - There are generic access operations for enum objects (using
434         MLRep.Signed.int).
435       - ml-nlffigen will generate a structure E_foo for each "enum foo".
436         * The structure contains the definition of type "mlrep" (the ML-side
437         representation type of the enum).  Normally, mlrep is the same
438         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
439         then mlrep will be defined as a datatype -- thus facilitating
440         pattern matching on mlrep values.
441         ("-ec" will be suppressed if there are duplicate values in an
442          enumeration.)
443         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
444         will be generated for each C enum constant xxx.
445         * Conversion functions m2i and i2m convert between mlrep and
446         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
447         * Coversion functions c and ml convert between mlrep and "tag enum".
448         * Access functions (get/set) fetch and store mlrep values.
449       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
450         enumerations are merged into one single enumeration represented by
451         structure E_'.
452    
453    ----------------------------------------------------------------------
454    Name: Allen Leung
455    Date: 2002/02/25 04:45:00 EST
456    Tag: leunga-20020225-cps-spill
457    
458    This is a new implementation of the CPS spill phase.
459    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
460    In case of problems, replace it with the old file spill.sml
461    
462    The current compiler runs into some serious performance problems when
463    constructing a large record.  This can happen when we try to compile a
464    structure with many items.  Even a very simple structure like the following
465    makes the compiler slow down.
466    
467        structure Foo = struct
468           val x_1 = 0w1 : Word32.int
469           val x_2 = 0w2 : Word32.int
470           val x_3 = 0w3 : Word32.int
471           ...
472           val x_N = 0wN : Word32.int
473        end
474    
475    The following table shows the compile time, from N=1000 to N=4000,
476    with the old compiler:
477    
478    N
479    1000   CPS 100 spill                           0.04u  0.00s  0.00g
480           MLRISC ra                               0.06u  0.00s  0.05g
481              (spills = 0 reloads = 0)
482           TOTAL                                   0.63u  0.07s  0.21g
483    
484    1100   CPS 100 spill                           8.25u  0.32s  0.64g
485           MLRISC ra                               5.68u  0.59s  3.93g
486              (spills = 0 reloads = 0)
487           TOTAL                                   14.71u  0.99s  4.81g
488    
489    1500   CPS 100 spill                           58.55u  2.34s  1.74g
490           MLRISC ra                               5.54u  0.65s  3.91g
491              (spills = 543 reloads = 1082)
492           TOTAL                                   65.40u  3.13s  6.00g
493    
494    2000   CPS 100 spill                           126.69u  4.84s  3.08g
495           MLRISC ra                               0.80u  0.10s  0.55g
496              (spills = 42 reloads = 84)
497           TOTAL                                   129.42u  5.10s  4.13g
498    
499    3000   CPS 100 spill                           675.59u  19.03s  11.64g
500           MLRISC ra                               2.69u  0.27s  1.38g
501              (spills = 62 reloads = 124)
502           TOTAL                                   682.48u  19.61s  13.99g
503    
504    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
505           MLRISC ra                               4.96u  0.27s  2.72g
506              (spills = 85 reloads = 170)
507           TOTAL                                   2375.26u  57.21s  48.00g
508    
509    As you can see the old cps spill module suffers from some serious
510    performance problem.  But since I cannot decipher the old code fully,
511    instead of patching the problems up, I'm reimplementing it
512    with a different algorithm.  The new code is more modular,
513    smaller when compiled, and substantially faster
514    (O(n log n) time and O(n) space).  Timing of the new spill module:
515    
516    4000  CPS 100 spill                           0.02u  0.00s  0.00g
517          MLRISC ra                               0.25u  0.02s  0.15g
518             (spills=1 reloads=3)
519          TOTAL                                   7.74u  0.34s  1.62g
520    
521    Implementation details:
522    
523    As far as I can tell, the purpose of the CPS spill module is to make sure the
524    number of live variables at any program point (the bandwidth)
525    does not exceed a certain limit, which is determined by the
526    size of the spill area.
527    
528    When the bandwidth is too large, we decrease the register pressure by
529    packing live variables into spill records.  How we achieve this is
530    completely different than what we did in the old code.
531    
532    First, there is something about the MLRiscGen code generator
533    that we should be aware of:
534    
535    o MLRiscGen performs code motion!
536    
537       In particular, it will move floating point computations and
538       address computations involving only the heap pointer to
539       their use sites (if there is only a single use).
540       What this means is that if we have a CPS record construction
541       statement
542    
543           RECORD(k,vl,w,e)
544    
545       we should never count the new record address w as live if w
546       has only one use (which is often the case).
547    
548       We should do something similar to floating point, but the transformation
549       there is much more complex, so I won't deal with that.
550    
551    Secondly, there are now two new cps primops at our disposal:
552    
553     1. rawrecord of record_kind option
554        This pure operator allocates some uninitialized storage from the heap.
555        There are two forms:
556    
557         rawrecord NONE [INT n]  allocates a tagless record of length n
558         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
559                                     and initializes the tag.
560    
561     2. rawupdate of cty
562          rawupdate cty (v,i,x)
563          Assigns to x to the ith component of record v.
564          The storelist is not updated.
565    
566    We use these new primops for both spilling and increment record construction.
567    
568     1. Spilling.
569    
570        This is implemented with a linear scan algorithm (but generalized
571        to trees).  The algorithm will create a single spill record at the
572        beginning of the cps function and use rawupdate to spill to it,
573        and SELECT or SELp to reload from it.  So both spills and reloads
574        are fine-grain operations.  In contrast, in the old algorithm
575        "spills" have to be bundled together in records.
576    
577        Ideally, we should sink the spill record construction to where
578        it is needed.  We can even split the spill record into multiple ones
579        at the places where they are needed.  But CPS is not a good
580        representation for global code motion, so I'll keep it simple and
581        am not attempting this.
582    
583     2. Incremental record construction (aka record splitting).
584    
585        Long records with many component values which are simulatenously live
586        (recall that single use record addresses are not considered to
587         be live) are constructed with rawrecord and rawupdate.
588        We allocate space on the heap with rawrecord first, then gradually
589        fill it in with rawupdate.  This is the technique suggested to me
590        by Matthias.
591    
592        Some restrictions on when this is applicable:
593        1. It is not a VECTOR record.  The code generator currently does not handle
594           this case. VECTOR record uses double indirection like arrays.
595        2. All the record component values are defined in the same "basic block"
596           as the record constructor.  This is to prevent speculative
597           record construction.
598    
599    ----------------------------------------------------------------------
600    Name: Allen Leung
601    Date: 2002/02/22 01:02:00 EST
602    Tag: leunga-20020222-mlrisc-tools
603    
604    Minor bug fixes in the parser and rewriter
605    
606    ----------------------------------------------------------------------
607    Name: Allen Leung
608    Date: 2002/02/21 20:20:00 EST
609    Tag: leunga-20020221-peephole
610    
611    Regenerated the peephole files.  Some contained typos in the specification
612    and some didn't compile because of pretty printing bugs in the old version
613    of 'nowhere'.
614    
615    ----------------------------------------------------------------------
616    Name: Allen Leung
617    Date: 2002/02/19 20:20:00 EST
618    Tag: leunga-20020219-mlrisc-tools
619    Description:
620    
621       Minor bug fixes to the mlrisc-tools library:
622    
623       1.  Fixed up parsing colon suffixed keywords
624       2.  Added the ability to shut the error messages up
625       3.  Reimplemented the pretty printer and fixed up/improved
626           the pretty printing of handle and -> types.
627       4.  Fixed up generation of literal symbols in the nowhere tool.
628       5.  Added some SML keywords to to sml.sty
629    
630    ----------------------------------------------------------------------
631    Name: Matthias Blume
632    Date: 2002/02/19 16:20:00 EST
633    Tag: blume-20020219-cmffi
634    Description:
635    
636    A wild mix of changes, some minor, some major:
637    
638    * All C FFI-related libraries are now anchored under $c:
639        $/c.cm      --> $c/c.cm
640        $/c-int.cm  --> $c/internals/c-int.cm
641        $/memory.cm --> $c/memory/memory.cm
642    
643    * "make" tool (in CM) now treats its argument pathname slightly
644      differently:
645        1. If the native expansion is an absolute name, then before invoking
646           the "make" command on it, CM will apply OS.Path.mkRelative
647           (with relativeTo = OS.FileSys.getDir()) to it.
648        2. The argument will be passed through to subsequent phases of CM
649           processing without "going native".  In particular, if the argument
650           was an anchored path, then "make" will not lose track of that anchor.
651    
652    * Compiler backends now "know" their respective C calling conventions
653      instead of having to be told about it by ml-nlffigen.  This relieves
654      ml-nlffigen from one of its burdens.
655    
656    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
657    
658    * Export C_DEBUG and C_Debug from $c/c.cm.
659    
660    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
661      subtyping relationship between incomplete pointers and their complete
662      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
663      with the convention of instantiating 'o with ('t, 'c) obj whenever
664      the pointer target type is complete.  In the incomplete case, 'o
665      will be instantiated with some "'c iobj" -- a type obtained by
666      using one of the functors PointerToIncompleteType or PointerToCompleteType.
667    
668      Operations that work on both incomplete and complete pointer types are
669      typed as taking an 'o ptr while operations that require the target to
670      be known are typed as taking some ('t, 'c) obj ptr.
671    
672      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
673      where void is an eqtype without any values.  This makes it possible
674      to work on voidptr values using functions meant to operate on light
675      incomplete pointers.
676    
677    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
678      been vastly simplified.
679    
680    ----------------------------------------------------------------------
681    Name: Matthias Blume
682    Date: 2002/02/19 10:48:00 EST
683    Tag: blume-20020219-pqfix
684    Description:
685    
686    Applied Chris Okasaki's bug fix for priority queues.
687    
688    ----------------------------------------------------------------------
689    Name: Matthias Blume
690    Date: 2002/02/15 17:05:00
691    Tag: Release_110_39
692    Description:
693    
694    Last-minute retagging is becoming a tradition... :-(
695    
696    This is the working release 110.39.
697    
698    ----------------------------------------------------------------------
699    Name: Matthias Blume
700    Date: 2002/02/15 16:00:00 EST
701    Tag: Release_110_39-orig
702    Description:
703    
704    Working release 110.39.  New bootfiles.
705    
706    (Update: There was a small bug in the installer so it wouldn't work
707    with all shells.  So I retagged. -Matthias)
708    
709    ----------------------------------------------------------------------
710    Name: Matthias Blume
711    Date: 2002/02/15 14:17:00 EST
712    Tag: blume-20020215-showbindings
713    Description:
714    
715    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
716    the latter can be useful for exploring what bindings are available at
717    the interactive prompt.  (The first function returns only the list
718    of symbols that are really bound, the second prints those but also the
719    ones that CM's autoloading mechanism knows about.)
720    
721    ----------------------------------------------------------------------
722    Name: Matthias Blume
723    Date: 2002/02/15 12:08:00 EST
724    Tag: blume-20020215-iptrs
725    Description:
726    
727    Two improvements to ml-nlffigen:
728    
729      1. Write files only if they do not exist or if their current contents
730         do not coincide with what's being written.  (That is, avoid messing
731         with the time stamps unless absolutely necessary.)
732    
733      2. Implement a "repository" mechanism for generated files related
734         to "incomplete pointer types".   See the README file for details.
735    
736    ----------------------------------------------------------------------
737    Name: Matthias Blume
738    Date: 2002/02/14 11:50:00 EST
739    Tag: blume-20020214-quote
740    Description:
741    
742    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
743    because of the new and improved tag generation scheme.  (Thanks to Allen
744    Leung for pointing it out.)
745    
746    ----------------------------------------------------------------------
747    Name: Lal George
748    Date: 2002/02/14 09:55:27 EST 2002
749    Tag: george-20020214-isabelle-bug
750    Description:
751    
752    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
753    of Isabelle on the x86.
754    
755    From Allen:
756    -----------
757     I've found the problem:
758    
759         in ra-core.sml, I use the counter "blocked" to keep track of the
760         true number of elements in the freeze queue.  When the counter goes
761         to zero, I skip examining the queue.  But I've messed up the
762         bookkeeping in combine():
763    
764             else ();
765             case !ucol of
766               PSEUDO => (if !cntv > 0 then
767                     (if !cntu > 0 then blocked := !blocked - 1 else ();
768                                        ^^^^^^^^^^^^^^^^^^^^^^^
769                      moveu := mergeMoveList(!movev, !moveu)
770                     )
771                  else ();
772    
773         combine() is called to coalesce two nodes u and v.
774         I think I was thinking that if the move counts of u and v are both
775         greater than zero then after they are coalesced then one node is
776         removed from the freeze queue.  Apparently I was thinking that
777         both u and v are of low degree, but that's clearly not necessarily true.
778    
779    
780    02/12/2002:
781        Here's the patch.  HOL now compiles.
782    
783        I don't know how this impact on performance (compile
784        time or runtime).  This bug caused the RA (especially on the x86)
785        to go thru the potential spill phase when there are still nodes on the
786        freeze queue.
787    
788    
789    
790    
791    ----------------------------------------------------------------------
792    Name: Matthias Blume
793    Date: 2002/02/13 22:40:00 EST
794    Tag: blume-20020213-fptr-rtti
795    Description:
796    
797    Fixed a bug in ml-nlffigen that was introduced with one of the previous
798    updates.
799    
800    ----------------------------------------------------------------------
801    Name: Matthias Blume
802    Date: 2002/02/13 16:41:00 EST
803    Tag: blume-20020213-cmlpq
804    Description:
805    
806    Added new priority queue export symbols (which have just been added to
807    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
808    not compile and the installer would choke.)
809    
810    ----------------------------------------------------------------------
811    Name: Matthias Blume
812    Date: 2002/02/13 16:15:00 EST
813    Tag: blume-20020213-various
814    Description:
815    
816    1. More tweaks to ml-nlffigen:
817    
818       - better internal datastructures (resulting in slight speedup)
819       - "-match" option requires exact match
820       - "localized" gensym counters (untagged structs/unions nested within
821         other structs/unions or within typedefs get a fresh counter; their
822         tag will be prefixed by a concatenation of their parents' tags)
823       - bug fixes (related to calculation of transitive closure of types
824         to be included in the output)
825    
826    2. Minor Basis updates:
827    
828       - added implementations for List.collate and Option.app
829    
830    ----------------------------------------------------------------------
831    Name: Matthias Blume
832    Date: 2002/02/11 15:55:00 EST
833    Tag: blume-20020211-gensym
834    Description:
835    
836    Added a "-gensym" option to command line of ml-nlffigen.  This can be
837    used to specify a "stem" -- a string that is inserted in all "gensym'd"
838    names (ML structure names that correspond to unnamed C structs, unions,
839    and enums), so that separate runs of ml-nlffigen do not clash.
