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 676, Sat Jun 24 03:37:03 2000 UTC revision 1175, Mon Mar 25 03:38:44 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: 2002/03/24 22:40:00 EST
18    Tag: blume-20020324-bootfiles
19    Description:
20    
21    Made the bootfiles that Allen asked for.
22    
23    ----------------------------------------------------------------------
24    Name: Allen Leung
25    Date: 2002/03/23 15:50:00 EST
26    Tag: leunga-20020323-flint-cps-rcc-primops
27    Description:
28    
29      1. Changes to FLINT primops:
30    
31        (* make a call to a C-function;
32         * The primop carries C function prototype information and specifies
33         * which of its (ML-) arguments are floating point. C prototype
34         * information is for use by the backend, ML information is for
35         * use by the CPS converter. *)
36      | RAW_CCALL of { c_proto: CTypes.c_proto,
37                       ml_args: ccall_type list,
38                       ml_res_opt: ccall_type option,
39                       reentrant : bool
40                     } option
41       (* Allocate uninitialized storage on the heap.
42        * The record is meant to hold short-lived C objects, i.e., they
43        * are not ML pointers.  With the tag, the representation is
44        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
45        *)
46      | RAW_RECORD of {tag:bool,sz:int}
47      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
48    
49      2.  These CPS primops are now overloaded:
50    
51           rawload of {kind:numkind}
52           rawstore of {kind:numkind}
53    
54          The one argument form is:
55    
56             rawload {kind} address
57    
58          The two argument form is:
59    
60             rawload {kind} [ml object, byte-offset]
61    
62      3. RAW_CCALL/RCC now takes two extra arguments:
63    
64         a. The first is whether the C call is reentrant, i.e., whether
65            ML state should be saved and restored.
66         b. The second argument is a string argument specifying the name of
67            library and the C function.
68    
69         These things are currently not handled in the code generator, yet.
70    
71      4. In CProto,
72    
73         An encoding type of "bool" means "ml object" and is mapped into
74         C prototype of PTR.  Note that "bool" is different than "string",
75         even though "string" is also mapped into PTR, because "bool"
76         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
77    
78      5. Pickler/unpicker
79    
80         Changed to handle RAW_RECORD and newest RAW_CCALL
81    
82      6. MLRiscGen,
83    
84         1. Changed to handle the new rawload/rawstore/rawrecord operators.
85         2. Code for handling C Calls has been moved to a new module CPSCCalls,
86            in the file CodeGen/cpscompile/cps-c-calls.sml
87    
88      7. Added the conditional move operator
89    
90             condmove of branch
91    
92         to cps.  Generation of this is still buggy so it is currently
93         disabled.
94    
95    ----------------------------------------------------------------------
96    Name: Lal George
97    Date: 2002/03/22 14:18:25 EST
98    Tag: george-20020322-cps-branch-prob
99    Description:
100    
101    Implemented the Ball-Larus branch prediction-heuristics, and
102    incorporated graphical viewers for control flow graphs.
103    
104    Ball-Larus Heuristics:
105    ---------------------
106    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
107    
108    By design it uses the Dempster-Shafer theory for combining
109    probabilities.  For example, in the function:
110    
111        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
112    
113    the ball-larus heuristics predicts that the n=0 is unlikely
114    (OH-heuristic), and the 'then' branch is unlikely because of the
115    RH-heuristic -- giving the 'then' branch an even lower combined
116    probability using the Dempster-Shafer theory.
117    
118    Finally, John Reppy's loop analysis in MLRISC, further lowers the
119    probability of the 'then' branch because of the loop in the else
120    branch.
121    
122    
123    Graphical Viewing:
124    ------------------
125    I merely plugged in Allen's graphical viewers into the compiler. The
126    additional code is not much. At the top level, saying:
127    
128            Control.MLRISC.getFlag "cfg-graphical-view" := true;
129    
130    will display the graphical view of the control flow graph just before
131    back-patching.  daVinci must be in your path for this to work. If
132    daVinci is not available, then the default viewer can be changed
133    using:
134    
135            Control.MLRISC.getString "viewer"
136    
137    which can be set to "dot" or "vcg" for the corresponding viewers. Of
138    course, these viewers must be in your path.
139    
140    The above will display the compilation unit at the level of clusters,
141    many of which are small, boring, and un-interesting. Also setting:
142    
143            Control.MLRISC.getInt "cfg-graphical-view_size"
144    
145    will display clusters that are larger than the value set by the above.
146    
147    
148    ----------------------------------------------------------------------
149    Name: Matthias Blume
150    Date: 2002/03/21 22:20:00 EST
151    Tag: blume-20020321-kmp-bugfix
152    Description:
153    
154    Changed the interface to the KMP routine in PreString and fixed
155    a minor bug in one place where it was used.
156    
157    ----------------------------------------------------------------------
158    Name: Allen Leung
159    Date: 2002/03/21 20:30:00 EST
160    Tag: leunga-20020321-cfg
161    Description:
162    
163      Fixed a potential problem in cfg edge splitting.
164    
165    ----------------------------------------------------------------------
166    Name: Allen Leung
167    Date: 2002/03/21 17:15:00 EST
168    Tag: leunga-20020321-x86-fp-cfg
169    Description:
170    
171      1. Recoded the buggy parts of x86-fp.
172    
173         a. All the block reordering code has been removed.
174            We now depend on the block placement phases to do this work.
175    
176         b. Critical edge splitting code has been simplified and moved into the
177            CFG modules, as where they belong.
178    
179         Both of these were quite buggy and complex.  The code is now much, much
180         simpler.
181    
182      2. X86 backend.
183    
184         a. Added instructions for 64-bit support.  Instruction selection for
185            64-bit has not been committed, however, since that
186            requires changes to MLTREE which haven't been approved by
187            Lal and John.
188    
189         b. Added support for FUCOMI and FUCOMIP when generating code for
190            PentiumPro and above.  We only generate these instructions in
191            the fast-fp mode.
192    
193         c. Added cases for JP and JNP in X86FreqProps.
194    
195      3. CFG
196    
197         CFG now has a bunch of methods for edge splitting and merging.
198    
199      4. Machine description.
200    
201         John's simplification of MLTREE_BASIS.fcond broke a few machine
202         description things:
203    
204         rtl-build.{sig,sml} and hppa.mdl fixed.
205    
206         NOTE: the machine description stuff in the repository is still broken.
207               Again, I can't put my fixes in because that involves
208               changes to MLTREE.
209    
210    ----------------------------------------------------------------------
211    Name: Matthias Blume
212    Date: 2002/03/20 15:55:00 EST
213    Tag: blume-20020320-kmp
214    Description:
215    
216    Implemented Knuth-Morris-Pratt string matching in PreString and used
217    it for String.isSubstring, Substring.isSubstring, and
218    Substring.position.
219    
220    (Might need some stress-testing.  Simple examples worked fine.)
221    
222    ----------------------------------------------------------------------
223    Name: Matthias Blume
224    Date: 2002/03/19 16:37:00 EST
225    Tag: blume-20020319-witnesses
226    Description:
227    
228    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
229    
230    This implements a generic mechanism for changing constness qualifiers
231    anywhere within big C types without resorting to outright "casts".
232    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
233    modify the constness at the outermost level.)
234    The implementation of "convert" is based on the idea of "witness"
235    values -- values that are not used by the operation but whose types
236    "testify" to their applicability.  On the implementation side, "convert"
237    is simply a projection (returning its second curried argument).  With
238    cross-module inlining, it should not result in any machine code being
239    generated.
240    
241    ----------------------------------------------------------------------
242    Name: Matthias Blume
243    Date: 2002/03/15 16:40:00 EST
244    Tag: blume-20020315-basis
245    Description:
246    
247    Provided (preliminary?) implementations for
248    
249      {String,Substring}.{concatWith,isSuffix,isSubstring}
250    
251    and
252    
253      Substring.full
254    
255    Those are in the Basis spec but they were missing in SML/NJ.
256    
257    ----------------------------------------------------------------------
258    Name: Matthias Blume
259    Date: 2002/03/14 21:30:00 EST
260    Tag: blume-20020314-controls
261    Description:
262    
263    Controls:
264    ---------
265    
266    1. Factored out the recently-added Controls : CONTROLS stuff and put
267       it into its own library $/controls-lib.cm.  The source tree for
268       this is under src/smlnj-lib/Controls.
269    
270    2. Changed the names of types and functions in this interface, so they
271       make a bit more "sense":
272    
273          module -> registry
274          'a registry -> 'a group
275    
276    3. The interface now deals in ref cells only.  The getter/setter interface
277       is (mostly) gone.
278    
279    4. Added a function that lets one register an already-existing ref cell.
280    
281    5. Made the corresponding modifications to the rest of the code so that
282       everything compiles again.
283    
284    6. Changed the implementation of Controls.MLRISC back to something closer
285       to the original.  In particular, this module (and therefore MLRISC)
286       does not depend on Controls.  There now is some link-time code in
287       int-sys.sml that registers the MLRISC controls with the Controls
288       module.
289    
290    CM:
291    ---
292    
293      * One can now specify the lambda-split aggressiveness in init.cmi.
294    
295    ----------------------------------------------------------------------
296    Name: Allen Leung
297    Date: 2002/03/13 17:30:00 EST
298    Tag: leunga-20020313-x86-fp-unary
299    Description:
300    
301    Bug fix for:
302    
303    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
304    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
305    > - fun f(x,(y,z)) = Real.~ y;
306    > [autoloading]
307    > [autoloading done]
308    >       fchsl   (%eax), 184(%esp)
309    > Error: MLRisc bug: X86MCEmitter.emitInstr
310    >
311    > uncaught exception Error
312    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
313    
314    The problem was that the code generator did not generate any fp registers
315    in this case, and the ra didn't know that it needed to run the X86FP phase to
316    translate the pseudo fp instruction.   This only happened with unary fp
317    operators in certain situations.