840    
841    ----------------------------------------------------------------------
842    Name: Matthias Blume
843    Date: 2002/02/11 12:05:00 EST
844    Tag: blume-20020211-gensml
845    Description:
846    
847    A quick fix for a problem with GenSML (in the pgraph-util library):
848    Make generation of toplevel "local" optional.  (Strictly speaking,
849    signature definitions within "local" are not legal SML.)
850    
851    Other than that: updates to INSTALL and cm/TODO.
852    
853    ----------------------------------------------------------------------
854    Name: Matthias Blume
855    Date: 2002/02/08 15:00:00 EST
856    Tag: blume-20020208-uniquepid
857    Description:
858    
859    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
860    
861    1. The installer (config/install.sh) has gotten smarter:
862    
863         - Configuration options are a bit easier to specify now
864           (in config/targets).
865         - Bug in recognizing .tar.bz2 files fixed.
866         - Installer automatically resolves dependencies between
867           configuration options (e.g., if you ask for eXene, you will
868           also get cml -- regardless whether you asked for it or not).
869         - Installer can run in "quieter mode" by setting the environment
870           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
871           "completely silent", though.
872         - Build HashCons library as part of smlnj-lib.
873    
874    2. A new scheme for assigning persistent identifiers to compilation
875       units (and, by extension, to types etc.) has been put into place.
876       This fixes a long-standing bug where types and even dynamic values
877       can get internally confused, thereby compromising type safety
878       (abstraction) and dynamic correctness.  See
879    
880         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
881    
882       for an example of how things could go wrong until now.
883    
884       The downside of the new scheme is that pids are not quite as
885       persistent as they used to be: CM will generate a fresh pid
886       for every compilation unit that it thinks it sees for the first
887       time.  That means that if you compile starting from a clean, fresh
888       source tree at two different times, you end up with different
889       binaries.
890    
891       Cutoff recompilation, however, has not been compromised because
892       CM keeps pid information in special caches between runs.
893    
894    ----------------------------------------------------------------------
895    Name: Lal George
896    Date: 2002/02/07 15:34:13 EST 2002
897    Tag: <none>
898    Description:
899    
900    Compilers that generate assembly code may produce  global labels
901    whose value is resolved at link time. The various peephole optimization
902    modules did not take this in account.
903    
904    TODO. The Labels.addrOf function should really return an option
905    type so that clients are forced to deal with this issue, rather
906    than an exception being raised.
907    
908    ----------------------------------------------------------------------
909    Name: Lal George
910    Date: 2002/02/06 13:55:02 EST
911    Tag: george-20020206-ra-breakup
912    Description:
913    
914    1. A bug fix from Allen.
915    
916        A typo causes extra fstp %st(0)'s to be generated at compensation
917        edges, which might cause stack underflow traps at runtime.  This
918        occurs in fft where there are extraneous fstps right before the 'into'
919        trap instruction (in this case they are harmless since none of the
920        integers overflow.)
921    
922    2. Pulled out various utility modules that were embedded in the modules
923       of the register allocator. I need these modules for other purposes, but
924       they are not complete enough to put into a library (just yet).
925    ----------------------------------------------------------------------
926    Name: Matthias Blume
927    Date: 2002/01/31 16:05:00 EST
928    Tag: blume-20020131-sparc-ccalls
929    Description:
930    
931    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
932       of extra stack space by mistake.  Fixed.
933    
934    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
935    
936    ----------------------------------------------------------------------
937    Name: Allen Leung
938    Date: 2002/01/30
939    Tag: leunga-20020130-nowhere-bug-fix
940    Description:
941    
942       MLRISC bug fixes:
943       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
944       2. MachineInt.fromString was negating its value.
945    
946    ----------------------------------------------------------------------
947    Name: Matthias Blume
948    Date: 2002/01/29
949    Tag: blume-20020129-INSTALL
950    Description:
951    
952    - Added somewhat detailed installation instructions (file INSTALL).
953    - Fixed curl-detection bug in config/install.sh.
954    - It is now possible to select the URL getter using the URLGETTER
955      environment variable:
956    
957          not set / "unknown"      --> automatic detection (script tries wget,
958                                       curl, and lynx)
959          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
960                                       how to properly invoke them)
961          other                    --> use $URLGETTER directly, it must take
962                                       precisely two command-line arguments
963                                       (source URL and destination file name)
964    
965    ----------------------------------------------------------------------
966    Name: Matthias Blume
967    Date: 2002/01/28
968    Tag: blume-20020128-sparc-ccalls
969    Description:
970    
971    - Fixed problem with calculation of "used" registers in sparc-c-calls.
972    - Make use of the allocParam argument in sparc-c-calls.
973    
974    ----------------------------------------------------------------------
975    Name: Matthias Blume
976    Date: 2002/01/28
977    Tag: blume-20020128-allocParam
978    Description:
979    
980    John Reppy:  Changes c-calls API to accept client-callback for
981    allocating extra stack space.
982    me: Corresponding changes to mlriscGen (using a dummy argument that
983        does not change the current behavior).
984    
985    ----------------------------------------------------------------------
986    Name: Matthias Blume
987    Date: 2002/01/28 12:00:00
988    Tag: Release_110_38
989    Description:
990    
991    This time for real!!!
992    
993    ----------------------------------------------------------------------
994    Name: Matthias Blume
995    Date: 2002/01/28 10:56:00 EST
996    Tag: blume-20020128-retraction
997    Description:
998    
999    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
1000       with blume-Release_110_38-retracted.)
1001    
1002    1. Fixed a problem with incorrect rounding modes in real64.sml.
1003       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
1004    
1005    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
1006       and enums fixed.  The naming of corresponding ML identifiers should
1007       now be consistent again.
1008    
1009    ----------------------------------------------------------------------
1010    Name: Allen Leung
1011    Date: 2002/01/27
1012    Tag: leunga-20020127-nowhere
1013    Description:
1014    
1015       Added a target called nowhere in the configuration scripts.
1016       Enabling this will build the MLRISC 'nowhere' tool (for translating
1017       programs with where-clauses into legal SML code) during installation.
1018    
1019    ----------------------------------------------------------------------
1020    Name: Matthias Blume
1021    Date: 2002/01/25 21:27:00 EST
1022    Tag: blume-Release_110_38-retracted
1023    Description:
1024    
1025    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
1026    
1027    README will be added later.
1028    
1029    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
1030    (Release_110_38) removed.  Reason: Last-minute bug fixes.
1031    
1032    ----------------------------------------------------------------------
1033    Name: Matthias Blume
1034    Date: 2002/01/25
1035    Tag: blume-20020125-ffi
1036    Description:
1037    
1038    A large number of tweaks and improvements to ml-nlffi-lib and
1039    ml-nlffigen:
1040    
1041       - ML represenation types have been streamlined
1042       - getter and setter functions work with concrete values, not abstract
1043         ones where possible
1044       - ml-nlffigen command line more flexible (see README file there)
1045       - some bugs have been fixed (hopefully)
1046    
1047    ----------------------------------------------------------------------
1048    Name: Lal George
1049    Date: 2002/01/24
1050    Tag: george-20020124-risc-ra-interface
1051    Description:
1052    
1053       There is a dramatic simplification in the interface to the
1054       register allocator for RISC architectures as a result of making
1055       parallel copy instructions explicit.
1056    
1057    ----------------------------------------------------------------------
1058    Name: Matthias Blume
1059    Date: 2002/01/22
1060    Tag: blume-20020122-x86-ccalls
1061    Description:
1062    
1063    Bug fix for c-calls on x86 (having to do with how char- and
1064    short-arguments are being handled).
1065    
1066    ----------------------------------------------------------------------
1067    Name: Matthias Blume
1068    Date: 2002/01/21
1069    Tag: blume-20020121-ff
1070    Description:
1071    
1072    Another day of fiddling with the FFI...
1073    
1074    1. Bug fix/workaround:  CKIT does not complain about negative array
1075       dimensions, so ml-nlffigen has to guard itself against this possibility.
1076       (Otherwise a negative dimension would send it into an infinite loop.)
1077    
1078    2. Some of the abstract types (light objects, light pointers, most "base"
1079       types) in structure C are now eqtypes.
1080    
1081    3. Added constructors and test functions for NULL function pointers.
1082    
1083    ----------------------------------------------------------------------
1084    Name: Matthias Blume
1085    Date: 2002/01/18
1086    Tag: blume-20020118-ready-for-new-release
1087    Description:
1088    
1089    Made config/srcarchiveurl point to a new place.  (Will provide boot
1090    files shortly.)
1091    
1092    Maybe we christen this to be 110.38?
1093    
1094    ----------------------------------------------------------------------
1095    Name: Matthias Blume
1096    Date: 2002/01/18
1097    Tag: blume-20020118-more-ffifiddle
1098    Description:
1099    
1100    Today's FFI fiddling:
1101    
1102      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
1103        C.Set.* that use concrete (MLRep.*) arguments and results instead
1104        of abstract ones.
1105    
1106      - Provided word-style bit operations etc. for "int" representation
1107        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
1108        and Long.
1109    
1110    ----------------------------------------------------------------------
1111    Name: Matthias Blume
1112    Date: 2002/01/18
1113    Tag: blume-20020118-use-x86-fp
1114    Description:
1115    
1116    Now that x86-fast-fp seems to be working, I turned it back on again
1117    by default.  (Seems to work fine now, even with the FFI.)
1118    
1119    Other than that, I added some documentation about the FFI to
1120    src/ml-nlffigen/README and updated the FFI test examples in
1121    src/ml-nlffi-lib/Tests/*.
1122    
1123    ----------------------------------------------------------------------
1124    Name: Allen Leung
1125    Date: 2002/01/17
1126    Tag: leunga-20020117-x86-fast-fp-call
1127    Description:
1128    
1129       1. Fixed a problem with handling return fp values when x86's fast fp
1130          mode is turned on.
1131    
1132       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
1133          of %f32.
1134    
1135       3. Added a constructor INT32lit to the ast of MLRISC tools.
1136    
1137    ----------------------------------------------------------------------
1138    Name: Matthias Blume
1139    Date: 2002/01/16
1140    Tag: blume-20020116-ffifiddle
1141    Description:
1142    
1143    More fiddling with the FFI interface:
1144    
1145     - Make constness 'c instead of rw wherever possible.  This eliminates
1146       the need for certain explicit coercions.  (However, due to ML's
1147       value polymorphism, there will still be many cases where explicit
1148       coercions are necessary.  Phantom types are not the whole answer
1149       to modeling a subtyping relationship in ML.)
1150    
1151     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
1152    
1153     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
1154       clumsy workaround for varargs problem.)
1155    
1156    ----------------------------------------------------------------------
1157    Name: Lal George
1158    Date: 2002/01/15
1159    Tag: <none>
1160    Description:
1161    
1162    1. Since COPY instructions are no longer native to the architecture,
1163       a generic functor can be used to implement the expandCopies function.
1164    
1165    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
1166       TEXT segment.
1167    
1168    ----------------------------------------------------------------------
1169    Name: Matthias Blume
1170    Date: 2002/01/15
1171    Tag: blume-20020115-ffiupdates
1172    Description:
1173    
1174    1. Fix for bug resulting in single-precision float values being returned
1175       incorrectly from FFI calls.
1176    
1177    2. Small modifications to C FFI API:
1178    
1179        - memory-allocation routines return straight objects (no options)
1180          and raise an exception in out-of-memory situations
1181        - unsafe extensions to cast between function pointers and pointers
1182          from/to ints
1183        - added structure C_Debug as an alternative to structure C where
1184          pointer-dereferencing (|*| and |*!) always check for null-pointers
1185        - added open_lib' to DynLinkage;  open_lib' works like open_lib
1186          but also takes a (possibly empty) list of existing library handles
1187          that the current library depends on
1188    
1189    ----------------------------------------------------------------------
1190    Name: Matthias Blume
1191    Date: 2002/01/10
1192    Tag: blume-20020110-newffigen
1193    Description:
1194    
1195    1. Updates to portable graph code.
1196    
1197    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
1198       more scalable now so that even huge interfaces such as the one
1199       for GTK compile in finite time and space. :-)
1200       See src/ml-nlffigen/README for details on what's new.
1201    
1202    ----------------------------------------------------------------------
1203    Name: Lal George
1204    Date: 2001/01/09 14:31:35 EST 2002
1205    Tag: george-20011206-rm-native-copy
1206    Description:
1207    
1208            Removed the native COPY and FCOPY instructions
1209            from all the architectures and replaced it with the
1210            explicit COPY instruction from the previous commit.
1211    
1212            It is now possible to simplify many of the optimizations
1213            modules that manipulate copies. This has not been
1214            done in this change.
1215    
1216    ----------------------------------------------------------------------
1217    Name: Lal George
1218    Date: 2001/12/06 16:50:13 EST 2001
1219    Tag: george-20011206-mlrisc-instruction
1220    Description:
1221    
1222    Changed the representation of instructions from being fully abstract
1223    to being partially concrete. That is to say:
1224    
1225      from
1226            type instruction
1227    
1228      to
1229            type instr                              (* machine instruction *)
1230    
1231            datatype instruction =
1232                LIVE of {regs: C.cellset, spilled: C.cellset}
1233              | KILL of {regs: C.cellset, spilled: C.cellset}
1234              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
1235              | ANNOTATION of {i: instruction, a: Annotations.annotation}
1236              | INSTR of instr
1237    
1238    This makes the handling of certain special instructions that appear on
1239    all architectures easier and uniform.
1240    
1241    LIVE and KILL say that a list of registers are live or killed at the
1242    program point where they appear. No spill code is generated when an
1243    element of the 'regs' field is spilled, but the register is moved to
1244    the 'spilled' (which is present, more for debugging than anything else).
1245    
1246    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
1247    We used to generate:
1248    
1249            DEFFREG f1
1250            f1 := f2 + f3
1251            trapb
1252    
1253    but now generate:
1254    
1255            f1 := f2 + f3
1256            trapb
1257            LIVE {regs=[f1,f2,f3], spilled=[]}
1258    
1259    Furthermore, the DEFFREG (hack) required that all floating point instruction
1260    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
1261    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
1262    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
1263    this architecture).
1264    
1265    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
1266    all the architectures. This will result in further simplification of the
1267    register allocator that must be aware of them for coalescing purposes, and
1268    will also simplify certain aspects of the machine description that provides
1269    callbacks related to parallel copies.
1270    
1271    ANNOTATION should be obvious, and now INSTR represents the honest to God
1272    machine instruction set!
1273    
1274    The <arch>/instructions/<arch>Instr.sml files define certain utility
1275    functions for making porting easier -- essentially converting upper case
1276    to lower case. All machine instructions (of type instr) are in upper case,
1277    and the lower case form generates an MLRISC instruction. For example on
1278    the alpha we have:
1279    
1280      datatype instr =
1281         LDA of {r:cell, b:cell, d:operand}
1282       | ...
1283    
1284      val lda : {r:cell, b:cell, d:operand} -> instruction
1285        ...
1286    
1287    where lda is just (INSTR o LDA), etc.
1288    
1289    ----------------------------------------------------------------------
1290    Name: Matthias Blume
1291    Date: 2001/11/22 21:40:00 EST
1292    Tag: Release_110_37
1293    Description:
1294    
1295    Release 110.37.  This time for real.
1296    
1297    ----------------------------------------------------------------------
1298    Name: Matthias Blume
1299    Date: 2001/11/21 16:35:00 EST
1300    Tag: blume-20011121-foot-in-mouth
1301    Description:
1302    
1303    Removed the "Release_110_37" tag because of a serious bug.
1304    This will be re-tagged once the bug is fixed.
1305    
1306    ----------------------------------------------------------------------
1307    Name: Matthias Blume
1308    Date: 2001/11/21 16:14:00 EST
1309    Tag: blume-20011121-forgottenfile
1310    Description:
1311    
1312    Forgot to add a file.  (Just a .tex-file -- part of
1313    the CM manual source.)
1314    
1315    ----------------------------------------------------------------------
1316    Name: Matthias Blume
1317    Date: 2001/11/21 16:10:00 EST
1318    Tag: blume-20011121-invalid_110_37
1319    Description:
1320    
1321    Note: I removed the original tag "Release_110_37" from this commit
1322          because we found a serious bug in all non-x86 backends.
1323          - Matthias
1324    
1325    1. Modifications to the SML/NJ code generator and to the runtime system
1326       so that code object name strings are directly inserted into code
1327       objects at code generation time.  The only business the runtime system
1328       has with this is now to read the name strings on occasions.
1329       (The encoding of the name string has also changed somewhat.)
1330    
1331    2. CM now implements a simple "set calculus" for specifying export lists.
1332       In particular, it is now possible to refer to the export lists of
1333       other libraries/groups/sources and form unions as well as differences.
1334       See the latest CM manual for details.
1335    
1336    3. An separate notion of "proxy" libraries has again be eliminated from
1337       CM's model.  (Proxy libraries are now simply a special case of using
1338       the export list calculus.)
1339    
1340    4. Some of the existing libraries now take advantage of the new set
1341       calculus.
1342       (Notice that not all libraries have been converted because some
1343       of the existing .cm-files are supposed to be backward compatible
1344       with 110.0.x.)
1345    
1346    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
1347       "General.exnMessage"!  The former relies on a certain hook to be
1348       initialized, and that often does not happen in the stand-alone case.)
1349    
1350    ----------------------------------------------------------------------
1351    Name: Lal George
1352    Date: 2001/11/21  13:56:18 EST
1353    Tag: george-2001121-pseudo-ops
1354    Description:
1355    
1356      Implemented a complete redesign of MLRISC pseudo-ops. Now there
1357      ought to never be any question of incompatabilities with
1358      pseudo-op syntax expected by host assemblers.
1359    
1360      For now, only modules supporting GAS syntax are implemented
1361      but more should follow, such as MASM, and vendor assembler
1362      syntax, e.g. IBM as, Sun as, etc.
1363    
1364    ----------------------------------------------------------------------
1365    Name: Matthias Blume
1366    Date: 2001/11/14 11:52:00 EST
1367    Tag: blume-20011114-srcname
1368    Description:
1369    
1370    1. Routed the name of the current source file to mlriscgen where it
1371       should be directly emitted into the code object.  (This last part
1372       is yet to be done.)
1373    
1374    2. Some cleanup of the pgraph code to make it match the proposal that
1375       I put out the other day.  (The proposal notwithstanding, things are
1376       still in flux here.)
1377    
1378    ----------------------------------------------------------------------
1379    Name: Lal George
1380    Date: 2001/11/14 09:44:04 EST
1381    Tag:
1382    Description:
1383    
1384      Fix for a backpatching bug reported by Allen.
1385    
1386      Because the boundary between short and long span-dependent
1387      instructions is +/- 128, there are an astounding number of
1388      span-dependent instructions whose size is over estimated.
1389    
1390      Allen came up with the idea of letting the size of span
1391      dependent instructions be non-monotonic, for a maxIter
1392      number of times, after which the size must be monotonically
1393      increasing.
1394    
1395      This table shows the number of span-dependent instructions
1396      whose size was over-estimated as a function of maxIter, for the
1397      file Parse/parse/ml.grm.sml:
1398    
1399         maxIter            # of instructions:
1400            10                      687
1401            20                      438
1402            30                      198
1403            40                        0
1404    
1405      In compiling the compiler, there is no significant difference in
1406      compilation speed between maxIter=10 and maxIter=40. Actually,
1407      my measurements showed that maxIter=40 was a tad faster than
1408      maxIter=10! Also 96% of the  files in the compiler reach a fix
1409      point within 13 iterations, so fixing maxIter at 40, while high,
1410      is okay.
1411    
1412    ----------------------------------------------------------------------
1413    Name: Matthias Blume
1414    Date: 2001/10/31 15:25:00 EST
1415    Tag: blume-20011031-pgraph
1416    Description:
1417    
1418    CKIT:
1419    * Changed the "Function" constructor of type Ast.ctype to carry optional
1420      argument identifiers.
1421    * Changed the return type of TypeUtil.getFunction accordingly.
1422    * Type equality ignores the argument names.
1423    * TypeUtil.composite tries to preserve argument names but gives up quickly
1424      if there is a mismatch.
1425    
1426    installation script:
1427    * attempts to use "curl" if available (unless "wget" is available as well)
1428    
1429    CM:
1430    * has an experimental implementation of "portable graphs" which I will
1431      soon propose as an implementation-independent library format
1432    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
1433    
1434    NLFFI-LIB:
1435    * some cleanup (all cosmetic)
1436    
1437    NLFFIGEN:
1438    * temporarily disabled the mechanism that suppresses ML output for
1439      C definitions whose identifiers start with an underscore character
1440    * generate val bindings for enum constants
1441    * user can request that only one style (light or heavy) is being used;
1442      default is to use both (command-line arguments: -heavy and -light)
1443    * fixed bug in handling of function types involving incomplete pointers
1444    * generate ML entry points that take record arguments (i.e., using
1445      named arguments) for C functions that have a prototype with named
1446      arguments
1447      (see changes to CKIT)
1448    
1449    ----------------------------------------------------------------------
1450    Name: Allen Leung
1451    Date: 2001/10/27 20:34:00 EDT
1452    Tag: leunga-20011027-x86-fast-fp-call
1453    Description:
1454    
1455       Fixed the bug described in blume-20010920-slowfp.
1456    
1457       The fix involves
1458          1. generating FCOPYs in FSTP in ia32-svid
1459          2. marking a CALL with the appropriate annotation
1460    
1461    ----------------------------------------------------------------------
1462    Name: Matthias Blume
1463    Date: 2001/10/16 11:32:00 EDT
1464    Tag: blume-20011016-netbsd
1465    Description:
1466    
1467    Underscore patch from Chris Richards (fixing problem with compiling
1468    runtime system under recent NetBSD).
1469    
1470    ----------------------------------------------------------------------
1471    Name: Allen Leung
1472    Date: 2001/10/12 17:18:32 EDT 2001
1473    Tag: leung-20011012-x86-printflowgraph
1474    Description:
1475    
1476    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
1477    
1478    ----------------------------------------------------------------------
1479    Name: Lal George
1480    Date: 2001/10/11 23:51:34 EDT
1481    Tag: george-20011011-too-many-instrs
1482    Description:
1483    
1484    The representation of a program point never expected to see more
1485    than 65536 instructions in a basic block!
1486    
1487    ----------------------------------------------------------------------
1488    Name: Lal George
1489    Date: 2001/10/09 09:41:37 EDT
1490    Tag: george-20011008-mlrisc-labels
1491    Description:
1492    
1493    Changed the machine description files to support printing of
1494    local and global labels in assembly code, based on host assembler
1495    conventions.
1496    
1497    ----------------------------------------------------------------------
1498    Name: Matthias Blume
1499    Date: 2001/09/25 15:25:00 EDT
1500    Tag: blume-20010925-exninfo
1501    Description:
1502    
1503    I provided a non-hook implementation of exnName (at the toplevel) and
1504    made the "dummy" implementation of exnMessage (at the toplevel) more
1505    useful: if nothing gets "hooked in", then at least you are going to
1506    see the exception name and a message indicating why you don't see more.
1507    
1508    [For the time being, programs that need exnMessage and want to use
1509    ml-build should either use General.exnMessage (strongly recommended) or
1510    refer to structure General at some other point so that CM sees a
1511    static dependency.]
1512    
1513    [Similar remarks go for "print" and "use":  If you want to use their
1514    functionality in stand-alone programs generated by ml-build, then use
1515    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
1516    
1517    ----------------------------------------------------------------------
1518    Name: Matthias Blume
1519    Date: 2001/09/20 17:28:00 EDT
1520    Tag: blume-20010920-slowfp
1521    Description:
1522    
1523    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
1524    
1525    ----------------------------------------------------------------------
1526    Name: Matthias Blume
1527    Date: 2001/09/20 17:20:00 EDT
1528    Tag: blume-20010920-canonicalpaths
1529    Description:
1530    
1531    0. Updated the BOOT file (something that I forgot to do earlier).
1532    
1533    1. Small internal change to CM so that it avoids "/../" in filenames
1534       as much as possible (but only where it is safe).
1535    
1536    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
1537       that arguments that contain delimiters are passed through correctly.
1538       This change also means that all "special" arguments of the form
1539       @SMLxxx... must come first.
1540    
1541    3. Changed install script to put relative anchor names for tool commands
1542       into pathconfig.
1543    
1544    ----------------------------------------------------------------------
1545    Name: Matthias Blume
1546    Date: 2001/09/18 15:35:00 EDT
1547    Tag: blume-20010918-readme11036
1548    Description:
1549    
1550    Added README files.
1551    
1552    ----------------------------------------------------------------------
1553    Name: Matthias Blume
1554    Date: 2001/09/18 11:45:00 EDT
1555    Tag: Release_110_36 (retag)
1556    Description:
1557    
1558    Fixed mistake in config/preloads. Retagged as 110.36.
1559    
1560    ----------------------------------------------------------------------
1561    Name: Matthias Blume
1562    Date: 2001/09/18 09:40:00 EDT
1563    Tag: Release_110_36_orig (tag changed)
1564    Description:
1565    
1566    New version (110.36).  New bootfiles.
1567    
1568    ----------------------------------------------------------------------
1569    Name: Matthias Blume
1570    Date: 2001/09/14 16:15:00 EDT
1571    Tag: blume-20010914-x86fastfp
1572    Description:
1573    
1574    John committed some changes that Allen made, in particular a (hopefully)
1575    correctly working version of the x86-fp module.
1576    
1577    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
1578    flag to "true".  Everything seems to compile to a fixpoint ok, and
1579    "mandelbrot" speeds up by about 15%.
1580    
1581    ----------------------------------------------------------------------
1582    Name: Matthias Blume
1583    Date: 2001/09/13 11:20:00 EDT
1584    Tag: blume-20010913-minimal
1585    Description:
1586    
1587    1. Stefan Monnier's patch to fix a miscompilation problem that
1588       was brought to light by John Reppy's work on Moby.
1589    
1590    2. Implemented a minimal "structure Compiler" that contains just
1591       "version" and "architecture".  The minimal version will be
1592       available when the full version is not.  This is for backward-
1593       compatibility with code that wants to test Compiler.version.
1594    
1595    ----------------------------------------------------------------------
1596    Name: Matthias Blume
1597    Date: 2001/08/28 14:03:00 EDT
1598    Tag: blume-20010828-ml-lex
1599    Description:
1600    
1601    Fix for bug 1581, received from Neophytos Michael.
1602    
1603    ----------------------------------------------------------------------
1604    Name: Matthias Blume
1605    Date: 2001/08/27 11:20:00 EDT
1606    Tag: blume-20010827-readme11035
1607    Description:
1608    
1609    Fleshed out the README file for 110.35.
1610    
1611    ----------------------------------------------------------------------
1612    Name: Matthias Blume
1613    Date: 2001/08/24 17:10:00 EDT
1614    Tag: Release_110_35
1615    Description:
1616    
1617    New version number (110.35).  New bootfiles.
1618    
1619    ----------------------------------------------------------------------
1620    Name: Lal George
1621    Date: 2001/08/24 13:47:18 EDT 2001
1622    Tag: george-20010824-MLRISC-graphs
1623    Description:
1624    
1625     removed clusters from MLRISC completely and replaced with graphs.
1626    
1627    ----------------------------------------------------------------------
1628    Name: Matthias Blume
1629    Date: 2001/08/23 17:50:00 EDT
1630    Tag: blume-20010823-toplevel
1631    Description:
1632    
1633    - some reorganization of the code that implements various kinds of
1634      environments in the compiler (static, dynamic, symbolic, combined)
1635    - re-implemented the EnvRef module so that evalStream works properly
1636      (if the stream contains references to "use", "CM.make", etc.)
1637    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
1638    
1639    ----------------------------------------------------------------------
1640    Name: Matthias Blume
1641    Date: 2001/08/20 15:50 EDT
1642    Tag: blume20010820-slipup
1643    Description:
1644    
1645    I forgot to commit a few files.  Here they are...
1646    
1647    ----------------------------------------------------------------------
1648    Name: Matthias Blume
1649    Date: 2001/08/20 15:35:00 EDT
1650    Tag: blume-20010820-debugprof
1651    Description:
1652    
1653    !!!! NEW BOOTFILES !!!!
1654    
1655    This is another round of reorganizing the compiler sources.  This
1656    time the main goal was to factor out all the "instrumentation"
1657    passes (for profiling and backtracing) into their own library.