318    
319    ----------------------------------------------------------------------
320    Name: Matthias Blume
321    Date: 2002/03/13 14:00:00 EST
322    Tag: blume-20020313-overload-etc
323    Description:
324    
325    1. Added _overload as a synonym for overload for backward compatibility.
326       (Control.overloadKW must be true for either version to be accepted.)
327    
328    2. Fixed bug in install script that caused more things to be installed
329       than what was requested in config/targets.
330    
331    3. Made CM aware of the (_)overload construct so that autoloading
332       works.
333    
334    ----------------------------------------------------------------------
335    Name: Matthias Blume
336    Date: 2002/03/12 22:03:00 EST
337    Tag: blume-20020312-url
338    Description:
339    
340    Forgot to update BOOT and srcarchiveurl.
341    
342    ----------------------------------------------------------------------
343    Name: Matthias Blume
344    Date: 2002/03/12 17:30:00 EST
345    Tag: blume-20020312-version110392
346    Description:
347    
348    Yet another version number bump (because of small changes to the
349    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
350    
351    Changes:
352    
353      The new pid generation scheme described a few weeks ago was overly
354      complicated.  I implemented a new mechanism that is simpler and
355      provides a bit more "stability":  Once CM has seen a compilation
356      unit, it keeps its identity constant (as long as you do not delete
357      those crucial CM/GUID/* files).  This means that when you change
358      an interface, compile, then go back to the old interface, and
359      compile again, you arrive at the original pid.
360    
361      There now also is a mechanism that instructs CM to use the plain
362      environment hash as a module's pid (effectively making its GUID
363      the empty string).  For this, "noguid" must be specified as an
364      option to the .sml file in question within its .cm file.
365      This is most useful for code that is being generated by tools such
366      as ml-nlffigen (because during development programmers tend to
367      erase the tool's entire output directory tree including CM's cached
368      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
369      revert to the old, broken behavior of SML/NJ, but in specific cases
370      where there is no danger of interface confusion, its use is ok
371      (I think).
372    
373      ml-nlffigen by default generates "noguid" annotations.  They can be
374      turned off by specifying -guid in its command line.
375    
376    ----------------------------------------------------------------------
377    Name: Lal George
378    Date: 2002/03/12 12 14:42:36 EST
379    Tag: george-20020312-frequency-computation
380    Description:
381    
382    Integrated jump chaining and static block frequency into the
383    compiler. More details and numbers later.
384    
385    ----------------------------------------------------------------------
386    Name: Lal George
387    Date: 2002/03/11 11 22:38:53 EST
388    Tag: george-20020311-jump-chain-elim
389    Description:
390    
391    Tested the jump chain elimination on all architectures (except the
392    hppa).  This is on by default right now and is profitable for the
393    alpha and x86, however, it may not be profitable for the sparc and ppc
394    when compiling the compiler.
395    
396    The gc test will typically jump to a label at the end of the cluster,
397    where there is another jump to an external cluster containing the actual
398    code to invoke gc. This is to allow factoring of common gc invocation
399    sequences. That is to say, we generate:
400    
401            f:
402               testgc
403               ja   L1      % jump if above to L1
404    
405            L1:
406               jmp L2
407    
408    
409    After jump chain elimination the 'ja L1' instructions is converted to
410    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
411    up being implemented in their long form (if L2 is far away) using:
412    
413            jbe     L3      % jump if below or equal to L3
414            jmp     L2
415         L3:
416            ...
417    
418    
419    For large compilation units L2  may be far away.
420    
421    
422    ----------------------------------------------------------------------
423    Name: Matthias Blume
424    Date: 2002/03/11 13:30:00 EST
425    Tag: blume-20020311-mltreeeval
426    Description:
427    
428    A functor parameter was missing.
429    
430    ----------------------------------------------------------------------
431    Name: Allen Leung
432    Date: 2002/03/11 10:30:00 EST
433    Tag: leunga-20020311-runtime-string0
434    Description:
435    
436       The representation of the empty string now points to a
437    legal null terminated C string instead of unit.  It is now possible
438    to convert an ML string into C string with InlineT.CharVector.getData.
439    This compiles into one single machine instruction.
440    
441    ----------------------------------------------------------------------
442    Name: Allen Leung
443    Date: 2002/03/10 23:55:00 EST
444    Tag: leunga-20020310-x86-call
445    Description:
446    
447       Added machine generation for CALL instruction (relative displacement mode)
448    
449    ----------------------------------------------------------------------
450    Name: Matthias Blume
451    Date: 2002/03/08 16:05:00
452    Tag: blume-20020308-entrypoints
453    Description:
454    
455    Version number bumped to 110.39.1.  NEW BOOTFILES!
456    
457    Entrypoints: non-zero offset into a code object where execution should begin.
458    
459    - Added the notion of an entrypoint to CodeObj.
460    - Added reading/writing of entrypoint info to Binfile.
461    - Made runtime system bootloader aware of entrypoints.
462    - Use the address of the label of the first function given to mlriscGen
463      as the entrypoint.  This address is currently always 0, but it will
464      not be 0 once we turn on block placement.
465    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
466      with entry points) from mlriscGen.
467    
468    ----------------------------------------------------------------------
469    Name: Allen Leung
470    Date: 2002/03/07 20:45:00 EST
471    Tag: leunga-20020307-x86-cmov
472    Description:
473    
474       Bug fixes for CMOVcc on x86.
475    
476       1. Added machine code generation for CMOVcc
477       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
478       3. CMOVcc cannot have an immediate operand as argument.
479    
480    ----------------------------------------------------------------------
481    Name: Matthias Blume
482    Date: 2002/03/07 16:15:00 EST
483    Tag: blume-20020307-controls
484    Description:
485    
486    This is a very large but mostly boring patch which makes (almost)
487    every tuneable compiler knob (i.e., pretty much everything under
488    Control.* plus a few other things) configurable via both the command
489    line and environment variables in the style CM did its configuration
490    until now.
491    
492    Try starting sml with '-h' (or, if you are brave, '-H')
493    
494    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
495    implements the underlying generic mechanism.
496    
497    The interface to some of the existing such facilities has changed somewhat.
498    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
499    (The getFoo interface is still there for backward-compatibility, but its
500    use is deprecated.)
501    
502    The ml-build script passes -Cxxx=yyy command-line arguments through so
503    that one can now twiddle the compiler settings when using this "batch"
504    compiler.
505    
506    TODO items:
507    
508    We should go through and throw out all controls that are no longer
509    connected to anything.  Moreover, we should go through and provide
510    meaningful (and correct!) documentation strings for those controls
511    that still are connected.
512    
513    Currently, multiple calls to Controls.new are accepted (only the first
514    has any effect).  Eventually we should make sure that every control
515    is being made (via Controls.new) exactly once.  Future access can then
516    be done using Controls.acc.
517    
518    Finally, it would probably be a good idea to use the getter-setter
519    interface to controls rather than ref cells.  For the time being, both
520    styles are provided by the Controls module, but getter-setter pairs are
521    better if thread-safety is of any concern because they can be wrapped.
522    
523    *****************************************
524    
525    One bug fix: The function blockPlacement in three of the MLRISC
526    backpatch files used to be hard-wired to one of two possibilities at
527    link time (according to the value of the placementFlag).  But (I
528    think) it should rather sense the flag every time.
529    
530    *****************************************
531    
532    Other assorted changes (by other people who did not supply a HISTORY entry):
533    
534    1. the cross-module inliner now works much better (Monnier)
535    2. representation of weights, frequencies, and probabilities in MLRISC
536       changed in preparation of using those for weighted block placement
537       (Reppy, George)
538    
539    ----------------------------------------------------------------------
540    Name: Lal George
541    Date: 2002/03/07 14:44:24 EST 2002
542    Tag: george-20020307-weighted-block-placement
543    
544    Tested the weighted block placement optimization on all architectures
545    (except the hppa) using AMPL to generate the block and edge frequencies.
546    Changes were required in the machine properties to correctly
547    categorize trap instructions. There is an MLRISC flag
548    "weighted-block-placement" that can be used to enable weighted block
549    placement, but this will be ineffective without block/edge
550    frequencies (coming soon).
551    
552    
553    ----------------------------------------------------------------------
554    Name: Lal George
555    Date: 2002/03/05 17:24:48 EST
556    Tag: george-20020305-linkage-cluster
557    
558    In order to support the block placement optimization, a new cluster
559    is generated as the very first cluster (called the linkage cluster).
560    It contains a single jump to the 'real' entry point for the compilation
561    unit. Block placement has no effect on the linkage cluster itself, but
562    all the other clusters  have full freedom in the manner in which they
563    reorder blocks or functions.
564    
565    On the x86 the typical linkage code that is generated is:
566       ----------------------
567            .align 2
568       L0:
569            addl    $L1-L0, 72(%esp)
570            jmp     L1
571    
572    
573            .align  2
574       L1:
575       ----------------------
576    
577    72(%esp) is the memory location for the stdlink register. This
578    must contain the address of the CPS function being called. In the
579    above example, it contains the address of  L0; before
580    calling L1 (the real entry point for the compilation unit), it
581    must contain the address for L1, and hence
582    
583            addl $L1-L0, 72(%esp)
584    
585    I have tested this on all architectures except the hppa.The increase
586    in code size is of course negligible
587    
588    ----------------------------------------------------------------------
589    Name: Allen Leung
590    Date: 2002/03/03 13:20:00 EST
591    Tag: leunga-20020303-mlrisc-tools
592    
593      Added #[ ... ] expressions to mlrisc tools
594    
595    ----------------------------------------------------------------------
596    Name: Matthias Blume
597    Date: 2002/02/27 12:29:00 EST
598    Tag: blume-20020227-cdebug
599    Description:
600    
601    - made types in structure C and C_Debug to be equal
602    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
603    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
604    
605    ----------------------------------------------------------------------
606    Name: Matthias Blume
607    Date: 2002/02/26 12:00:00 EST
608    Tag: blume-20020226-ffi
609    Description:
610    
611    1. Fixed a minor bug in CM's "noweb" tool:
612       If numbering is turned off, then truly don't number (i.e., do not
613       supply the -L option to noweb).  The previous behavior was to supply
614       -L'' -- which caused noweb to use the "default" line numbering scheme.
615       Thanks to Chris Richards for pointing this out (and supplying the fix).
616    
617    2. Once again, I reworked some aspects of the FFI:
618    
619       A. The incomplete/complete type business:
620    
621       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
622         gone!
623       - ML types representing an incomplete type are now *equal* to
624         ML types representing their corresponding complete types (just like
625         in C).  This is still safe because ml-nlffigen will not generate
626         RTTI for incomplete types, nor will it generate functions that
627         require access to such RTTI.   But when ML code generated from both
628         incomplete and complete versions of the C type meet, the ML types
629         are trivially interoperable.
630    
631         NOTE:  These changes restore the full generality of the translation
632         (which was previously lost when I eliminated functorization)!
633    
634       B. Enum types:
635    
636       - Structure C now has a type constructor "enum" that is similar to
637         how the "su" constructor works.  However, "enum" is not a phantom
638         type because each "T enum" has values (and is isomorphic to
639         MLRep.Signed.int).
640       - There are generic access operations for enum objects (using
641         MLRep.Signed.int).
642       - ml-nlffigen will generate a structure E_foo for each "enum foo".
643         * The structure contains the definition of type "mlrep" (the ML-side
644         representation type of the enum).  Normally, mlrep is the same
645         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
646         then mlrep will be defined as a datatype -- thus facilitating
647         pattern matching on mlrep values.
648         ("-ec" will be suppressed if there are duplicate values in an
649          enumeration.)
650         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
651         will be generated for each C enum constant xxx.
652         * Conversion functions m2i and i2m convert between mlrep and
653         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
654         * Coversion functions c and ml convert between mlrep and "tag enum".
655         * Access functions (get/set) fetch and store mlrep values.
656       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
657         enumerations are merged into one single enumeration represented by
658         structure E_'.
659    
660    ----------------------------------------------------------------------
661    Name: Allen Leung
662    Date: 2002/02/25 04:45:00 EST
663    Tag: leunga-20020225-cps-spill
664    
665    This is a new implementation of the CPS spill phase.
666    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
667    In case of problems, replace it with the old file spill.sml
668    
669    The current compiler runs into some serious performance problems when
670    constructing a large record.  This can happen when we try to compile a
671    structure with many items.  Even a very simple structure like the following
672    makes the compiler slow down.
673    
674        structure Foo = struct
675           val x_1 = 0w1 : Word32.int
676           val x_2 = 0w2 : Word32.int
677           val x_3 = 0w3 : Word32.int
678           ...
679           val x_N = 0wN : Word32.int
680        end
681    
682    The following table shows the compile time, from N=1000 to N=4000,
683    with the old compiler:
684    
685    N
686    1000   CPS 100 spill                           0.04u  0.00s  0.00g
687           MLRISC ra                               0.06u  0.00s  0.05g
688              (spills = 0 reloads = 0)
689           TOTAL                                   0.63u  0.07s  0.21g
690    
691    1100   CPS 100 spill                           8.25u  0.32s  0.64g
692           MLRISC ra                               5.68u  0.59s  3.93g
693              (spills = 0 reloads = 0)
694           TOTAL                                   14.71u  0.99s  4.81g
695    
696    1500   CPS 100 spill                           58.55u  2.34s  1.74g
697           MLRISC ra                               5.54u  0.65s  3.91g
698              (spills = 543 reloads = 1082)
699           TOTAL                                   65.40u  3.13s  6.00g
700    
701    2000   CPS 100 spill                           126.69u  4.84s  3.08g
702           MLRISC ra                               0.80u  0.10s  0.55g
703              (spills = 42 reloads = 84)
704           TOTAL                                   129.42u  5.10s  4.13g
705    
706    3000   CPS 100 spill                           675.59u  19.03s  11.64g
707           MLRISC ra                               2.69u  0.27s  1.38g
708              (spills = 62 reloads = 124)
709           TOTAL                                   682.48u  19.61s  13.99g
710    
711    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
712           MLRISC ra                               4.96u  0.27s  2.72g
713              (spills = 85 reloads = 170)
714           TOTAL                                   2375.26u  57.21s  48.00g
715    
716    As you can see the old cps spill module suffers from some serious
717    performance problem.  But since I cannot decipher the old code fully,
718    instead of patching the problems up, I'm reimplementing it
719    with a different algorithm.  The new code is more modular,
720    smaller when compiled, and substantially faster
721    (O(n log n) time and O(n) space).  Timing of the new spill module:
722    
723    4000  CPS 100 spill                           0.02u  0.00s  0.00g
724          MLRISC ra                               0.25u  0.02s  0.15g
725             (spills=1 reloads=3)
726          TOTAL                                   7.74u  0.34s  1.62g
727    
728    Implementation details:
729    
730    As far as I can tell, the purpose of the CPS spill module is to make sure the
731    number of live variables at any program point (the bandwidth)
732    does not exceed a certain limit, which is determined by the
733    size of the spill area.
734    
735    When the bandwidth is too large, we decrease the register pressure by
736    packing live variables into spill records.  How we achieve this is
737    completely different than what we did in the old code.
738    
739    First, there is something about the MLRiscGen code generator
740    that we should be aware of:
741    
742    o MLRiscGen performs code motion!
743    
744       In particular, it will move floating point computations and
745       address computations involving only the heap pointer to
746       their use sites (if there is only a single use).
747       What this means is that if we have a CPS record construction
748       statement
749    
750           RECORD(k,vl,w,e)
751    
752       we should never count the new record address w as live if w
753       has only one use (which is often the case).
754    
755       We should do something similar to floating point, but the transformation
756       there is much more complex, so I won't deal with that.
757    
758    Secondly, there are now two new cps primops at our disposal:
759    
760     1. rawrecord of record_kind option
761        This pure operator allocates some uninitialized storage from the heap.
762        There are two forms:
763    
764         rawrecord NONE [INT n]  allocates a tagless record of length n
765         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
766                                     and initializes the tag.
767    
768     2. rawupdate of cty
769          rawupdate cty (v,i,x)
770          Assigns to x to the ith component of record v.
771          The storelist is not updated.
772    
773    We use these new primops for both spilling and increment record construction.
774    
775     1. Spilling.
776    
777        This is implemented with a linear scan algorithm (but generalized
778        to trees).  The algorithm will create a single spill record at the
779        beginning of the cps function and use rawupdate to spill to it,
780        and SELECT or SELp to reload from it.  So both spills and reloads
781        are fine-grain operations.  In contrast, in the old algorithm
782        "spills" have to be bundled together in records.
783    
784        Ideally, we should sink the spill record construction to where
785        it is needed.  We can even split the spill record into multiple ones
786        at the places where they are needed.  But CPS is not a good
787        representation for global code motion, so I'll keep it simple and
788        am not attempting this.
789    
790     2. Incremental record construction (aka record splitting).
791    
792        Long records with many component values which are simulatenously live
793        (recall that single use record addresses are not considered to
794         be live) are constructed with rawrecord and rawupdate.
795        We allocate space on the heap with rawrecord first, then gradually
796        fill it in with rawupdate.  This is the technique suggested to me
797        by Matthias.
798    
799        Some restrictions on when this is applicable:
800        1. It is not a VECTOR record.  The code generator currently does not handle
801           this case. VECTOR record uses double indirection like arrays.
802        2. All the record component values are defined in the same "basic block"
803           as the record constructor.  This is to prevent speculative
804           record construction.
805    
806    ----------------------------------------------------------------------
807    Name: Allen Leung
808    Date: 2002/02/22 01:02:00 EST
809    Tag: leunga-20020222-mlrisc-tools
810    
811    Minor bug fixes in the parser and rewriter
812    
813    ----------------------------------------------------------------------
814    Name: Allen Leung
815    Date: 2002/02/21 20:20:00 EST
816    Tag: leunga-20020221-peephole
817    
818    Regenerated the peephole files.  Some contained typos in the specification
819    and some didn't compile because of pretty printing bugs in the old version
820    of 'nowhere'.
821    
822    ----------------------------------------------------------------------
823    Name: Allen Leung
824    Date: 2002/02/19 20:20:00 EST
825    Tag: leunga-20020219-mlrisc-tools
826    Description:
827    
828       Minor bug fixes to the mlrisc-tools library:
829    
830       1.  Fixed up parsing colon suffixed keywords
831       2.  Added the ability to shut the error messages up
832       3.  Reimplemented the pretty printer and fixed up/improved
833           the pretty printing of handle and -> types.
834       4.  Fixed up generation of literal symbols in the nowhere tool.