1658    The difficulty was to do it in such a way that it does not depend
1659    on elaborate.cm but only on elabdata.cm.
1660    
1661    Therefore there have been further changes to both elaborate.cm and
1662    elabdata.cm -- more "generic" things have been moved from the former
1663    to the latter.  As a result, I was forced to split the assignment
1664    of numbers indicating "primtyc"s into two portions: SML-generic and
1665    SML/NJ-specific.  Since it would have been awkward to maintain,
1666    I bit the bullet and actually _changed_ the mapping between these
1667    numbers and primtycs.  The bottom line of this is that you need
1668    a new set of bin- and bootfiles.
1669    
1670    I have built new bootfiles for all architectures, so doing a fresh
1671    checkout and config/install.sh should be all you need.
1672    
1673    The newly created library's name is
1674    
1675        $smlnj/viscomp/debugprof.cm
1676    
1677    and its sources live under
1678    
1679        src/compiler/DebugProf
1680    
1681    ----------------------------------------------------------------------
1682    Name: Matthias Blume
1683    Date: 2001/08/15 17:15:00 EDT
1684    Tag: blume-20010815-compreorg
1685    Description:
1686    
1687    This is a first cut at reorganizing the CM libraries that make up the
1688    core of the compiler.  The idea is to separate out pieces that could
1689    be used independently by tools, e.g., the parser, the typechecker, etc.
1690    
1691    The current status is a step in this direction, but it is not quite
1692    satisfactory yet.  Expect more changes in the future.
1693    
1694    Here is the current (new) organization...
1695    
1696        What used to be $smlnj/viscomp/core.cm is now divided into
1697        six CM libraries:
1698    
1699             $smlnj/viscomp/basics.cm
1700                           /parser.cm
1701                           /elabdata.cm
1702                           /elaborate.cm
1703                           /execute.cm
1704                           /core.cm
1705    
1706        The CM files for these libraries live under src/system/smlnj/viscomp.
1707        All these libraries are proxy libraries that contain precisely
1708        one CM library component.  Here are the locations of the components
1709        (all within the src/compiler tree):
1710    
1711             Basics/basics.cm
1712             Parse/parser.cm
1713             ElabData/elabdata.cm
1714             Elaborator/elaborate.cm
1715             Execution/execute.cm
1716             core.cm
1717    
1718         [This organization is the same that has been used already
1719         for a while for the architecture-specific parts of the visible
1720         compiler and for the old version of core.cm.]
1721    
1722         As you will notice, many source files have been moved from their
1723         respective original locations to a new home in one of the above
1724         subtrees.
1725    
1726         The division of labor between the new libraries is the following:
1727    
1728             basics.cm:
1729                - Simple, basic definitions that pertain to many (or all) of
1730                  the other libraries.
1731             parser.cm:
1732                - The SML parser, producing output of type Ast.dec.
1733                - The type family for Ast is also defined and exported here.
1734             elabdata.cm:
1735                - The datatypes that describe input and output of the elaborator.
1736                  This includes types, absyn, and static environments.
1737             elaborator.cm:
1738                - The SML/NJ type checker and elaborator.
1739                  This maps an Ast.dec (with a given static environment) to
1740                  an Absyn.dec (with a new static environment).
1741                - This libraries implements certain modules that used to be
1742                  structures as functors (to remove dependencies on FLINT).
1743             execute.cm:
1744                - Everything having to do with executing binary code objects.
1745                - Dynamic environments.
1746             core.cm:
1747                - SML/NJ-specific instantiations of the elaborator and MLRISC.
1748                - Top-level modules.
1749                - FLINT (this should eventually become its own library)
1750    
1751    Notes:
1752    
1753    I am not 100% happy with the way I separated the elaborator (and its
1754    data structures) from FLINT.  Two instances of the same problem:
1755    
1756        1. Data structures contain certain fields that carry FLINT-specific
1757           information.  I hacked around this using exn and the property list
1758           module from smlnj-lib.  But the fact that there are middle-end
1759           specific fields around at all is a bit annoying.
1760    
1761        2. The elaborator calculates certain FLINT-related information.  I tried
1762           to make this as abstract as I could using functorization, but, again,
1763           the fact that the elaborator has to perform calculations on behalf
1764           of the middle-end at all is not nice.
1765    
1766        3. Having to used exn and property lists is unfortunate because it
1767           weakens type checking.  The other alternative (parameterizing
1768           nearly *everything*) is not appealing, though.
1769    
1770    I removed the "rebinding =" warning hack because due to the new organization
1771    it was awkward to maintain it.  As a result, the compiler now issues some of
1772    these warnings when compiling init.cmi during bootstrap compilation. On
1773    the plus side, you also get a warning when you do, for example:
1774       val op = = Int32.+
1775    which was not the case up to now.
1776    
1777    I placed "assign" and "deref" into the _Core structure so that the
1778    code that deals with the "lazy" keyword can find them there.  This
1779    removes the need for having access to the primitive environment
1780    during elaboration.
1781    
1782    ----------------------------------------------------------------------
1783    Name: Matthias Blume
1784    Date: 2001/08/13
1785    Tag: blume-20010813-closures
1786    Description:
1787    
1788    This fix was sent to us by Zhong Shao.  It is supposed to improve the
1789    performance of certain loops by avoiding needless closure allocation.
1790    
1791    ----------------------------------------------------------------------
1792    Name: Lal George
1793    Date: 2001/07/31 10:03:23 EDT 2001
1794    Tag: george-20010731-x86-fmalloc
1795    Description: Fixed bug in x86 calls
1796    
1797        There was a bug where call instructions would mysteriously
1798        vanish. The call instruction had to be one that returned
1799        a floating point value.
1800    
1801    ----------------------------------------------------------------------
1802    Name: Lal George
1803    Date: 2001/07/19 16:36:29 EDT 2001
1804    Tag: george-20010719-simple-cells
1805    Description:
1806    
1807    I have dramatically simplified the interface for CELLS in MLRISC.
1808    
1809    In summary, the cells interface is broken up into three parts:
1810    
1811      1. CellsBasis : CELLS_BASIS
1812    
1813            CellsBasis is a top level structure and common for all
1814            architectures.  it contains the definitions of basic datatypes
1815            and utility  functions over these types.
1816    
1817      2. functor Cells() : CELLS
1818    
1819            Cells generates an interface for CELLS that incorporates the
1820            specific resources on the target architecture, such as the
1821            presence of special register classes, their number and size,
1822            and various useful substructures.
1823    
1824      3. <ARCH>CELLS
1825    
1826            e.g. SparcCells: SPARCCELLS
1827    
1828            <ARCH>CELLS usually contains additional bindings for special
1829            registers  on the architecture, such as:
1830    
1831                    val r0 : cell           (* register zero *)
1832                    val y : cell            (* Y register *)
1833                    val psr : cell          (* processor status register *)
1834                    ...
1835    
1836            The structure returned by applying the Cells functor is opened
1837            in this interface.
1838    
1839    The main implication of all this is that the datatypes for cells is
1840    split between CellsBasis and CELLS -- a fairly simple change for user
1841    code.
1842    
1843    In the old scheme the CELLS interface had a definitional binding of
1844    the form:
1845    
1846            signature CELLS = sig
1847    
1848               structure CellsBasis = CellsBasis
1849    
1850               ...
1851    
1852            end
1853    
1854    With all the sharing constraints that goes on in MLRISC, this old
1855    design  quickly leads to errors such as:
1856    
1857            "structure definition spec inside of sharing ... "
1858    
1859    
1860    and appears to require an unacceptable amount of sharing and where
1861    constraint hackery.
1862    
1863    I think this error message (the interaction of definitional specs and
1864    sharing) requires more explanation on our web page.
1865    
1866    ----------------------------------------------------------------------
1867    Name: Matthias Blume
1868    Date: 2001/07/19 15:00:00 EDT
1869    Tag: blume-20010719-libreorg
1870    Description:
1871    
1872    This update puts together a fairly extensive but straightforward change
1873    to the way the libraries that implement the interactive system are
1874    organized:
1875    
1876       The biggest change is the elimination of structure Compiler.  As a
1877       replacement for this structure, there is now a CM library
1878       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
1879       that exports all the substructures of the original structure Compiler
1880       directly.  So instead of saying Compiler.Foo.bar one now simply
1881       says Foo.bar.  (The CM libraries actually export a collection of
1882       structures that is richer than the collection of substructures of
1883       structure Compiler.)
1884    
1885       To make the transition smooth, there is a separate library called
1886       $smlnj/compiler/compiler.cm which puts together and exports the
1887       original structure Compiler (or at least something very close to it).
1888    
1889       There are five members of the original structure Compiler
1890       that are not exported directly but which instead became members
1891       of a new structure Backend (described by signature BACKEND).  These are:
1892       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
1893       Interact (: INTERACT), structure Machine (: MACHINE), and val
1894       architecture (: string).
1895    
1896       Structure Compiler.Version has become structure CompilerVersion.
1897    
1898       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
1899       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
1900       or x86, respectively.
1901       Each of these exports the same frontend structures that
1902       $smlnj/compiler.cm exports.  But they do not have a structure Backend
1903       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
1904       Hppa, PPC, Sparc, or X86, respectively.
1905    
1906       Library $smlnj/compiler/all.cm exports the union of the exports of
1907       $smlnj/compiler/<arch>.cm
1908    
1909       There are no structures <Arch>Compiler anymore, use
1910       $smlnj/compiler/<arch>.cm instead.
1911    
1912       Library host-compiler-0.cm is gone.  Instead, the internal library
1913       that instantiates CM is now called cm0.cm.  Selection of the host
1914       compiler (backend) is no longer done here but. (Responsibility for it
1915       now lies with $smlnj/compiler/current.cm.  This seems to be more
1916       logical.)
1917    
1918       Many individual files have been moved or renamed.  Some files have
1919       been split into multiple files, and some "dead" files have been deleted.
1920    
1921    Aside from these changes to library organization, there are also changes
1922    to the way the code itself is organized:
1923    
1924       Structure Binfile has been re-implemented in such a way that it no
1925       longer needs any knowledge of the compiler.  It exclusively deals
1926       with the details of binfile layout.  It no longer invokes the
1927       compiler (for the purpose of creating new prospective binfile
1928       content), and it no longer has any knowledge of how to interpret
1929       pickles.
1930    
1931       Structure Compile (: COMPILE) has been stripped down to the bare
1932       essentials of compilation.  It no longer deals with linking/execution.
1933       The interface has been cleaned up considerably.
1934    
1935       Utility routines for dealing with linking and execution have been
1936       moved into their own substructures.
1937    
1938       (The ultimate goal of these changes is to provide a light-weight
1939       binfile loader/linker (at least for, e.g., stable libraries) that
1940       does not require CM or the compiler to be present.)
1941    
1942    CM documentation has been updated to reflect the changes to library
1943    organization.
1944    
1945    ----------------------------------------------------------------------
1946    Name: Matthias Blume
1947    Date: 2001/07/10 17:30:00 EDT
1948    Tag: Release_110_34
1949    Description:
1950    
1951    Minor tweak to 110.34 (re-tagged):
1952    
1953      - README.html file added to CVS repository
1954      - runtime compiles properly under FreeBSD 3.X and 4.X
1955    
1956    ----------------------------------------------------------------------
1957    Name: Matthias Blume
1958    Date: 2001/07/10 17:30:00 EDT
1959    Tag: Release_110_34
1960    Description:
1961    
1962    New version number (110.34). New bootfiles.
1963    
1964    ----------------------------------------------------------------------
1965    Name: Matthias Blume
1966    Date: 2001/07/09 16:00:00 EDT
1967    Tag: blume-20010709-more-varargs
1968    Description:
1969    
1970    I changed the handling of varargs in ml-nlffigen again:
1971    The ellipsis ... will now simply be ignored (with an accompanying warning).
1972    
1973    The immediate effect is that you can actually call a varargs function
1974    from ML -- but you can't actually supply any arguments beyond the ones
1975    specified explicitly.  (For example, you can call printf with its format
1976    string, but you cannot pass additional arguments.)
1977    
1978    This behavior is only marginally more useful than the one before, but
1979    it has the advantage that a function or, more importantly, a function
1980    type never gets dropped on the floor, thus avoiding follow-up problems with
1981    other types that refer to the offending one.
1982    
1983    ----------------------------------------------------------------------
1984    Name: Matthias Blume
1985    Date: 2001/07/09 11:25:00 EDT
1986    Tag: blume-20010709-varargs
1987    Description:
1988    
1989    1. ckit-lib.cm now exports structure Error
1990    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1991       with a warning accompanied by a source location.  Moreover, it
1992       merely skips the offending function or type and proceeds with the
1993       rest of its work.u  As a result, one can safely feed C code containing
1994       "..." to ml-nlffigen.
1995    3. There are some internal improvements to CM, providing slightly
1996       more general string substitutions in the tools subsystem.
1997    
1998    ----------------------------------------------------------------------
1999    Name: Matthias Blume
2000    Date: 2001/06/27 15:10:00 EDT
2001    Tag: blume-20010627-concur
2002    Description:
2003    
2004    Fixed a small bug in CM's handling of parallel compilation.
2005    (You could observe the bug by Control-C-interrupting an ordinary
2006    CMB.make or CM.stabilize and then attaching some compile servers.
2007    The result was that all of a sudden the previously interrupted
2008    compilation would continue on its own.  This was because of
2009    an over-optimization: CM did not bother to clean out certain queues
2010    when no servers were attached "anyway", resulting in the contents
2011    of these queues to grab control when new servers did get attached.)
2012    
2013    There is also another minor update to the CM manual.
2014    
2015    ----------------------------------------------------------------------
2016    Name: Matthias Blume
2017    Date: 2001/06/26 16:15:00 EDT
2018    Tag: blume-20010626-cmdoc
2019    Description:
2020    
2021    Minor typo fixed in CM manual (syntax diagram for libraries).
2022    
2023    ----------------------------------------------------------------------
2024    Name: Matthias Blume
2025    Date: 2001/06/25 22:55:00 EDT
2026    Tag: blume-20010625-x86pc
2027    Description:
2028    
2029    Fixed a nasty bug in the X86 assembly code that caused signal
2030    handlers to fail (crash) randomly.
2031    
2032    ----------------------------------------------------------------------
2033    Name: Matthias Blume
2034    Date: 2001/06/25 12:05:00 EDT
2035    Tag: blume-20010625-nlffigen
2036    Description:
2037    
2038    This update fixes a number of minor bugs in ml-nlffigen as reported by
2039    Nick Carter <nbc@andrew.cmu.edu>.
2040    
2041      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
2042      2. Default names for generated files are now derived from the name of
2043         the C file *without its directory*.  In particular, this causes generated
2044         files to be placed locally even if the C file is in some system directory.
2045      3. Default names for generated signatures and structures are also derived
2046         from the C file name without its directory.  This avoids silly things
2047         like "structure GL/GL".
2048         (Other silly names are still possible because ml-nlffigen does not do
2049          a thorough check of whether generated names are legal ML identifiers.
2050          When in doubt, use command line arguments to force particular names.)
2051    
2052    ----------------------------------------------------------------------
2053    Name: Matthias Blume
2054    Date: 2001/06/21 12:25:00 EDT
2055    Tag: blume-20010621-eXene
2056    Description:
2057    
2058    eXene now compiles and (sort of) works again.
2059    
2060    The library name (for version > 110.33) is $/eXene.cm.
2061    
2062    I also added an new example in src/eXene/examples/nbody.  See the
2063    README file there for details.
2064    
2065    ----------------------------------------------------------------------
2066    Name: Matthias Blume
2067    Date: 2001/06/20 16:40:00 EDT
2068    Tag: blume-20010620-cml
2069    Description:
2070    
2071    CML now compiles and works again.
2072    
2073    Libraries (for version > 110.33):
2074    
2075      $cml/cml.cm            Main CML library.
2076      $cml/basis.cm          CML's version of $/basis.cm.
2077      $cml/cml-internal.cm   Internal helper library.
2078      $cml/core-cml.cm       Internal helper library.
2079      $cml-lib/trace-cml.cm  Tracing facility.
2080      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
2081    
2082    The installer (config/install.sh) has been taught how to properly
2083    install this stuff.
2084    
2085    ----------------------------------------------------------------------
2086    Name: Matthias Blume
2087    Date: 2001/06/19 17:55:00 EDT
2088    Tag: blume-20010619-instantiate
2089    Description:
2090    
2091    This un-breaks the fix for bug 1432.
2092    (The bug was originally fixed in 110.9 but I broke it again some
2093    time after that.)
2094    
2095    ----------------------------------------------------------------------
2096    Name: Matthias Blume
2097    Date: 2001/06/19 17:25:00 EDT
2098    Tag: blume-20010619-signals
2099    Description:
2100    
2101    This should (hopefully) fix the long-standing signal handling bug.
2102    (The runtime system was constructing a continuation record with an
2103    incorrect descriptor which would cause the GC to drop data on the floor...)
2104    
2105    ----------------------------------------------------------------------
2106    Name: Matthias Blume
2107    Date: 2001/06/15 15:05:00 EDT
2108    Tag: blume-20010615-moresparc
2109    Description:
2110    
2111    Here is a short late-hour update related to Sparc c-calls:
2112    
2113     -- made handling of double-word arguments a bit smarter
2114    
2115     -- instruction selection phase tries to collapse certain clumsily
2116        constructed ML-Trees; typical example:
2117    
2118            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
2119    
2120        This currently has no further impact on SML/NJ since mlriscGen does
2121        not seem to generate such patterns in the first place, and c-calls
2122        (which did generate them in the beginning) has meanwhile been fixed
2123        so as to avoid them as well.
2124    
2125    ----------------------------------------------------------------------
2126    Name: Matthias Blume
2127    Date: 2001/06/15 15:05:00 EDT
2128    Tag: blume-20010615-sparc
2129    Description:
2130    
2131    The purpose of this update is to provide an implementation of NLFFI
2132    on Sparc machines.
2133    
2134    Here are the changes in detail:
2135    
2136       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
2137       the Sparc implementation of the c-calls API.
2138       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
2139       for accessing the ML frame.  Thus, we have a real frame pointer and
2140       can freely modify %sp without need for an omit-frame-ptr phase.
2141       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
2142       this case.
2143       * ml-nlffigen has been taught to produce code for different architectures
2144       and calling conventions.
2145       * In a way similar to what was done in the x86 case, the Sparc
2146       backend uses its own specific extension to mltree.  (For example,
2147       it needs to be able to generate UNIMP instructions which are part
2148       of the calling convention.)
2149       * ml-nlffi-lib was reorganized to make it more modular (in particular,
2150       to make it easier to plug in new machine- and os-dependent parts).
2151    
2152    There are some other fairly unrelated bug fixes and cleanups as well:
2153    
2154       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
2155       that they properly share their libraries with existing SML/NJ libraries.
2156       * I fixed a minor cosmetic bug in CM, supressing certain spurious
2157       follow-up error messages.
2158       * Updates to CM/CMB documentation.
2159    
2160    TODO items:
2161    
2162       * MLRISC should use a different register as its asmTemp on the Sparc.
2163         (The current %o2 is a really bad choice because it is part of the
2164         calling conventions, so things might interfere in unexpected ways.)
2165    
2166    ----------------------------------------------------------------------
2167    Name: Matthias Blume
2168    Date: 2001/06/07
2169    Tag: blume-20010607-calls
2170    Description:
2171    
2172    A number of internal changes related to C calls and calling conventions:
2173    
2174    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
2175       number of bytes popped implicitly (by the callee).  In most cases
2176       this field is 0 but on x86/win32 it is some non-zero value.  This
2177       is information provided for the benefit of the "omit-frameptr" pass.
2178    2. The CALL instruction on the x86 carries a similar "pops" field.
2179       The instruction selection phase copies its value from the ML-Tree
2180       CALL statement.
2181    3. On all other architectures, the instruction selection phase checks
2182       whether "pops=0" and complains if not.
2183    4. The c-calls implementation for x86 now accepts two calling conventions:
2184       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
2185       up after the call and pops is set to 0.  For "stdcall", the caller
2186       does nothing, leaving the cleanup to the callee; pops is set to
2187       the number of bytes that were pushed onto the stack.
2188    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
2189       distinguish between "ccall" and "stdcall".
2190    6. The UNIMP instruction has been added to the supported Sparc instruction
2191       set. (This is needed for implementing the official C calling convention
2192       on this architecture.)
2193    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
2194       work with the latest CM.
2195    
2196    ----------------------------------------------------------------------
2197    Name: Matthias Blume
2198    Date: 2001/06/05 15:10:00 EDT
2199    Tag: blume-20010605-cm-index
2200    Description:
2201    
2202    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
2203    
2204    1. CM can now generate "index files".  These are human-readable files
2205       that list on a per-.cm-file basis each toplevel symbol defined or
2206       imported.  The location of the index file for
2207       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
2208       To enable index-file generation, set CM.Control.generate_index to true
2209       or export an environment-symbol: export CM_GENERATE_INDEX=true.
2210    
2211       The CM manual has been updated accordingly.
2212    
2213    2. I made some slight modifications to the c-calls API in MLRISC.
2214    
2215         a) There is now a callback to support saving/restoring of
2216            dedicated but caller-save registers around the actual call
2217            instruction.
2218         b) One can optionally specify a comment-annotation for the
2219            call instruction.
2220    
2221    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
2222       (For example, the comment annotation shows the C prototype of
2223        the function being called.)
2224    
2225    ----------------------------------------------------------------------
2226    Name: Matthias Blume
2227    Date: 2001/06/01 13:30:00 EDT
2228    Tag: blume-20010601-nlffi-cleanup
2229    Description:
2230    
2231    This is mostly a cleanup of MLFFI stuff:
2232    
2233       - some signature files have been put into a more exposed place
2234       - the ugly 'f type parameter is gone (simplifies types tremendously!)
2235       - ml-nlffigen changed accordingly
2236       - tutorial updated
2237    
2238    Other changes:
2239    
2240       - author's affiliation in CM manual(s) updated
2241       - some more recognized keywords added to Allen's sml.sty
2242    
2243    ----------------------------------------------------------------------
2244    Name: Matthias Blume
2245    Date: 2001/05/25 15:30:00 EDT
2246    Tag: blume-20010525-iptr
2247    Description:
2248    
2249      - put the official 110.33-README (as it appears on the ftp server) under
2250        CVS
2251      - fixed a small bug related to incomplete pointer types in
2252        ml-nlffigen
2253      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
2254        (it does not need the 'f type parameter)
2255    
2256    ----------------------------------------------------------------------
2257    Name: Matthias Blume
2258    Date: 2001/05/23 14:30:00 EDT
2259    Tag: Release_110_33
2260    Description:
2261    
2262    New version number (110.33).  New bootfiles.
2263    
2264    ----------------------------------------------------------------------
2265    Name: Matthias Blume
2266    Date: 2001/05/22 18:06:00 EDT
2267    Tag: blume-20010522-targets
2268    Description:
2269    
2270    Made install.sh use file config/targets.customized if it exists, falling
2271    back to config/targets if it doesn't.  This way one can have a customized
2272    version of the targets file without touching the "real thing", thus
2273    eliminating the constant fear of accidentally checking something bogus
2274    back into the CVS repository...  (File config/targets.customized must
2275    not be added to the repository!)
2276    
2277    ----------------------------------------------------------------------
2278    Name: Matthias Blume
2279    Date: 2001/05/22 16:30:00 EDT
2280    Tag: blume-20010522-minitut
2281    Description:
2282    
2283    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
2284       struct returns.
2285    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
2286       incomplete, preliminary documentation for NLFFI.
2287    
2288    ----------------------------------------------------------------------
2289    Name: Matthias Blume
2290    Date: 2001/05/14 11:30:00 EDT
2291    Tag: blume-20010514-script
2292    Description:
2293    
2294    Some bugs in install script fixed.
2295    
2296    In addition to that I also made a slight change to the NLFFI API:
2297    Functors generated by ml-nlffigen now take the dynamic library as a
2298    straight functor argument, not as a suspended one.  (The original
2299    functor code used to force the suspension right away anyway, so there
2300    was nothing gained by this complication of the interface.)
2301    
2302    ----------------------------------------------------------------------
2303    Name: Matthias Blume
2304    Date: 2001/05/11 14:35:00 EDT
2305    Tag: blume-20010511-ml-nlffi
2306    Description:
2307    
2308    I finally took the plunge and added my new FFI code to the main
2309    repository.  For x86-linux it is now ready for prime-time.
2310    
2311    There are two new subdirectories of "src":
2312    
2313      - ml-nlffi-lib:
2314           The utility library for programs using the FFI interface.
2315           Here is the implementation of $/c.cm and its associated low-level
2316           partners $/c-int.cm and $/memory.cm.
2317      - ml-nlffigen:
2318           A stand-alone program for generating ML glue code from C source
2319           code.
2320    
2321    Building ml-nlffigen requires $/ckit-lib.cm.
2322    
2323    The config/install.sh script has been updates to do the Right Thing
2324    (hopefully).
2325    
2326    Notice that the source tree for the C-Kit will not be put under "src"
2327    but directly under the installation root directory.  (This is the
2328    structure that currently exists on the CVS server when you check out
2329    module "sml".)  Fortunately, config/install.sh knows about this oddity.
2330    
2331    Bugs: No documentation yet.
2332    
2333    ----------------------------------------------------------------------
2334    Name: Matthias Blume
2335    Date: 2001/05/09 16:35:00 EDT
2336    Tag: blume-20010509-cpscontract
2337    Description:
2338    
2339    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
2340    wrapper/unwrapper elimination did not decrement usage counts and some
2341    dead variables got overlooked by the dead-up logic.)
2342    
2343    ----------------------------------------------------------------------
2344    Name: Lal George
2345    Date: 2001/05/08  17:26:09 EDT
2346    Tag: george-20010508-omit-frameptr
2347    Description:
2348    
2349    Changes to implement the omit-frame-pointer optimization to support
2350    raw C calls. For now, there is only support on the Intel x86, but
2351    other architectures will follow as more experience is gained with this.
2352    
2353    
2354    ----------------------------------------------------------------------
2355    Name: Matthias Blume
2356    Date: 2001/05/07 14:40:00 EDT
2357    Tag: blume-20010507-proxies
2358    Description:
2359    
2360    I made into "proxy libraries" all libraries that qualify for such a
2361    change.  (A qualifying library is a library that has another library or
2362    groups as its sole member and repeats that member's export list
2363    verbatim.  A proxy library avoids this repetition by omitting its export
2364    list, effectively inheriting the list that its (only) member exports.
2365    See the CM manual for more explanation.)
2366    The main effect is that explicit export lists for these libraries
2367    do not have to be kepts in sync, making maintenance a bit easier.
2368    
2369    I also added copyright notices to many .cm-files.
2370    
2371    Last but not least, I made a new set of bootfiles.
2372    
2373    ----------------------------------------------------------------------
2374    Name: Matthias Blume
2375    Date: 2001/05/04 17:00:00 EDT
2376    Tag: blume-20010504-cm-lsplit
2377    Description:
2378    
2379    0. John merged pending changes to $/smlnj-lib.cm
2380    
2381    1. Allen's previous change accidentally backed out of one of Lal's
2382       earlier changes.  I undid this mistake (re-introducing Lal's change).
2383    
2384    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
2385       within the compiler where applicable.  There is some code simplification
2386       because of that.
2387    
2388    3. The "split" phase (in FLINT) is now part of the default list of phases.
2389       Compiler.Control.LambdaSplitting.* can be used to globally control the
2390       lambda-splitting (cross-module-inlining) engine.  In addition to that,
2391       it can now also be controlled on a per-source basis: CM has been taught
2392       a new tool parameter applicable to ML source files.
2393    
2394       - To turn lambda-splitting off completely:
2395            local open Compiler.Control.LambdaSplitting in
2396                val _ = set Off
2397            end
2398       - To make "no lambda-splitting" the global default (but allow per-source
2399         overriding); this is the initial setting:
2400            local open Compiler.Control.LambdaSplitting in
2401                val _ = set (Default NONE)
2402            end
2403       - To make "lambda-splitting with aggressiveness a" the global default
2404         (and allow per-source overriding):
2405            local open Compiler.Control.LambdaSplitting in
2406                val _ = set (Default (SOME a))
2407            end
2408    
2409       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
2410         write (in the respective .cm-file):
2411            a.sml (lambdasplitting:off)
2412       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
2413            a.sml (lambdasplitting:on)
2414       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
2415         <a> is a decimal non-negative integer):
2416            a.sml (lambdasplitting:<a>)
2417       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
2418            a.sml (lambdasplitting:infinity)
2419       - To use the global default for a.sml:
2420            a.sml (lambdasplitting:default)
2421         or simply
2422            a.sml
2423    
2424    ----------------------------------------------------------------------
2425    Name: Allen Leung
2426    Date: 2001/05/04 01:57:00 EDT
2427    Tag: leunga-20010504-sync
2428    Description:
2429    
2430      MLRISC features.
2431    
2432      1. Fix to CMPXCHG instructions.