835       5.  Added some SML keywords to to sml.sty
836    
837    ----------------------------------------------------------------------
838    Name: Matthias Blume
839    Date: 2002/02/19 16:20:00 EST
840    Tag: blume-20020219-cmffi
841    Description:
842    
843    A wild mix of changes, some minor, some major:
844    
845    * All C FFI-related libraries are now anchored under $c:
846        $/c.cm      --> $c/c.cm
847        $/c-int.cm  --> $c/internals/c-int.cm
848        $/memory.cm --> $c/memory/memory.cm
849    
850    * "make" tool (in CM) now treats its argument pathname slightly
851      differently:
852        1. If the native expansion is an absolute name, then before invoking
853           the "make" command on it, CM will apply OS.Path.mkRelative
854           (with relativeTo = OS.FileSys.getDir()) to it.
855        2. The argument will be passed through to subsequent phases of CM
856           processing without "going native".  In particular, if the argument
857           was an anchored path, then "make" will not lose track of that anchor.
858    
859    * Compiler backends now "know" their respective C calling conventions
860      instead of having to be told about it by ml-nlffigen.  This relieves
861      ml-nlffigen from one of its burdens.
862    
863    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
864    
865    * Export C_DEBUG and C_Debug from $c/c.cm.
866    
867    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
868      subtyping relationship between incomplete pointers and their complete
869      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
870      with the convention of instantiating 'o with ('t, 'c) obj whenever
871      the pointer target type is complete.  In the incomplete case, 'o
872      will be instantiated with some "'c iobj" -- a type obtained by
873      using one of the functors PointerToIncompleteType or PointerToCompleteType.
874    
875      Operations that work on both incomplete and complete pointer types are
876      typed as taking an 'o ptr while operations that require the target to
877      be known are typed as taking some ('t, 'c) obj ptr.
878    
879      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
880      where void is an eqtype without any values.  This makes it possible
881      to work on voidptr values using functions meant to operate on light
882      incomplete pointers.
883    
884    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
885      been vastly simplified.
886    
887    ----------------------------------------------------------------------
888    Name: Matthias Blume
889    Date: 2002/02/19 10:48:00 EST
890    Tag: blume-20020219-pqfix
891    Description:
892    
893    Applied Chris Okasaki's bug fix for priority queues.
894    
895    ----------------------------------------------------------------------
896    Name: Matthias Blume
897    Date: 2002/02/15 17:05:00
898    Tag: Release_110_39
899    Description:
900    
901    Last-minute retagging is becoming a tradition... :-(
902    
903    This is the working release 110.39.
904    
905    ----------------------------------------------------------------------
906    Name: Matthias Blume
907    Date: 2002/02/15 16:00:00 EST
908    Tag: Release_110_39-orig
909    Description:
910    
911    Working release 110.39.  New bootfiles.
912    
913    (Update: There was a small bug in the installer so it wouldn't work
914    with all shells.  So I retagged. -Matthias)
915    
916    ----------------------------------------------------------------------
917    Name: Matthias Blume
918    Date: 2002/02/15 14:17:00 EST
919    Tag: blume-20020215-showbindings
920    Description:
921    
922    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
923    the latter can be useful for exploring what bindings are available at
924    the interactive prompt.  (The first function returns only the list
925    of symbols that are really bound, the second prints those but also the
926    ones that CM's autoloading mechanism knows about.)
927    
928    ----------------------------------------------------------------------
929    Name: Matthias Blume
930    Date: 2002/02/15 12:08:00 EST
931    Tag: blume-20020215-iptrs
932    Description:
933    
934    Two improvements to ml-nlffigen:
935    
936      1. Write files only if they do not exist or if their current contents
937         do not coincide with what's being written.  (That is, avoid messing
938         with the time stamps unless absolutely necessary.)
939    
940      2. Implement a "repository" mechanism for generated files related
941         to "incomplete pointer types".   See the README file for details.
942    
943    ----------------------------------------------------------------------
944    Name: Matthias Blume
945    Date: 2002/02/14 11:50:00 EST
946    Tag: blume-20020214-quote
947    Description:
948    
949    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
950    because of the new and improved tag generation scheme.  (Thanks to Allen
951    Leung for pointing it out.)
952    
953    ----------------------------------------------------------------------
954    Name: Lal George
955    Date: 2002/02/14 09:55:27 EST 2002
956    Tag: george-20020214-isabelle-bug
957    Description:
958    
959    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
960    of Isabelle on the x86.
961    
962    From Allen:
963    -----------
964     I've found the problem:
965    
966         in ra-core.sml, I use the counter "blocked" to keep track of the
967         true number of elements in the freeze queue.  When the counter goes
968         to zero, I skip examining the queue.  But I've messed up the
969         bookkeeping in combine():
970    
971             else ();
972             case !ucol of
973               PSEUDO => (if !cntv > 0 then
974                     (if !cntu > 0 then blocked := !blocked - 1 else ();
975                                        ^^^^^^^^^^^^^^^^^^^^^^^
976                      moveu := mergeMoveList(!movev, !moveu)
977                     )
978                  else ();
979    
980         combine() is called to coalesce two nodes u and v.
981         I think I was thinking that if the move counts of u and v are both
982         greater than zero then after they are coalesced then one node is
983         removed from the freeze queue.  Apparently I was thinking that
984         both u and v are of low degree, but that's clearly not necessarily true.
985    
986    
987    02/12/2002:
988        Here's the patch.  HOL now compiles.
989    
990        I don't know how this impact on performance (compile
991        time or runtime).  This bug caused the RA (especially on the x86)
992        to go thru the potential spill phase when there are still nodes on the
993        freeze queue.
994    
995    
996    
997    
998    ----------------------------------------------------------------------
999    Name: Matthias Blume
1000    Date: 2002/02/13 22:40:00 EST
1001    Tag: blume-20020213-fptr-rtti
1002    Description:
1003    
1004    Fixed a bug in ml-nlffigen that was introduced with one of the previous
1005    updates.
1006    
1007    ----------------------------------------------------------------------
1008    Name: Matthias Blume
1009    Date: 2002/02/13 16:41:00 EST
1010    Tag: blume-20020213-cmlpq
1011    Description:
1012    
1013    Added new priority queue export symbols (which have just been added to
1014    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
1015    not compile and the installer would choke.)
1016    
1017    ----------------------------------------------------------------------
1018    Name: Matthias Blume
1019    Date: 2002/02/13 16:15:00 EST
1020    Tag: blume-20020213-various
1021    Description:
1022    
1023    1. More tweaks to ml-nlffigen:
1024    
1025       - better internal datastructures (resulting in slight speedup)
1026       - "-match" option requires exact match
1027       - "localized" gensym counters (untagged structs/unions nested within
1028         other structs/unions or within typedefs get a fresh counter; their
1029         tag will be prefixed by a concatenation of their parents' tags)
1030       - bug fixes (related to calculation of transitive closure of types
1031         to be included in the output)
1032    
1033    2. Minor Basis updates:
1034    
1035       - added implementations for List.collate and Option.app
1036    
1037    ----------------------------------------------------------------------
1038    Name: Matthias Blume
1039    Date: 2002/02/11 15:55:00 EST
1040    Tag: blume-20020211-gensym
1041    Description:
1042    
1043    Added a "-gensym" option to command line of ml-nlffigen.  This can be
1044    used to specify a "stem" -- a string that is inserted in all "gensym'd"
1045    names (ML structure names that correspond to unnamed C structs, unions,
1046    and enums), so that separate runs of ml-nlffigen do not clash.
1047    
1048    ----------------------------------------------------------------------
1049    Name: Matthias Blume
1050    Date: 2002/02/11 12:05:00 EST
1051    Tag: blume-20020211-gensml
1052    Description:
1053    
1054    A quick fix for a problem with GenSML (in the pgraph-util library):
1055    Make generation of toplevel "local" optional.  (Strictly speaking,
1056    signature definitions within "local" are not legal SML.)
1057    
1058    Other than that: updates to INSTALL and cm/TODO.
1059    
1060    ----------------------------------------------------------------------
1061    Name: Matthias Blume
1062    Date: 2002/02/08 15:00:00 EST
1063    Tag: blume-20020208-uniquepid
1064    Description:
1065    
1066    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
1067    
1068    1. The installer (config/install.sh) has gotten smarter:
1069    
1070         - Configuration options are a bit easier to specify now
1071           (in config/targets).
1072         - Bug in recognizing .tar.bz2 files fixed.
1073         - Installer automatically resolves dependencies between
1074           configuration options (e.g., if you ask for eXene, you will
1075           also get cml -- regardless whether you asked for it or not).
1076         - Installer can run in "quieter mode" by setting the environment
1077           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
1078           "completely silent", though.
1079         - Build HashCons library as part of smlnj-lib.
1080    
1081    2. A new scheme for assigning persistent identifiers to compilation
1082       units (and, by extension, to types etc.) has been put into place.
1083       This fixes a long-standing bug where types and even dynamic values
1084       can get internally confused, thereby compromising type safety
1085       (abstraction) and dynamic correctness.  See
1086    
1087         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
1088    
1089       for an example of how things could go wrong until now.
1090    
1091       The downside of the new scheme is that pids are not quite as
1092       persistent as they used to be: CM will generate a fresh pid
1093       for every compilation unit that it thinks it sees for the first
1094       time.  That means that if you compile starting from a clean, fresh
1095       source tree at two different times, you end up with different
1096       binaries.
1097    
1098       Cutoff recompilation, however, has not been compromised because
1099       CM keeps pid information in special caches between runs.
1100    
1101    ----------------------------------------------------------------------
1102    Name: Lal George
1103    Date: 2002/02/07 15:34:13 EST 2002
1104    Tag: <none>
1105    Description:
1106    
1107    Compilers that generate assembly code may produce  global labels
1108    whose value is resolved at link time. The various peephole optimization
1109    modules did not take this in account.