2433      2. Changed RA interface to allow annotations in callbacks.
2434      3. Added a new method to the stream interface to allow annotations updates.
2435    
2436    ----------------------------------------------------------------------
2437    Name: Matthias Blume
2438    Date: 2001/05/01 11:45:00 EDT
2439    Tag: blume-20010501-pcedittmp
2440    Description:
2441    
2442    Changed install.sh to use the current working directory instead of
2443    /usr/tmp for a temporary file (pcedittmp).  The previous choice
2444    of /usr/tmp caused trouble with MacOS X because of file premission
2445    problems.
2446    
2447    ----------------------------------------------------------------------
2448    Name: Matthias Blume
2449    Date: 2001/04/20 11:10:00 EDT
2450    Tag: blume-20010420-inMLflag
2451    Description:
2452    
2453     - added vp_limitPtrMask to vproc-state.h
2454       (for use by the raw-C-calls mechanism to implement proper interrupt
2455        handling)
2456     - made the ML compiler aware of various data-structure offsets so it
2457       can generate code for accessing the vp_inML flag and vp_limitPtrMask
2458     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
2459       raw C-calls
2460    
2461    ----------------------------------------------------------------------
2462    Name: Lal George
2463    Date: 2001/04/20 09:15:28 EDT
2464    Tag: george-20010420-macosX
2465    Description:
2466    
2467     - Changes to port to Mac OS X; Darwin.
2468    
2469     - In the process I found that sqrt was broken on the PPC, because the
2470       fsqrt instruction is not implemented.
2471    
2472    ----------------------------------------------------------------------
2473    Name: Matthias Blume
2474    Date: 2001/04/18 12:45:00 EDT
2475    Tag: blume-20010418-ccalls
2476    Description:
2477    
2478     - fixed two off-by-4 errors in the x86-specific c-calls implementation
2479       (this bug prevented structure arguments containing pointers from being
2480        passed correctly)
2481     - changed the raw-C-call code in mlriscGen.sml in such a way that
2482       structure arguments are represented as a pointer to the beginning
2483       of the structure (instead of having a series of synthesized arguments,
2484       one for each structure member)
2485    
2486     - made makeml script's verbosity level configurable via environment
2487       variable (MAKEML_VERBOSITY)
2488    
2489     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
2490       in rawmem-x86.sml; we are now using the real thing
2491    
2492    ----------------------------------------------------------------------
2493    Name: Matthias Blume
2494    Date: 2001/03/22 16:25:00 EST
2495    Tag: blume-20010322-bootfiles
2496    Description:
2497    
2498    Created a new set of bootfiles (for your automatic installation convenience).
2499    
2500    ----------------------------------------------------------------------
2501    Name: Matthias Blume
2502    Date: 2001/03/22 15:10:00 EST
2503    Tag: blume-20010322-rawmem-parcm
2504    Description:
2505    
2506    1. All "raw memory access" primitives for the new FFI are implemented now
2507       (at least on the x86).
2508    2. Some further cleanup of CM's parallel make mechanism.
2509    
2510    ----------------------------------------------------------------------
2511    Name: Matthias Blume
2512    Date: 2001/03/19 17:53:00 EST
2513    Tag: blume-20010319-parallel
2514    Description:
2515    
2516    Parallel make (using compile servers) now works again.
2517    
2518    To this end, CM.stabilize and CMB.make have been modified to work in
2519    two passes when compile servers are attached:
2520       1. Compile everything, do not perform stabilization; this pass
2521          uses compile servers
2522       2. Stabilize everything; this pass does not use compile servers
2523    If there are no compile servers, the two passes are combined into one
2524    (as before).  Splitting the passes increases the inherent parallelism
2525    in the dependency graph because the entire graph including all
2526    libraries is available at the same time.  This, in turn, improves
2527    server utilization.  The downside is that the master process will
2528    have to do some extra work after compilation is done (because for
2529    technical reasons it must re-read all the binfiles during stabilization).
2530    
2531    ----------------------------------------------------------------------
2532    Name: Matthias Blume
2533    Date: 2001/03/16 12:22:00 EST
2534    Tag: blume-20010316-bootfiles
2535    Description:
2536    
2537    Created a new set of bootfiles (for your automatic installation convenience).
2538    
2539    ----------------------------------------------------------------------
2540    Name: Matthias Blume
2541    Date: 2001/03/16 11:00:00 EST
2542    Tag: blume-20010316-MLTREE-fixup
2543    Description:
2544    
2545    This is a minor fixup for an (untagged) earlier commit by Allen.
2546    (A file was missing).
2547    
2548    ----------------------------------------------------------------------
2549    Name: Allen Leung
2550    Date: Mon Mar  5 18:54:57 EST 2001
2551    Tag: leunga-20010305-cut-support
2552    
2553    1. New support for alternative control-flow in MLTREE.
2554       Currently we support
2555    
2556          FLOW_TO(CALL ...., [k1,...,kn])
2557    
2558       This is needed for 'cuts to' in C-- and try/handle-like constructs
2559       in Moby
2560    
2561       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
2562    
2563    2. Register Allocator
2564    
2565       Changes in interface [from Fermin, John]
2566    
2567    3. Alpha 8-bit SLL support [Fermin]
2568    
2569    4. All architectures
2570    
2571       A new module (ClusterExpandCopies) for expanding parallel copies.
2572    
2573    ----------------------------------------------------------------------
2574    Name: Allen Leung
2575    Date: 2001/02/27 23:07:00 EST
2576    Tag: leunga-20010227-minor-stuff
2577    
2578    1. Alpha bug fix for CMOVNE
2579    2. Handle mltree COND(..,FCMP ...,...)
2580    3. Bug fix in simplifier
2581    
2582    ----------------------------------------------------------------------
2583    Name: Matthias Blume
2584    Date: 2001/01/30 17:50:00 EST
2585    Tag: blume-20010130-sync
2586    Description:
2587    
2588    This is just a minor update to sync my devel branch with the main brach.
2589    The only visible change is the addition of some README files.
2590    
2591    ----------------------------------------------------------------------
2592    Name: Matthias Blume
2593    Date: 2001/01/12 23:30:00 JST
2594    Tag: blume-20010112-bootfiles
2595    Description:
2596    
2597    Made a new set of bootfiles that goes with the current state of the
2598    repository.
2599    
2600    ----------------------------------------------------------------------
2601    Name: Matthias Blume
2602    Date: 2001/01/12 21:20:00 JST
2603    Tag: blume-20010112-sync
2604    Description:
2605    
2606    I am just flushing out some minor changes that had accumulated in
2607    my private branch in order to sync with the main tree.  (This is
2608    mainly because I had CVS trouble when trying to merge _into_ my
2609    private branch.)
2610    
2611    Most people should be completely unaffected by this.
2612    
2613    ----------------------------------------------------------------------
2614    Name: Allen Leung
2615    Date: Thu Jan 11 21:03:00 EST 2001
2616    Tag: leunga-20010111-labexp=mltree
2617    Description:
2618    
2619    1.  Removed the type LabelExp and replace it by MLTree.
2620    2.  Rewritten mltree-simplify with the pattern matcher tool.
2621    3.  There were some bugs in alpha code generator which would break
2622        64-bit code generation.
2623    4.  Redo the tools to generate code with the
2624    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
2625        are now generated by perl scripts.
2626    
2627    ----------------------------------------------------------------------
2628    Name: Matthias Blume
2629    Date: 2001/01/10 21:55:00 JST
2630    Tag: blume-20010110-rcc
2631    Description:
2632    
2633    The RCC stuff now seems to work (but only on the x86).
2634    This required hacking of the c-calls interface (and -implementation) in
2635    MLRISC.
2636    
2637    Normal compiler users should be unaffected.
2638    
2639    ----------------------------------------------------------------------
2640    Name: Matthias Blume
2641    Date: 2001/01/09 01:20:00 JST
2642    Tag: blume-20010109-rcc
2643    Description:
2644    
2645    This is a fairly big patch, flushing out a large number of pending
2646    changes that I made to my development copy over the last couple of days.
2647    
2648    Of practical relevance at this moment is a workaround for a pickling
2649    bug that Allen ran into the other day.  The cause of the bug itself is
2650    still unknown and it might be hard to fix it properly, but the
2651    workaround has some merits of its own (namely somewhat reducing pickling
2652    overhead for certain libraries).  Therefore, I think this solution should
2653    be satisfactory at this time.
2654    
2655    The rest of the changes (i.e., the vast majority) has to do with my
2656    ongoing efforts of providing direct support for C function calls from
2657    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
2658    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
2659    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
2660    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
2661    handle RCC.
2662    
2663    The last part (the changes to mlriscGen.sml) are still known to be
2664    wrong on the x86 and not implemented on all other architectures.  But
2665    the infrastructure is in place. I had to change a few functor
2666    signatures in the backend to be able to route the CCalls interface
2667    from MLRISC there, and I had to specialize the mltree type (on the
2668    x86) to include the necessary extensions. (The extensions themselves
2669    were already there and redy to go in MLRISC/x86).
2670    
2671    Everything should be very happy as soon as someone helps me with
2672    mlriscGen.sml...
2673    
2674    In any case, nothing of this should matter to anyone as long as the
2675    new primop is not being used (which is going to be the case unless you
2676    find it where I hid it :). The rest of the compiler is completely
2677    unaffected.
2678    
2679    ----------------------------------------------------------------------
2680    Name: Matthias Blume
2681    Date: 2001/01/05 00:30:00 JST
2682    Tag: blume-20010105-primops
2683    Description:
2684    
2685    Added some experimental support for work that I am doing right now.
2686    These changes mostly concern added primops, but there is also a new
2687    experimental C library in the runtime system (but currently not enabled
2688    anywhere except on Linux/X86).
2689    
2690    In the course of adding primops (and playing with them), I discovered that
2691    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
2692    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
2693    To recover, I made type information in INL_PRIM mandatory and changed
2694    prim.sml as well as built-in.sml accordingly.  The InLine structure now
2695    has complete, correct type information (i.e., no bottom types).
2696    
2697    Since all these changes mean that we need new binfiles, I also bumped the
2698    version number to 110.32.1.
2699    
2700    ----------------------------------------------------------------------
2701    Name: Matthias Blume
2702    Date: 2000/12/30 22:10:00 JST
2703    Tag: blume-20001230-various
2704    Description:
2705    
2706    Added proxy libraries for MLRISC and let MLRISC libraries refer
2707    to each other using path anchors.  (See CM manual for explanation.)
2708    
2709    Updated CM documentation.
2710    
2711    Fixed some bugs in CM.
2712    
2713    Implemented "proxy" libraries (= syntactic sugar for CM).
2714    
2715    Added "-quiet" option to makeml and changed runtime system accordingly.
2716    
2717    Added cleanup handler for exportML to reset timers and compiler stats.
2718    
2719    ----------------------------------------------------------------------
2720    Name: Lal George
2721    Date: 2000/12/22 22:22:58 EST 2000
2722    Tag: Release_110_32
2723    Description:
2724    
2725            Infinite precision used throughout MLRISC.
2726            see MLRISC/mltree/machine-int.sig
2727    
2728    ----------------------------------------------------------------------
2729    Name: Matthias Blume
2730    Date: 2000/12/22 23:16:00 JST
2731    Tag: blume-20001222-warn
2732    Description:
2733    
2734    Corrected wording and formatting of some CM warning message which I
2735    broke in my previous patch.
2736    
2737    ----------------------------------------------------------------------
2738    Name: Matthias Blume
2739    Date: 2000/12/22 21:20:00 JST
2740    Tag: blume-20001222-anchorenv
2741    Description:
2742    
2743    Fixed CM's handling of anchor environments in connection with CMB.make.
2744    
2745    ----------------------------------------------------------------------
2746    Name: Matthias Blume
2747    Date: 2000/12/22 13:15:00 JST
2748    Tag: blume-20001222-cleanup
2749    Description:
2750    
2751    Removed src/cm/ffi which does not (and did not) belong here.
2752    
2753    ----------------------------------------------------------------------
2754    Name: Matthias Blume
2755    Date: 2000/12/21 23:55:00 JST
2756    Tag: blume-20001221-exn
2757    Description:
2758    
2759    Probably most important: CM no longer silently swallows all exceptions
2760    in the compiler.
2761    Plus: some other minor CM changes.  For example, CM now reports some
2762    sizes for generated binfiles (code, data, envpickle, lambdapickle).
2763    
2764    ----------------------------------------------------------------------
2765    Name: Matthias Blume
2766    Date: 2000/12/15 00:01:05 JST
2767    Tag: blume-20001215-dirtool
2768    Description:
2769    
2770    - "dir" tool added.
2771    - improvements and cleanup to Tools structure
2772    - documentation updates
2773    
2774    ----------------------------------------------------------------------
2775    Name: Allen Leung
2776    Date: Thu Dec 14 03:45:24 EST 2000
2777    Description:
2778    Tag:  leunga-20001214-int-inf
2779    Description:
2780    
2781       In IntInf, added these standard functions, which are missing from our
2782    implementation:
2783    
2784        andb : int * int -> int
2785        xorb : int * int -> int
2786        orb  : int * int -> int
2787        notb : int -> int
2788         <<   : int * word -> int
2789        ~>>  : int * word -> int
2790    
2791       Not tested, I hope they are correct.
2792    
2793    ----------------------------------------------------------------------
2794    Name: Allen Leung
2795    Date: Fri Dec  8 19:23:26 EST 2000
2796    Description:
2797    Tag:  leunga-20001208-nowhere
2798    Description:
2799    
2800      Slight improvements to the 'nowhere' tool to handle OR-patterns,
2801    to generate better error messages etc.  Plus a brief manual.
2802    
2803    ----------------------------------------------------------------------
2804    Name: Lal George
2805    Date: 2000/12/08 09:54:02 EST 2000
2806    Tag: Release_110_31
2807    Description:
2808    
2809    - Version 110.31
2810    ----------------------------------------------------------------------
2811    Name: Allen Leung
2812    Date: Thu Dec  7 22:01:04 EST 2000
2813    Tag:  leunga-20001207-cell-monster-hack
2814    Description:
2815    
2816    Major MLRISC internal changes.  Affect all clients.
2817    Summary:
2818    
2819    1.  Type CELLS.cell = int is now replaced by a datatype.