1110    
1111    TODO. The Labels.addrOf function should really return an option
1112    type so that clients are forced to deal with this issue, rather
1113    than an exception being raised.
1114    
1115    ----------------------------------------------------------------------
1116    Name: Lal George
1117    Date: 2002/02/06 13:55:02 EST
1118    Tag: george-20020206-ra-breakup
1119    Description:
1120    
1121    1. A bug fix from Allen.
1122    
1123        A typo causes extra fstp %st(0)'s to be generated at compensation
1124        edges, which might cause stack underflow traps at runtime.  This
1125        occurs in fft where there are extraneous fstps right before the 'into'
1126        trap instruction (in this case they are harmless since none of the
1127        integers overflow.)
1128    
1129    2. Pulled out various utility modules that were embedded in the modules
1130       of the register allocator. I need these modules for other purposes, but
1131       they are not complete enough to put into a library (just yet).
1132    ----------------------------------------------------------------------
1133    Name: Matthias Blume
1134    Date: 2002/01/31 16:05:00 EST
1135    Tag: blume-20020131-sparc-ccalls
1136    Description:
1137    
1138    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
1139       of extra stack space by mistake.  Fixed.
1140    
1141    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
1142    
1143    ----------------------------------------------------------------------
1144    Name: Allen Leung
1145    Date: 2002/01/30
1146    Tag: leunga-20020130-nowhere-bug-fix
1147    Description:
1148    
1149       MLRISC bug fixes:
1150       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
1151       2. MachineInt.fromString was negating its value.
1152    
1153    ----------------------------------------------------------------------
1154    Name: Matthias Blume
1155    Date: 2002/01/29
1156    Tag: blume-20020129-INSTALL
1157    Description:
1158    
1159    - Added somewhat detailed installation instructions (file INSTALL).
1160    - Fixed curl-detection bug in config/install.sh.
1161    - It is now possible to select the URL getter using the URLGETTER
1162      environment variable:
1163    
1164          not set / "unknown"      --> automatic detection (script tries wget,
1165                                       curl, and lynx)
1166          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
1167                                       how to properly invoke them)
1168          other                    --> use $URLGETTER directly, it must take
1169                                       precisely two command-line arguments
1170                                       (source URL and destination file name)
1171    
1172    ----------------------------------------------------------------------
1173    Name: Matthias Blume
1174    Date: 2002/01/28
1175    Tag: blume-20020128-sparc-ccalls
1176    Description:
1177    
1178    - Fixed problem with calculation of "used" registers in sparc-c-calls.
1179    - Make use of the allocParam argument in sparc-c-calls.
1180    
1181    ----------------------------------------------------------------------
1182    Name: Matthias Blume
1183    Date: 2002/01/28
1184    Tag: blume-20020128-allocParam
1185    Description:
1186    
1187    John Reppy:  Changes c-calls API to accept client-callback for
1188    allocating extra stack space.
1189    me: Corresponding changes to mlriscGen (using a dummy argument that
1190        does not change the current behavior).
1191    
1192    ----------------------------------------------------------------------
1193    Name: Matthias Blume
1194    Date: 2002/01/28 12:00:00
1195    Tag: Release_110_38
1196    Description:
1197    
1198    This time for real!!!
1199    
1200    ----------------------------------------------------------------------
1201    Name: Matthias Blume
1202    Date: 2002/01/28 10:56:00 EST
1203    Tag: blume-20020128-retraction
1204    Description:
1205    
1206    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
1207       with blume-Release_110_38-retracted.)
1208    
1209    1. Fixed a problem with incorrect rounding modes in real64.sml.
1210       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
1211    
1212    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
1213       and enums fixed.  The naming of corresponding ML identifiers should
1214       now be consistent again.
1215    
1216    ----------------------------------------------------------------------
1217    Name: Allen Leung
1218    Date: 2002/01/27
1219    Tag: leunga-20020127-nowhere
1220    Description:
1221    
1222       Added a target called nowhere in the configuration scripts.
1223       Enabling this will build the MLRISC 'nowhere' tool (for translating
1224       programs with where-clauses into legal SML code) during installation.
1225    
1226    ----------------------------------------------------------------------
1227    Name: Matthias Blume
1228    Date: 2002/01/25 21:27:00 EST
1229    Tag: blume-Release_110_38-retracted
1230    Description:
1231    
1232    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
1233    
1234    README will be added later.
1235    
1236    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
1237    (Release_110_38) removed.  Reason: Last-minute bug fixes.
1238    
1239    ----------------------------------------------------------------------
1240    Name: Matthias Blume
1241    Date: 2002/01/25
1242    Tag: blume-20020125-ffi
1243    Description:
1244    
1245    A large number of tweaks and improvements to ml-nlffi-lib and
1246    ml-nlffigen:
1247    
1248       - ML represenation types have been streamlined
1249       - getter and setter functions work with concrete values, not abstract
1250         ones where possible
1251       - ml-nlffigen command line more flexible (see README file there)
1252       - some bugs have been fixed (hopefully)
1253    
1254    ----------------------------------------------------------------------
1255    Name: Lal George
1256    Date: 2002/01/24
1257    Tag: george-20020124-risc-ra-interface
1258    Description:
1259    
1260       There is a dramatic simplification in the interface to the
1261       register allocator for RISC architectures as a result of making
1262       parallel copy instructions explicit.
1263    
1264    ----------------------------------------------------------------------
1265    Name: Matthias Blume
1266    Date: 2002/01/22
1267    Tag: blume-20020122-x86-ccalls
1268    Description:
1269    
1270    Bug fix for c-calls on x86 (having to do with how char- and
1271    short-arguments are being handled).
1272    
1273    ----------------------------------------------------------------------
1274    Name: Matthias Blume
1275    Date: 2002/01/21
1276    Tag: blume-20020121-ff
1277    Description:
1278    
1279    Another day of fiddling with the FFI...
1280    
1281    1. Bug fix/workaround:  CKIT does not complain about negative array
1282       dimensions, so ml-nlffigen has to guard itself against this possibility.
1283       (Otherwise a negative dimension would send it into an infinite loop.)
1284    
1285    2. Some of the abstract types (light objects, light pointers, most "base"
1286       types) in structure C are now eqtypes.
1287    
1288    3. Added constructors and test functions for NULL function pointers.
1289    
1290    ----------------------------------------------------------------------
1291    Name: Matthias Blume
1292    Date: 2002/01/18
1293    Tag: blume-20020118-ready-for-new-release
1294    Description:
1295    
1296    Made config/srcarchiveurl point to a new place.  (Will provide boot
1297    files shortly.)
1298    
1299    Maybe we christen this to be 110.38?
1300    
1301    ----------------------------------------------------------------------
1302    Name: Matthias Blume
1303    Date: 2002/01/18
1304    Tag: blume-20020118-more-ffifiddle
1305    Description:
1306    
1307    Today's FFI fiddling:
1308    
1309      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
1310        C.Set.* that use concrete (MLRep.*) arguments and results instead
1311        of abstract ones.
1312    
1313      - Provided word-style bit operations etc. for "int" representation
1314        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
1315        and Long.
1316    
1317    ----------------------------------------------------------------------
1318    Name: Matthias Blume
1319    Date: 2002/01/18
1320    Tag: blume-20020118-use-x86-fp
1321    Description:
1322    
1323    Now that x86-fast-fp seems to be working, I turned it back on again
1324    by default.  (Seems to work fine now, even with the FFI.)
1325    
1326    Other than that, I added some documentation about the FFI to
1327    src/ml-nlffigen/README and updated the FFI test examples in
1328    src/ml-nlffi-lib/Tests/*.
1329    
1330    ----------------------------------------------------------------------
1331    Name: Allen Leung
1332    Date: 2002/01/17
1333    Tag: leunga-20020117-x86-fast-fp-call
1334    Description:
1335    
1336       1. Fixed a problem with handling return fp values when x86's fast fp
1337          mode is turned on.
1338    
1339       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
1340          of %f32.
1341    
1342       3. Added a constructor INT32lit to the ast of MLRISC tools.
1343    
1344    ----------------------------------------------------------------------
1345    Name: Matthias Blume
1346    Date: 2002/01/16
1347    Tag: blume-20020116-ffifiddle
1348    Description:
1349    
1350    More fiddling with the FFI interface:
1351    
1352     - Make constness 'c instead of rw wherever possible.  This eliminates
1353       the need for certain explicit coercions.  (However, due to ML's
1354       value polymorphism, there will still be many cases where explicit
1355       coercions are necessary.  Phantom types are not the whole answer
1356       to modeling a subtyping relationship in ML.)
1357    
1358     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
1359    
1360     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
1361       clumsy workaround for varargs problem.)
1362    
1363    ----------------------------------------------------------------------
1364    Name: Lal George
1365    Date: 2002/01/15
1366    Tag: <none>
1367    Description:
1368    
1369    1. Since COPY instructions are no longer native to the architecture,
1370       a generic functor can be used to implement the expandCopies function.
1371    
1372    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
1373       TEXT segment.
1374    
1375    ----------------------------------------------------------------------
1376    Name: Matthias Blume
1377    Date: 2002/01/15
1378    Tag: blume-20020115-ffiupdates
1379    Description:
1380    
1381    1. Fix for bug resulting in single-precision float values being returned
1382       incorrectly from FFI calls.
1383    
1384    2. Small modifications to C FFI API:
1385    
1386        - memory-allocation routines return straight objects (no options)
1387          and raise an exception in out-of-memory situations
1388        - unsafe extensions to cast between function pointers and pointers
1389          from/to ints
1390        - added structure C_Debug as an alternative to structure C where
1391          pointer-dereferencing (|*| and |*!) always check for null-pointers
1392        - added open_lib' to DynLinkage;  open_lib' works like open_lib
1393          but also takes a (possibly empty) list of existing library handles
1394          that the current library depends on
1395    
1396    ----------------------------------------------------------------------
1397    Name: Matthias Blume
1398    Date: 2002/01/10
1399    Tag: blume-20020110-newffigen
1400    Description:
1401    
1402    1. Updates to portable graph code.
1403    
1404    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
1405       more scalable now so that even huge interfaces such as the one
1406       for GTK compile in finite time and space. :-)
1407       See src/ml-nlffigen/README for details on what's new.
1408    
1409    ----------------------------------------------------------------------
1410    Name: Lal George
1411    Date: 2001/01/09 14:31:35 EST 2002
1412    Tag: george-20011206-rm-native-copy
1413    Description:
1414    
1415            Removed the native COPY and FCOPY instructions
1416            from all the architectures and replaced it with the
1417            explicit COPY instruction from the previous commit.
1418    
1419            It is now possible to simplify many of the optimizations
1420            modules that manipulate copies. This has not been
1421            done in this change.
1422    
1423    ----------------------------------------------------------------------
1424    Name: Lal George
1425    Date: 2001/12/06 16:50:13 EST 2001
1426    Tag: george-20011206-mlrisc-instruction
1427    Description:
1428    
1429    Changed the representation of instructions from being fully abstract
1430    to being partially concrete. That is to say:
1431    
1432      from
1433            type instruction
1434    
1435      to
1436            type instr                              (* machine instruction *)
1437    
1438            datatype instruction =
1439                LIVE of {regs: C.cellset, spilled: C.cellset}
1440              | KILL of {regs: C.cellset, spilled: C.cellset}
1441              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
1442              | ANNOTATION of {i: instruction, a: Annotations.annotation}
1443              | INSTR of instr
1444    
1445    This makes the handling of certain special instructions that appear on
1446    all architectures easier and uniform.
1447    
1448    LIVE and KILL say that a list of registers are live or killed at the
1449    program point where they appear. No spill code is generated when an
1450    element of the 'regs' field is spilled, but the register is moved to
1451    the 'spilled' (which is present, more for debugging than anything else).
1452    
1453    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
1454    We used to generate:
1455    
1456            DEFFREG f1
1457            f1 := f2 + f3
1458            trapb
1459    
1460    but now generate:
1461    
1462            f1 := f2 + f3
1463            trapb
1464            LIVE {regs=[f1,f2,f3], spilled=[]}
1465    
1466    Furthermore, the DEFFREG (hack) required that all floating point instruction
1467    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
1468    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
1469    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
1470    this architecture).
1471    
1472    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
1473    all the architectures. This will result in further simplification of the
1474    register allocator that must be aware of them for coalescing purposes, and
1475    will also simplify certain aspects of the machine description that provides
1476    callbacks related to parallel copies.
1477    
1478    ANNOTATION should be obvious, and now INSTR represents the honest to God
1479    machine instruction set!
1480    
1481    The <arch>/instructions/<arch>Instr.sml files define certain utility
1482    functions for making porting easier -- essentially converting upper case
1483    to lower case. All machine instructions (of type instr) are in upper case,
1484    and the lower case form generates an MLRISC instruction. For example on
1485    the alpha we have:
1486    
1487      datatype instr =
1488         LDA of {r:cell, b:cell, d:operand}
1489       | ...
1490    
1491      val lda : {r:cell, b:cell, d:operand} -> instruction
1492        ...
1493    
1494    where lda is just (INSTR o LDA), etc.
1495    
1496    ----------------------------------------------------------------------
1497    Name: Matthias Blume
1498    Date: 2001/11/22 21:40:00 EST
1499    Tag: Release_110_37
1500    Description:
1501    
1502    Release 110.37.  This time for real.
1503    
1504    ----------------------------------------------------------------------
1505    Name: Matthias Blume
1506    Date: 2001/11/21 16:35:00 EST
1507    Tag: blume-20011121-foot-in-mouth
1508    Description:
1509    
1510    Removed the "Release_110_37" tag because of a serious bug.
1511    This will be re-tagged once the bug is fixed.
1512    
1513    ----------------------------------------------------------------------
1514    Name: Matthias Blume
1515    Date: 2001/11/21 16:14:00 EST
1516    Tag: blume-20011121-forgottenfile
1517    Description:
1518    
1519    Forgot to add a file.  (Just a .tex-file -- part of
1520    the CM manual source.)
1521    
1522    ----------------------------------------------------------------------
1523    Name: Matthias Blume
1524    Date: 2001/11/21 16:10:00 EST
1525    Tag: blume-20011121-invalid_110_37
1526    Description:
1527    
1528    Note: I removed the original tag "Release_110_37" from this commit
1529          because we found a serious bug in all non-x86 backends.
1530          - Matthias
1531    
1532    1. Modifications to the SML/NJ code generator and to the runtime system
1533       so that code object name strings are directly inserted into code
1534       objects at code generation time.  The only business the runtime system
1535       has with this is now to read the name strings on occasions.
1536       (The encoding of the name string has also changed somewhat.)
1537    
1538    2. CM now implements a simple "set calculus" for specifying export lists.
1539       In particular, it is now possible to refer to the export lists of
1540       other libraries/groups/sources and form unions as well as differences.
1541       See the latest CM manual for details.
1542    
1543    3. An separate notion of "proxy" libraries has again be eliminated from
1544       CM's model.  (Proxy libraries are now simply a special case of using
1545       the export list calculus.)
1546    
1547    4. Some of the existing libraries now take advantage of the new set
1548       calculus.
1549       (Notice that not all libraries have been converted because some
1550       of the existing .cm-files are supposed to be backward compatible
1551       with 110.0.x.)
1552    
1553    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
1554       "General.exnMessage"!  The former relies on a certain hook to be
1555       initialized, and that often does not happen in the stand-alone case.)
1556    
1557    ----------------------------------------------------------------------
1558    Name: Lal George
1559    Date: 2001/11/21  13:56:18 EST
1560    Tag: george-2001121-pseudo-ops
1561    Description:
1562    
1563      Implemented a complete redesign of MLRISC pseudo-ops. Now there
1564      ought to never be any question of incompatabilities with
1565      pseudo-op syntax expected by host assemblers.
1566    
1567      For now, only modules supporting GAS syntax are implemented
1568      but more should follow, such as MASM, and vendor assembler
1569      syntax, e.g. IBM as, Sun as, etc.
1570    
1571    ----------------------------------------------------------------------
1572    Name: Matthias Blume
1573    Date: 2001/11/14 11:52:00 EST
1574    Tag: blume-20011114-srcname
1575    Description:
1576    
1577    1. Routed the name of the current source file to mlriscgen where it
1578       should be directly emitted into the code object.  (This last part
1579       is yet to be done.)
1580    
1581    2. Some cleanup of the pgraph code to make it match the proposal that
1582       I put out the other day.  (The proposal notwithstanding, things are
1583       still in flux here.)
1584    
1585    ----------------------------------------------------------------------
1586    Name: Lal George
1587    Date: 2001/11/14 09:44:04 EST
1588    Tag:
1589    Description:
1590    
1591      Fix for a backpatching bug reported by Allen.
1592    
1593      Because the boundary between short and long span-dependent
1594      instructions is +/- 128, there are an astounding number of
1595      span-dependent instructions whose size is over estimated.
1596    
1597      Allen came up with the idea of letting the size of span
1598      dependent instructions be non-monotonic, for a maxIter
1599      number of times, after which the size must be monotonically
1600      increasing.
1601    
1602      This table shows the number of span-dependent instructions
1603      whose size was over-estimated as a function of maxIter, for the
1604      file Parse/parse/ml.grm.sml:
1605    
1606         maxIter            # of instructions:
1607            10                      687
1608            20                      438
1609            30                      198
1610            40                        0
1611    
1612      In compiling the compiler, there is no significant difference in
1613      compilation speed between maxIter=10 and maxIter=40. Actually,
1614      my measurements showed that maxIter=40 was a tad faster than
1615      maxIter=10! Also 96% of the  files in the compiler reach a fix
1616      point within 13 iterations, so fixing maxIter at 40, while high,
1617      is okay.
1618    
1619    ----------------------------------------------------------------------
1620    Name: Matthias Blume
1621    Date: 2001/10/31 15:25:00 EST
1622    Tag: blume-20011031-pgraph
1623    Description:
1624    
1625    CKIT:
1626    * Changed the "Function" constructor of type Ast.ctype to carry optional
1627      argument identifiers.
1628    * Changed the return type of TypeUtil.getFunction accordingly.
1629    * Type equality ignores the argument names.
1630    * TypeUtil.composite tries to preserve argument names but gives up quickly
1631      if there is a mismatch.