2820        As a result, the old regmap is now gone.  Almost all interfaces
2821        in MLRISC change as a consequence.
2822    
2823    2.  A new brand version of machine description tool (v3.0) that generates
2824        modules expecting the new interface.  The old version is removed.
2825    
2826    3.  The RA interface has been further abstracted into two new functors.
2827        RISC_RA and X86RA.  These functors have much simpler interfaces.
2828        [See also directory MLRISC/demo.]
2829    
2830    4.  Some other new source->source code generation tools are available:
2831    
2832        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
2833        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
2834           I use this tool to generate the peephole optimizers---with the new
2835           cell type changes, peephole rules are becoming difficult to write
2836           without conditional pattern matching.
2837    
2838    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
2839        cover the entire MLRISC source tree so many things broke.
2840    
2841    6.  CM files have been moved to the subdirectory MLRISC/cm.
2842        They are moved because there are a lot of them and they clutter up the
2843        root dir.
2844    
2845    7.  More detailed documentation to come...
2846    
2847        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
2848        a makeml -rebuild first.  This is because of other other
2849        changes that Matthias has made (see below).
2850    
2851    
2852    ----------------------------------------------------------------------
2853    Name: Matthias Blume
2854    Date: 2000/11/30 23:12:00 JST
2855    Tag: blume-20001130-filereorg
2856    Description:
2857    
2858    Some manual updates and some file reorganizations in CM.
2859    
2860    ----------------------------------------------------------------------
2861    Name: Matthias Blume
2862    Date: 2000/11/24 17:45:00 JST
2863    Tag: blume-20001124-link
2864    Description:
2865    
2866    Drastically improved link traversal code for the case that the dynamic
2867    value was already loaded at bootstrap time.  As a result, CM and CMB
2868    now both load blazingly fast -- even on a very slow machine.  Also,
2869    memory consumption has been further reduced by this.
2870    
2871    Warning: The format of the PIDMAP file has changed.  THerefore, to
2872    bootstrap you have to do this:
2873    
2874    1. Run CMB.make
2875    2. Make a symbolic link for the boot directory:
2876         ln -s sml.boot.ARCH-OS xxx
2877    3. "Rebuild" the boot directory:
2878         ./makeml -boot xxx -rebuild sml ; rm xxx
2879    4. Boot normally:
2880          ./makeml
2881    
2882    ----------------------------------------------------------------------
2883    Name: Matthias Blume
2884    Date: 2000/11/21 21:20:00 JST
2885    Tag: blume-20001121-tools
2886    Description:
2887    
2888    Continued hacking on autoloading problem -- with success this time.
2889    Also changed tool-plugin mechanism.  See new CM manual.
2890    
2891    ----------------------------------------------------------------------
2892    Name: Matthias Blume
2893    Date: 2000/11/19 14:30:00 JST
2894    Tag:  blume-20001119-autoload
2895    Description:
2896    
2897    Some hacking to make autoloading faster.  Success for CMB, no success
2898    so far for CM.  There is a reduced structure CM' that autoloads faster.
2899    (This is a temporary, non-documented hack to be eliminated again when
2900    the general problem is solved.)
2901    
2902    ----------------------------------------------------------------------
2903    Name: Matthias Blume
2904    Date: 2000/11/17 14:10:00 JST
2905    Tag: blume-20001117-pickle-lib
2906    Description:
2907    
2908    1. Eliminated comp-lib.cm
2909    2. Made pickle-lib.cm
2910    3. Eliminated all uses of intset.sml (from comp-lib.cm)
2911    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
2912       equivalent constructs from smlnj-lib.cm (INtHashTable).
2913    5. Point 4. also goes for those uses of intmap.* in MLRISC.
2914       Duplicated intmap modules thrown out.
2915    6. Hunted down all duplicated SCC code and replaced it with
2916       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
2917    7. Rewrote Feedback module.
2918    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
2919       should be thrown out and equivalent modules from smlnj-lib.cm
2920       should be used (IntRedBlackSet, IntListSet, ...).
2921    
2922    Confirmed that compiler compiles to fixpoint.
2923    
2924    ----------------------------------------------------------------------
2925    Name: Allen Leung
2926    Date: 2000/11/10 18:00:00
2927    Tag: leunga-20001110-new-x86-fp
2928    
2929    A new x86 floating point code generator has been added.
2930    By default this is turned off.  To turn this on, do:
2931    
2932        CM.autoload "$smlnj/compiler.cm";
2933        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
2934    
2935    Changes:
2936    
2937    1.  Changed FTAN to FPTAN so that the assembly output is correct.
2938    2.  Changed the extension callback for FTANGENT to generate:
2939    
2940              fptan
2941              fstp  %st(0)
2942        instead of
2943              fptan
2944              fstpl ftempmem
2945    
2946    3.  Numerous assembly fixes for x86.
2947    
2948    5.  Cleaned up the machine code output module x86/x86MC.sml and added
2949        support for a whole bunch of instructions and addressing modes:
2950    
2951          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
2952          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
2953          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
2954          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
2955          fxch %st(n)
2956          fld %st(n)
2957          fst %st(n)
2958          fst mem
2959          fstp %st(n)
2960          fucom %st(n)
2961          fucomp %st(n)
2962    
2963        All these are now generated when the fast fp mode is turned on.
2964    
2965    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
2966    
2967    ----------------------------------------------------------------------
2968    Name: Matthias Blume
2969    Date: 2000/11/09 11:20:00 JST
2970    Tag: blume-20001109-scc
2971    Description:
2972    
2973    Eliminated some code duplication:
2974    
2975    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
2976       (Otherwise the functor is useless.)
2977    2. Used GraphSCCFn where SCCUtilFun was used previously.
2978    3. Got rid of SCCUtilFun (in comp-lib.cm).
2979    
2980    ----------------------------------------------------------------------
2981    Name: Lal George
2982    Date: 2000/11/06 09:02:21 EST 2000
2983    Tag: Release_110_30
2984    Description:
2985    
2986    - Version 110.30
2987    ----------------------------------------------------------------------
2988    Name: Matthias Blume
2989    Date: 2000/11/04 14:45:00
2990    Tag: blume-20001104-mlbuild
2991    Description:
2992    
2993    - Made ml-build faster on startup.
2994    - Documentation fixes.
2995    
2996    ----------------------------------------------------------------------
2997    Name: Matthias Blume
2998    Date: 2000/11/02 17:00:00 JST
2999    Tag: blume-20001102-condcomp
3000    Description:
3001    
3002    - Small tweaks to pickler -- new BOOTFILES!
3003    - Version bumped to 110.29.2.
3004    - Added conditional compilation facility to init.cmi (see comment there).
3005    ----------------------------------------------------------------------
3006    Name: Allen Leung
3007    Date: 2000/10/23 19:31:00
3008    Tag: leunga-20001023-demo-ra
3009    
3010    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
3011    2. Test programs for the graph library updated
3012    3. Some new MLRISC demo programs added
3013    
3014    ----------------------------------------------------------------------
3015    Name: Matthias Blume
3016    Date: 2000/08/31 22:15:00 JST
3017    Tag: blume-20001017-errmsg
3018    Description:
3019    
3020    More error message grief: Where there used to be no messages, there
3021    now were some that had bogus error regions.  Fixed.
3022    
3023    ----------------------------------------------------------------------
3024    Name: Matthias Blume
3025    Date: 2000/08/31 17:30:00 JST
3026    Tag: blume-20001017-v110p29p1
3027    Description:
3028    
3029    I made a version 110.29.1 with new bootfiles.
3030    
3031    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
3032              CM documentation changes and a small bugfix in CM's error reporting.
3033    
3034    ----------------------------------------------------------------------
3035    Name: Lal George
3036    Date: 2000/09/27 14:42:35 EDT
3037    Tag: george-20000927-nodestatus
3038    Description:
3039    
3040    Changed the type of the nodestatus, so that:
3041    
3042            SPILLED(~1)             is now SPILLED
3043            SPILLED(m) where m>=0   is now MEMREG(m)
3044            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
3045    
3046    ----------------------------------------------------------------------
3047    Name: Matthias Blume
3048    Date: 2000/09/07 14:45:00 JST
3049    Tag: blume-20000907-cmerrmsg
3050    Description:
3051    
3052    Small tweak to CM to avoid getting ML syntax error messages twice.
3053    
3054    ----------------------------------------------------------------------
3055    Name: Matthias Blume
3056    Date: 2000/08/31 18:00:00 JST
3057    Tag: blume-20000831-cvsbootfiles
3058    Description:
3059    
3060    New URL for boot files (because the 110.29 files on the BL server do
3061    now work correctly with my updated install scripts for yacc and lex).
3062    
3063    ----------------------------------------------------------------------
3064    Name: Matthias Blume
3065    Date: 2000/08/08 12:33:00 JST
3066    Tag: blume-20000808-manual
3067    Description:
3068    
3069    Tiny update to CM manual.
3070    
3071    ----------------------------------------------------------------------
3072    Name: Allen Leung
3073    Date: 2000/08/7 19:31:00
3074    Tag: leunga-20000807-a-whole-bunch-of-stuff
3075    
3076      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
3077    and Mobdy.
3078    
3079    1.  x86
3080    
3081       a.  Fixes to peephole module by John and Dan.
3082       b.  Assembly fix to SETcc by Allen.
3083       c.  Fix to c-call by John.
3084       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
3085       e.  Instruction selection optimization to SETcc as suggested by John.
3086    
3087           For example,
3088    
3089            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
3090    
3091           should generate:
3092    
3093            MOVL a, x
3094            SUBL b, x
3095            SHRL 31, x
3096    
3097    2.  IR stuff
3098    
3099         A bunch of new DJ-graph related algorithms added.  These
3100         speed up SSA construction.
3101    
3102    3.  SSA + Scheduling
3103    
3104         Added code for SSA and scheduling to the repository
3105    
3106    ----------------------------------------------------------------------
3107    Name: Lal George
3108    Date: 2000/07/27 11:53:14 EDT
3109    
3110    Tag: lal-20000727-linux-ppc
3111    Description:
3112    
3113     Made changes to support Linux PPC.
3114     p.s. I have confirmation that the 110.29 boot files work fine.
3115    
3116    ----------------------------------------------------------------------
3117    Name: Matthias Blume
3118    Date: 2000/07/27 17:40:00 JST
3119    Tag: blume-20000727-scripts
3120    Description:
3121    
3122    !!!! WARNING !!!!
3123    You must recompile the runtime system!
3124    !!!! WARNING !!!!
3125    
3126    This is basically another round of script-enhancements:
3127    
3128    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
3129       and undefine CM preprocessor symbols.
3130    
3131    2. ml-build avoids generating a new heap image if it finds that the
3132       existing one is still ok.  (The condition is that no ML file had to
3133       be recompiled and all ML files are found to be older that the heap
3134       file.)
3135    
3136       To make this work smoothly, I also hacked the runtime system as
3137       well as SMLofNJ.SysInfo to get access to the heap image suffix
3138       (.sparc-solaris, ...) that is currently being used.
3139    
3140       Moreover, the signature of CM.mk_standalone has changed.  See the
3141       CM manual.
3142    
3143    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
3144       CM manual for details.)
3145    
3146    4. More CM manual updates:
3147        - all of the above has been documented.
3148        - there is now a section describing the (CM-related) command line
3149          arguments that are accepted by the "sml" command
3150    
3151    ----------------------------------------------------------------------
3152    Name: Matthias Blume
3153    Date: 2000/07/25 16:20:00 JST
3154    Tag: blume-20000725-makedepend
3155    Description:
3156    
3157    Added a script called ml-makedepend.  This can be used in makefiles
3158    for Unix' make in a way very similar to the "makedepend" command for
3159    C.
3160    
3161    The script internally uses function CM.sources.
3162    
3163    Synopsis:
3164    
3165        ml-makedepend [-f makefile] cmfile targetname
3166    
3167    The default for the makefile is "makefile" (or "Makefile" should
3168    "makefile" not exist).
3169    
3170    ml-makedepend adds a cmfile/targetname-specific section to this
3171    makefile (after removing the previous version of this section).  The
3172    section contains a single dependency specification with targetname on
3173    the LHS (targetname is an arbitrary name), and a list of files derived
3174    from the cmfile on the RHS.  Some of the files on the RHS are
3175    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
3176    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
3177    path names.  The makefile writer is responsible for making sure that
3178    these variables have correct at the time "make" is invoked.
3179    
3180    ----------------------------------------------------------------------
3181    Name: Matthias Blume
3182    Date: 2000/07/22 23:30:00 JST
3183    Tag: blume-20000722-urlupdate
3184    Description:
3185    
3186    Changed BOOT and config/srcarchiveurl to point to BL server:
3187    
3188        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
3189    
3190    ----------------------------------------------------------------------
3191    Name: Matthias Blume
3192    Date: 2000/07/18 18:00:00 JST
3193    Tag: blume-20000718-Version_110_29
3194    Description:
3195    
3196    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
3197    
3198    2. Updated config/version to 110.29
3199    
3200    3. Updated config/srcarchiveurl
3201    
3202    3. New boot files!
3203       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
3204    
3205    ----------------------------------------------------------------------
3206    Name: Matthias Blume
3207    Date: 2000/07/11 13:58:00 JST
3208    Tag: blume-20000711-doctypo
3209    Description:
3210    
3211    Fixed a few typos in CM manual.
3212    
3213    ----------------------------------------------------------------------
3214    Name: Allen Leung
3215    Date: 2000/06/15 00:38:00
3216    Tag: leunga-20000704-sparc-x86
3217    
3218    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
3219    2. fix to x86 RET bug [found by Dan Grossman]
3220    3. sparc assembly bug fix for ticc instructions [found by Fermin]
3221    
3222       Affects c-- and moby only
3223    
3224    ----------------------------------------------------------------------
3225    Name: Matthias Blume
3226    Date: 2000/07/04 15:26:00
3227    Tag: blume-20000704-trigger
3228    Description:
3229    
3230    1. Improvements to CM manual.
3231    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
3232       of getting a back-trace.  The function, when called, raises an
3233       internal exception which explicitly carries the full back-trace history,
3234       so it is unaffected by any intervening handle-raise pairs ("trivial"
3235       or not).  The interactive loop will print that history once it arrives
3236       at top level.