1632    
1633    installation script:
1634    * attempts to use "curl" if available (unless "wget" is available as well)
1635    
1636    CM:
1637    * has an experimental implementation of "portable graphs" which I will
1638      soon propose as an implementation-independent library format
1639    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
1640    
1641    NLFFI-LIB:
1642    * some cleanup (all cosmetic)
1643    
1644    NLFFIGEN:
1645    * temporarily disabled the mechanism that suppresses ML output for
1646      C definitions whose identifiers start with an underscore character
1647    * generate val bindings for enum constants
1648    * user can request that only one style (light or heavy) is being used;
1649      default is to use both (command-line arguments: -heavy and -light)
1650    * fixed bug in handling of function types involving incomplete pointers
1651    * generate ML entry points that take record arguments (i.e., using
1652      named arguments) for C functions that have a prototype with named
1653      arguments
1654      (see changes to CKIT)
1655    
1656    ----------------------------------------------------------------------
1657    Name: Allen Leung
1658    Date: 2001/10/27 20:34:00 EDT
1659    Tag: leunga-20011027-x86-fast-fp-call
1660    Description:
1661    
1662       Fixed the bug described in blume-20010920-slowfp.
1663    
1664       The fix involves
1665          1. generating FCOPYs in FSTP in ia32-svid
1666          2. marking a CALL with the appropriate annotation
1667    
1668    ----------------------------------------------------------------------
1669    Name: Matthias Blume
1670    Date: 2001/10/16 11:32:00 EDT
1671    Tag: blume-20011016-netbsd
1672    Description:
1673    
1674    Underscore patch from Chris Richards (fixing problem with compiling
1675    runtime system under recent NetBSD).
1676    
1677    ----------------------------------------------------------------------
1678    Name: Allen Leung
1679    Date: 2001/10/12 17:18:32 EDT 2001
1680    Tag: leung-20011012-x86-printflowgraph
1681    Description:
1682    
1683    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
1684    
1685    ----------------------------------------------------------------------
1686    Name: Lal George
1687    Date: 2001/10/11 23:51:34 EDT
1688    Tag: george-20011011-too-many-instrs
1689    Description:
1690    
1691    The representation of a program point never expected to see more
1692    than 65536 instructions in a basic block!
1693    
1694    ----------------------------------------------------------------------
1695    Name: Lal George
1696    Date: 2001/10/09 09:41:37 EDT
1697    Tag: george-20011008-mlrisc-labels
1698    Description:
1699    
1700    Changed the machine description files to support printing of
1701    local and global labels in assembly code, based on host assembler
1702    conventions.
1703    
1704    ----------------------------------------------------------------------
1705    Name: Matthias Blume
1706    Date: 2001/09/25 15:25:00 EDT
1707    Tag: blume-20010925-exninfo
1708    Description:
1709    
1710    I provided a non-hook implementation of exnName (at the toplevel) and
1711    made the "dummy" implementation of exnMessage (at the toplevel) more
1712    useful: if nothing gets "hooked in", then at least you are going to
1713    see the exception name and a message indicating why you don't see more.
1714    
1715    [For the time being, programs that need exnMessage and want to use
1716    ml-build should either use General.exnMessage (strongly recommended) or
1717    refer to structure General at some other point so that CM sees a
1718    static dependency.]
1719    
1720    [Similar remarks go for "print" and "use":  If you want to use their
1721    functionality in stand-alone programs generated by ml-build, then use
1722    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
1723    
1724    ----------------------------------------------------------------------
1725    Name: Matthias Blume
1726    Date: 2001/09/20 17:28:00 EDT
1727    Tag: blume-20010920-slowfp
1728    Description:
1729    
1730    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
1731    
1732    ----------------------------------------------------------------------
1733    Name: Matthias Blume
1734    Date: 2001/09/20 17:20:00 EDT
1735    Tag: blume-20010920-canonicalpaths
1736    Description:
1737    
1738    0. Updated the BOOT file (something that I forgot to do earlier).
1739    
1740    1. Small internal change to CM so that it avoids "/../" in filenames
1741       as much as possible (but only where it is safe).
1742    
1743    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
1744       that arguments that contain delimiters are passed through correctly.
1745       This change also means that all "special" arguments of the form
1746       @SMLxxx... must come first.
1747    
1748    3. Changed install script to put relative anchor names for tool commands
1749       into pathconfig.
1750    
1751    ----------------------------------------------------------------------
1752    Name: Matthias Blume
1753    Date: 2001/09/18 15:35:00 EDT
1754    Tag: blume-20010918-readme11036
1755    Description:
1756    
1757    Added README files.
1758    
1759    ----------------------------------------------------------------------
1760    Name: Matthias Blume
1761    Date: 2001/09/18 11:45:00 EDT
1762    Tag: Release_110_36 (retag)
1763    Description:
1764    
1765    Fixed mistake in config/preloads. Retagged as 110.36.
1766    
1767    ----------------------------------------------------------------------
1768    Name: Matthias Blume
1769    Date: 2001/09/18 09:40:00 EDT
1770    Tag: Release_110_36_orig (tag changed)
1771    Description:
1772    
1773    New version (110.36).  New bootfiles.
1774    
1775    ----------------------------------------------------------------------
1776    Name: Matthias Blume
1777    Date: 2001/09/14 16:15:00 EDT
1778    Tag: blume-20010914-x86fastfp
1779    Description:
1780    
1781    John committed some changes that Allen made, in particular a (hopefully)
1782    correctly working version of the x86-fp module.
1783    
1784    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
1785    flag to "true".  Everything seems to compile to a fixpoint ok, and
1786    "mandelbrot" speeds up by about 15%.
1787    
1788    ----------------------------------------------------------------------
1789    Name: Matthias Blume
1790    Date: 2001/09/13 11:20:00 EDT
1791    Tag: blume-20010913-minimal
1792    Description:
1793    
1794    1. Stefan Monnier's patch to fix a miscompilation problem that
1795       was brought to light by John Reppy's work on Moby.
1796    
1797    2. Implemented a minimal "structure Compiler" that contains just
1798       "version" and "architecture".  The minimal version will be
1799       available when the full version is not.  This is for backward-
1800       compatibility with code that wants to test Compiler.version.
1801    
1802    ----------------------------------------------------------------------
1803    Name: Matthias Blume
1804    Date: 2001/08/28 14:03:00 EDT
1805    Tag: blume-20010828-ml-lex
1806    Description:
1807    
1808    Fix for bug 1581, received from Neophytos Michael.
1809    
1810    ----------------------------------------------------------------------
1811    Name: Matthias Blume
1812    Date: 2001/08/27 11:20:00 EDT
1813    Tag: blume-20010827-readme11035
1814    Description:
1815    
1816    Fleshed out the README file for 110.35.
1817    
1818    ----------------------------------------------------------------------
1819    Name: Matthias Blume
1820    Date: 2001/08/24 17:10:00 EDT
1821    Tag: Release_110_35
1822    Description:
1823    
1824    New version number (110.35).  New bootfiles.
1825    
1826    ----------------------------------------------------------------------
1827    Name: Lal George
1828    Date: 2001/08/24 13:47:18 EDT 2001
1829    Tag: george-20010824-MLRISC-graphs
1830    Description:
1831    
1832     removed clusters from MLRISC completely and replaced with graphs.
1833    
1834    ----------------------------------------------------------------------
1835    Name: Matthias Blume
1836    Date: 2001/08/23 17:50:00 EDT
1837    Tag: blume-20010823-toplevel
1838    Description:
1839    
1840    - some reorganization of the code that implements various kinds of
1841      environments in the compiler (static, dynamic, symbolic, combined)
1842    - re-implemented the EnvRef module so that evalStream works properly
1843      (if the stream contains references to "use", "CM.make", etc.)
1844    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
1845    
1846    ----------------------------------------------------------------------
1847    Name: Matthias Blume
1848    Date: 2001/08/20 15:50 EDT
1849    Tag: blume20010820-slipup
1850    Description:
1851    
1852    I forgot to commit a few files.  Here they are...
1853    
1854    ----------------------------------------------------------------------
1855    Name: Matthias Blume
1856    Date: 2001/08/20 15:35:00 EDT
1857    Tag: blume-20010820-debugprof
1858    Description:
1859    
1860    !!!! NEW BOOTFILES !!!!
1861    
1862    This is another round of reorganizing the compiler sources.  This
1863    time the main goal was to factor out all the "instrumentation"
1864    passes (for profiling and backtracing) into their own library.
1865    The difficulty was to do it in such a way that it does not depend
1866    on elaborate.cm but only on elabdata.cm.
1867    
1868    Therefore there have been further changes to both elaborate.cm and
1869    elabdata.cm -- more "generic" things have been moved from the former
1870    to the latter.  As a result, I was forced to split the assignment
1871    of numbers indicating "primtyc"s into two portions: SML-generic and
1872    SML/NJ-specific.  Since it would have been awkward to maintain,
1873    I bit the bullet and actually _changed_ the mapping between these
1874    numbers and primtycs.  The bottom line of this is that you need
1875    a new set of bin- and bootfiles.
1876    
1877    I have built new bootfiles for all architectures, so doing a fresh
1878    checkout and config/install.sh should be all you need.
1879    
1880    The newly created library's name is
1881    
1882        $smlnj/viscomp/debugprof.cm
1883    
1884    and its sources live under
1885    
1886        src/compiler/DebugProf
1887    
1888    ----------------------------------------------------------------------
1889    Name: Matthias Blume
1890    Date: 2001/08/15 17:15:00 EDT
1891    Tag: blume-20010815-compreorg
1892    Description:
1893    
1894    This is a first cut at reorganizing the CM libraries that make up the
1895    core of the compiler.  The idea is to separate out pieces that could
1896    be used independently by tools, e.g., the parser, the typechecker, etc.
1897    
1898    The current status is a step in this direction, but it is not quite
1899    satisfactory yet.  Expect more changes in the future.
1900    
1901    Here is the current (new) organization...
1902    
1903        What used to be $smlnj/viscomp/core.cm is now divided into
1904        six CM libraries:
1905    
1906             $smlnj/viscomp/basics.cm
1907                           /parser.cm
1908                           /elabdata.cm
1909                           /elaborate.cm
1910                           /execute.cm
1911                           /core.cm
1912    
1913        The CM files for these libraries live under src/system/smlnj/viscomp.
1914        All these libraries are proxy libraries that contain precisely
1915        one CM library component.  Here are the locations of the components
1916        (all within the src/compiler tree):
1917    
1918             Basics/basics.cm
1919             Parse/parser.cm
1920             ElabData/elabdata.cm
1921             Elaborator/elaborate.cm
1922             Execution/execute.cm
1923             core.cm
1924    
1925         [This organization is the same that has been used already
1926         for a while for the architecture-specific parts of the visible
1927         compiler and for the old version of core.cm.]