3237       Short of having all exceptions implicitly carry the full history, the
3238       recommended way of using this facility is:
3239         - compile your program with instrumentation "on"
3240         - run it, when it raises an exception, look at the history
3241         - if the history is "cut off" because of some handler, go and modify
3242           your program so that it explicitly calls BTrace.trigger
3243         - recompile (still instrumented), and rerun; look at the full history
3244    
3245    ----------------------------------------------------------------------
3246    Name: Matthias Blume
3247    Date: 2000/07/03 15:36:00 JST
3248    Tag: blume-20000702-manual
3249    Description:
3250    
3251    Small corrections and updates to CM manual.
3252    
3253    ----------------------------------------------------------------------
3254    Name: Matthias Blume
3255    Date: 2000/06/29 16:04:00 JST
3256    Tag: blume-20000629-yacctool
3257    Description:
3258    
3259    Changes:
3260    
3261    1. Class "mlyacc" now takes separate arguments to pass options to
3262       generated .sml- and .sig-files independently.
3263    2. Corresponding CM manual updates.
3264    3. BTrace module now also reports call sites.  (However, for loop clusters
3265       it only shows from where the cluster was entered.)  There are associated
3266       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
3267    
3268    ----------------------------------------------------------------------
3269    Name: Matthias Blume
3270    Date: 2000/06/27 16:51:00 JST
3271    Tag: blume-20000627-noweb
3272    Description:
3273    
3274    Changes:
3275    
3276     1. Implemented "subdir" and "witness" options for noweb tool.
3277        This caused some slight internal changes in CM's tool implementation.
3278     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
3279        some remaining issues from earlier path anchor changes.
3280     3. Updated CM manual accordingly.
3281    
3282     4. Changed implementation of back-tracing so that I now consider it
3283        ready for prime-time.
3284    
3285        In particular, you don't have to explicitly trigger the back-trace
3286        anymore.  Instead, if you are running BTrace-instrumented code and
3287        there is an uncaught exception (regardless of whether or not it was
3288        raised in instrumented code), the top-level evalloop will print
3289        the back-trace.
3290    
3291        Features:
3292    
3293          - Instrumented and uninstrumented code work together seemlessly.
3294            (Of course, uninstrumented code is never mentioned in actual
3295             back-traces.)
3296    
3297          - Asymptotic time- and space-complexity of instrumented code is
3298            equal to that of uninstrumented code.  (This means that
3299            tail-recursion is preserved by the instrumentation phase.)
3300    
3301          - Modules whose code has been instrumented in different sessions
3302            work together without problem.
3303    
3304          - There is no penalty whatsoever on uninstrumented code.
3305    
3306          - There is no penalty on "raise" expressions, even in
3307            instrumented code.
3308    
3309        A potential bug (or perhaps it is a feature, too):
3310    
3311          A back-trace reaches no further than the outermost instrumented
3312          non-trivial "raise".  Here, a "trivial" raise is one that is the
3313          sole RHS of a "handle" rule.  Thus, back-traces reach trough
3314    
3315               <exp> handle e => raise e
3316    
3317          and even
3318    
3319               <exp> handle Foo => raise Bar
3320    
3321          and, of course, through
3322    
3323               <exp> handle Foo => ...
3324    
3325         if the exception was not Foo.
3326    
3327         Back-traces always reach right through any un-instrumented code
3328         including any of its "handle" expressions, trivial or not.
3329    
3330       To try this out, do the following:
3331    
3332         - Erase all existing binfiles for your program.
3333           (You may keep binfiles for those modules where you think you
3334            definitely don't need back-tracing.)
3335         - Turn on back-trace instrumentation:
3336              SMLofNJ.Internals.BTrace.mode (SOME true);
3337         - Recompile your program.  (I.e., run "CM.make" or "use".)
3338         - You may now turn instrumentation off again (if you want):
3339              SMLofNJ.Internals.BTrace.mode (SOME false);
3340         - Run your program as usual.  If it raises an exception that
3341           reaches the interactive toplevel, then a back-trace will
3342           automatically be printed.  After that, the toplevel loop
3343           will print the exception history as usual.
3344    
3345    ----------------------------------------------------------------------
3346    Name: Matthias Blume
3347    Date: 2000/06/26 09:56:46 JST
3348    Tag: blume-20000626-setup
3349    Description:
3350    
3351    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
3352          ML code before and after compiling a file (e.g., to set compiler
3353          flags)
3354    
3355    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
3356              - associated changes to btrace.sml (BTrace instrumentation pass)
3357              - cleaner implementation of btimp.sml (BTrace tracing and report
3358                module)
3359    
3360    CM manual: * new path encoding documented
3361               * description of setup-parameter to "sml" added
3362    
3363    The biggest user-visible change to back-tracing is that it is no
3364    longer necessary to compile all traced modules within the same
3365    session.  (This was a real limitation.)
3366    
3367    ----------------------------------------------------------------------
3368    Name: Matthias Blume
3369    Date: 2000/06/24 12:40:00 JST
3370    Tag: blume-20000624-startup
3371    Description:
3372    
3373    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
3374    bit too often -- to put it mildly. :)
3375    
3376    ----------------------------------------------------------------------
3377    Name: Matthias Blume
3378    Date: 2000/06/23 18:20:00 JST
3379    Tag: blume-20000623-btrace
3380    Description:
3381    
3382    This updates adds a backtrace facility to aid programmers in debugging
3383    their programs.  This involves the following changes:
3384    
3385    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
3386       keeping track of the current call stack.  When programs are compiled
3387       in a special mode, the compiler will insert calls to these hooks
3388       into the user program.
3389       "Hook" means that it is possible for different implementations of
3390       back-tracing to register themselves (at different times).
3391    
3392    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
3393       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
3394       It can be turned on using this call:
3395         SMLofNJ.Internals.BTrace.mode (SOME true);
3396       Turning it off again:
3397         SMLofNJ.Internals.BTrace.mode (SOME false);
3398       Querying the current status:
3399         SMLofNJ.Internals.BTrace.mode NONE;
3400       Annotated programs are about twice as big as normal ones, and they
3401       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
3402       where all hooks do nothing).  The slowdown with a plugin that is
3403       actually useful (such as the one supplied by default) is even greater,
3404       but in the case of the default plugin it is still only an constant
3405       factor (amortized).
3406    
3407    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
3408       with a sub-structure BTrace for controlling back-tracing.  In particular,
3409       the above-mentioned function "mode" controls whether the annotation
3410       phase is invoked by the compiler.  Another important function is
3411       "trigger": when called it aborts the current execution and causes
3412       the top-level loop to print a full back-trace.
3413    
3414    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
3415       for back-tracing.  It keeps track of the dynamic call stack and in
3416       addition to that it keeps a partial history at each "level" of that
3417       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
3418       d and b (at separate times, dynamically), then the report will show:
3419    
3420       GOTO   d
3421             /c
3422       GOTO  \b
3423       CALL   a
3424    
3425       This shows that there was an initial non-tail call of a, then a
3426       tail-call to b or c, looping behavior in a cluster of functions that
3427       consist of b and c, and then a goto from that cluster (i.e., either from
3428       b or from c) to d.
3429    
3430       Note that (depending on the user program) the amount of information
3431       that the back-trace module has to keep track of at each level is bounded
3432       by a constant.  Thus, the whole implementation has the same asymptotical
3433       complexity as the original program (both in space and in time).
3434    
3435    5. compiler/TopLevel/interact/evalloop.sml has been modified to
3436       handle the special exception SMLofNJ.Internals.BTrace.BTrace
3437       which is raised by the "trigger" function mentioned above.
3438    
3439    Notes on usage:
3440    
3441    - Annotated code works well together with unannotated code:
3442    Unannotated calls simply do not show up at all in the backtrace.
3443    
3444    - It is not a good idea to let modules that were annotated during
3445    different sessions run at the same time.  This is because the compiler
3446    chooses small integers to identify individual functions, and there
3447    will be clashes if different modules were compiled in separate sessions.
3448    (Nothing will crash, and you will even be told about the clashes, but
3449    back-trace information will in general not be useful.)
3450    
3451    - Back-tracing can be confused by callcc and capture.
3452    
3453    - The only way of getting a back-trace right now is to explicitly
3454    invoke the "trigger" function from your user program.  Eventually, we
3455    should make every exception carry back-trace information (if
3456    available).  But since this creates more overhead at "raise"-time
3457    (similar to the current exnHistory overhead), I have not yet
3458    implemented this.  (The implementation will be rather easy.)  With
3459    exceptions carrying back-trace information, this facility will be even
3460    more useful because users don't need to modify their programs...
3461    
3462    - While it is possible to compile the compiler with back-trace
3463    annotations turned on (I did it to get some confidence in
3464    correctness), you must make absolutely sure that core.sml and
3465    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
3466    be compiled with annotation because there is no core access yet, but
3467    if you compile btimp.sml with annotation, then the system will go into
3468    an infinite recursion and crash.)
3469    Since CM currently does not know about BTrace, the only way to turn
3470    annotations on and off for different modules of the compiler is to
3471    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
3472    this is awkward and clumsy.
3473    
3474    Sample sessions:
3475    
3476    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
3477    - SMLofNJ.Internals.BTrace.mode (SOME true);
3478    [autoloading]
3479    [autoloading done]
3480    val it = false : bool
3481    - structure X = struct
3482    -     fun main n = let
3483    -         fun a (x, 0) = d x
3484    -           | a (x, n) = b (x, n - 1)
3485    -         and b (x, n) = c (x, n)
3486    -         and c (x, n) = a (x, n)
3487    -         and d x = e (x, 3)
3488    -         and e (x, 0) = f x
3489    -           | e (x, n) = e (x, n - 1)
3490    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
3491    -           | f n = n * g (n - 1)
3492    -         and g n = a (n, 3)
3493    -     in
3494    -         f n
3495    -     end
3496    - end;
3497    structure X : sig val main : int -> int end
3498    - X.main 3;
3499    *** BACK-TRACE ***
3500    GOTO   stdIn:4.2-13.20: X.main[2].f
3501    GOTO-( stdIn:4.2-13.20: X.main[2].e
3502    GOTO   stdIn:4.2-13.20: X.main[2].d
3503         / stdIn:4.2-13.20: X.main[2].a
3504         | stdIn:4.2-13.20: X.main[2].b
3505    GOTO-\ stdIn:4.2-13.20: X.main[2].c
3506    CALL   stdIn:4.2-13.20: X.main[2].g
3507    GOTO   stdIn:4.2-13.20: X.main[2].f
3508    GOTO-( stdIn:4.2-13.20: X.main[2].e
3509    GOTO   stdIn:4.2-13.20: X.main[2].d
3510         / stdIn:4.2-13.20: X.main[2].a
3511         | stdIn:4.2-13.20: X.main[2].b
3512    GOTO-\ stdIn:4.2-13.20: X.main[2].c
3513    CALL   stdIn:4.2-13.20: X.main[2].g
3514    GOTO   stdIn:4.2-13.20: X.main[2].f
3515    GOTO-( stdIn:4.2-13.20: X.main[2].e
3516    GOTO   stdIn:4.2-13.20: X.main[2].d
3517         / stdIn:4.2-13.20: X.main[2].a
3518         | stdIn:4.2-13.20: X.main[2].b
3519    GOTO-\ stdIn:4.2-13.20: X.main[2].c
3520    CALL   stdIn:4.2-13.20: X.main[2].g
3521    GOTO   stdIn:4.2-13.20: X.main[2].f
3522    CALL   stdIn:2.15-17.4: X.main[2]
3523    -
3524    
3525    (Note that because of a FLINt bug the above code currently does not
3526    compile without BTrace turned on.)
3527    
3528    Here is another example, using my modified Tiger compiler:
3529    
3530    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
3531    - SMLofNJ.Internals.BTrace.mode (SOME true);
3532    [autoloading]
3533    [autoloading done]
3534    val it = false : bool
3535    - CM.make "sources.cm";
3536    [autoloading]
3537    ...
3538    [autoloading done]
3539    [scanning sources.cm]
3540    [parsing (sources.cm):parse.sml]
3541    [creating directory CM/SKEL ...]
3542    [parsing (sources.cm):tiger.lex.sml]
3543    ...
3544    [wrote CM/sparc-unix/semant.sml]
3545    [compiling (sources.cm):main.sml]
3546    [wrote CM/sparc-unix/main.sml]
3547    [New bindings added.]
3548    val it = true : bool
3549    - Main.compile ("../testcases/merge.tig", "foo.out");
3550    *** BACK-TRACE ***
3551    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
3552    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3553    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
3554    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
3555    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3556    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3557    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
3558         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
3559    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
3560    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3561    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
3562    CALL   main.sml:1.18-118.4: Main.compile[2]
3563    -
3564    
3565    ----------------------------------------------------------------------
3566    Name: Matthias Blumen
3567    Date: 2000/06/21 18:00:00 JST
3568    Tag: blume-20000621-manual
3569    Description:
3570    
3571    CM manual update: Path environments documented.
3572    
3573    ----------------------------------------------------------------------
3574    Name: Matthias Blume
3575    Date: 2000/06/19 13:40:00
3576    Tag: blume-20000619-manual
3577    Description:
3578    
3579    CM manual and system/README update.  This only covers the fact that
3580    there are no more implicit anchors.  (Path environments and the "bind"
3581    option to "cm" have yet to be documented.)
3582    
3583    ----------------------------------------------------------------------
3584    Name: Matthias Blume
3585    Date: 2000/06/19 11:05:00 JST
3586    Tag: blume-20000619-chdir-bugfix
3587    Description:
3588    
3589    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
3590    
3591    ----------------------------------------------------------------------
3592    Name: Matthias Blume
3593    Date: 2000/06/18 22:00:10 JST
3594    Tag: blume-20000618-implicit-anchors-really-gone
3595    Description:
3596    
3597    I updates the previous HISTORY entry where I forgot to mention that
3598    implicit anchors are no longer with us.
3599    
3600    The current update also gets rid of the (now useless) controller
3601    CM.Control.implicit_anchors.
3602    
3603    ----------------------------------------------------------------------
3604    Name: Matthias Blume
3605    Date: 2000/06/16 17:30:00 JST
3606    Tag: blume-20000616-anchorenv
3607    Description:
3608    
3609    This patch implements the long anticipated (just kidding :) "anchor
3610    environment" mechanism.  In the course of doing this, I also
3611    re-implemented CM's internal "SrcPath" module from scratch.  The new
3612    one should be more robust in certain boundary cases.  In any case, it
3613    is a lot cleaner than its predecessor (IMHO).
3614    
3615    This time, although there is yet another boot file format change, I
3616    kept the unpickler backward-compatible.  As a result, no new bootfiles
3617    are necessary and bootstrapping is straightforward.  (You cannot read