1928    
1929         As you will notice, many source files have been moved from their
1930         respective original locations to a new home in one of the above
1931         subtrees.
1932    
1933         The division of labor between the new libraries is the following:
1934    
1935             basics.cm:
1936                - Simple, basic definitions that pertain to many (or all) of
1937                  the other libraries.
1938             parser.cm:
1939                - The SML parser, producing output of type Ast.dec.
1940                - The type family for Ast is also defined and exported here.
1941             elabdata.cm:
1942                - The datatypes that describe input and output of the elaborator.
1943                  This includes types, absyn, and static environments.
1944             elaborator.cm:
1945                - The SML/NJ type checker and elaborator.
1946                  This maps an Ast.dec (with a given static environment) to
1947                  an Absyn.dec (with a new static environment).
1948                - This libraries implements certain modules that used to be
1949                  structures as functors (to remove dependencies on FLINT).
1950             execute.cm:
1951                - Everything having to do with executing binary code objects.
1952                - Dynamic environments.
1953             core.cm:
1954                - SML/NJ-specific instantiations of the elaborator and MLRISC.
1955                - Top-level modules.
1956                - FLINT (this should eventually become its own library)
1957    
1958    Notes:
1959    
1960    I am not 100% happy with the way I separated the elaborator (and its
1961    data structures) from FLINT.  Two instances of the same problem:
1962    
1963        1. Data structures contain certain fields that carry FLINT-specific
1964           information.  I hacked around this using exn and the property list
1965           module from smlnj-lib.  But the fact that there are middle-end
1966           specific fields around at all is a bit annoying.
1967    
1968        2. The elaborator calculates certain FLINT-related information.  I tried
1969           to make this as abstract as I could using functorization, but, again,
1970           the fact that the elaborator has to perform calculations on behalf
1971           of the middle-end at all is not nice.
1972    
1973        3. Having to used exn and property lists is unfortunate because it
1974           weakens type checking.  The other alternative (parameterizing
1975           nearly *everything*) is not appealing, though.
1976    
1977    I removed the "rebinding =" warning hack because due to the new organization
1978    it was awkward to maintain it.  As a result, the compiler now issues some of
1979    these warnings when compiling init.cmi during bootstrap compilation. On
1980    the plus side, you also get a warning when you do, for example:
1981       val op = = Int32.+
1982    which was not the case up to now.
1983    
1984    I placed "assign" and "deref" into the _Core structure so that the
1985    code that deals with the "lazy" keyword can find them there.  This
1986    removes the need for having access to the primitive environment
1987    during elaboration.
1988    
1989    ----------------------------------------------------------------------
1990    Name: Matthias Blume
1991    Date: 2001/08/13
1992    Tag: blume-20010813-closures
1993    Description:
1994    
1995    This fix was sent to us by Zhong Shao.  It is supposed to improve the
1996    performance of certain loops by avoiding needless closure allocation.
1997    
1998    ----------------------------------------------------------------------
1999    Name: Lal George
2000    Date: 2001/07/31 10:03:23 EDT 2001
2001    Tag: george-20010731-x86-fmalloc
2002    Description: Fixed bug in x86 calls
2003    
2004        There was a bug where call instructions would mysteriously
2005        vanish. The call instruction had to be one that returned
2006        a floating point value.
2007    
2008    ----------------------------------------------------------------------
2009    Name: Lal George
2010    Date: 2001/07/19 16:36:29 EDT 2001
2011    Tag: george-20010719-simple-cells
2012    Description:
2013    
2014    I have dramatically simplified the interface for CELLS in MLRISC.
2015    
2016    In summary, the cells interface is broken up into three parts:
2017    
2018      1. CellsBasis : CELLS_BASIS
2019    
2020            CellsBasis is a top level structure and common for all
2021            architectures.  it contains the definitions of basic datatypes
2022            and utility  functions over these types.
2023    
2024      2. functor Cells() : CELLS
2025    
2026            Cells generates an interface for CELLS that incorporates the
2027            specific resources on the target architecture, such as the
2028            presence of special register classes, their number and size,
2029            and various useful substructures.
2030    
2031      3. <ARCH>CELLS
2032    
2033            e.g. SparcCells: SPARCCELLS
2034    
2035            <ARCH>CELLS usually contains additional bindings for special
2036            registers  on the architecture, such as:
2037    
2038                    val r0 : cell           (* register zero *)
2039                    val y : cell            (* Y register *)
2040                    val psr : cell          (* processor status register *)
2041                    ...
2042    
2043            The structure returned by applying the Cells functor is opened
2044            in this interface.
2045    
2046    The main implication of all this is that the datatypes for cells is
2047    split between CellsBasis and CELLS -- a fairly simple change for user
2048    code.
2049    
2050    In the old scheme the CELLS interface had a definitional binding of
2051    the form:
2052    
2053            signature CELLS = sig
2054    
2055               structure CellsBasis = CellsBasis
2056    
2057               ...
2058    
2059            end
2060    
2061    With all the sharing constraints that goes on in MLRISC, this old
2062    design  quickly leads to errors such as:
2063    
2064            "structure definition spec inside of sharing ... "
2065    
2066    
2067    and appears to require an unacceptable amount of sharing and where
2068    constraint hackery.
2069    
2070    I think this error message (the interaction of definitional specs and
2071    sharing) requires more explanation on our web page.
2072    
2073    ----------------------------------------------------------------------
2074    Name: Matthias Blume
2075    Date: 2001/07/19 15:00:00 EDT
2076    Tag: blume-20010719-libreorg
2077    Description:
2078    
2079    This update puts together a fairly extensive but straightforward change
2080    to the way the libraries that implement the interactive system are
2081    organized:
2082    
2083       The biggest change is the elimination of structure Compiler.  As a
2084       replacement for this structure, there is now a CM library
2085       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
2086       that exports all the substructures of the original structure Compiler
2087       directly.  So instead of saying Compiler.Foo.bar one now simply
2088       says Foo.bar.  (The CM libraries actually export a collection of
2089       structures that is richer than the collection of substructures of
2090       structure Compiler.)
2091    
2092       To make the transition smooth, there is a separate library called
2093       $smlnj/compiler/compiler.cm which puts together and exports the
2094       original structure Compiler (or at least something very close to it).
2095    
2096       There are five members of the original structure Compiler
2097       that are not exported directly but which instead became members
2098       of a new structure Backend (described by signature BACKEND).  These are:
2099       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
2100       Interact (: INTERACT), structure Machine (: MACHINE), and val
2101       architecture (: string).
2102    
2103       Structure Compiler.Version has become structure CompilerVersion.
2104    
2105       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
2106       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
2107       or x86, respectively.
2108       Each of these exports the same frontend structures that
2109       $smlnj/compiler.cm exports.  But they do not have a structure Backend
2110       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
2111       Hppa, PPC, Sparc, or X86, respectively.
2112    
2113       Library $smlnj/compiler/all.cm exports the union of the exports of
2114       $smlnj/compiler/<arch>.cm
2115    
2116       There are no structures <Arch>Compiler anymore, use
2117       $smlnj/compiler/<arch>.cm instead.
2118    
2119       Library host-compiler-0.cm is gone.  Instead, the internal library
2120       that instantiates CM is now called cm0.cm.  Selection of the host
2121       compiler (backend) is no longer done here but. (Responsibility for it
2122       now lies with $smlnj/compiler/current.cm.  This seems to be more
2123       logical.)
2124    
2125       Many individual files have been moved or renamed.  Some files have
2126       been split into multiple files, and some "dead" files have been deleted.
2127    
2128    Aside from these changes to library organization, there are also changes
2129    to the way the code itself is organized:
2130    
2131       Structure Binfile has been re-implemented in such a way that it no
2132       longer needs any knowledge of the compiler.  It exclusively deals
2133       with the details of binfile layout.  It no longer invokes the
2134       compiler (for the purpose of creating new prospective binfile
2135       content), and it no longer has any knowledge of how to interpret
2136       pickles.
2137    
2138       Structure Compile (: COMPILE) has been stripped down to the bare
2139       essentials of compilation.  It no longer deals with linking/execution.
2140       The interface has been cleaned up considerably.
2141    
2142       Utility routines for dealing with linking and execution have been
2143       moved into their own substructures.
2144    
2145       (The ultimate goal of these changes is to provide a light-weight
2146       binfile loader/linker (at least for, e.g., stable libraries) that
2147       does not require CM or the compiler to be present.)
2148    
2149    CM documentation has been updated to reflect the changes to library
2150    organization.
2151    
2152    ----------------------------------------------------------------------
2153    Name: Matthias Blume
2154    Date: 2001/07/10 17:30:00 EDT
2155    Tag: Release_110_34
2156    Description:
2157    
2158    Minor tweak to 110.34 (re-tagged):
2159    
2160      - README.html file added to CVS repository
2161      - runtime compiles properly under FreeBSD 3.X and 4.X
2162    
2163    ----------------------------------------------------------------------
2164    Name: Matthias Blume
2165    Date: 2001/07/10 17:30:00 EDT
2166    Tag: Release_110_34
2167    Description:
2168    
2169    New version number (110.34). New bootfiles.
2170    
2171    ----------------------------------------------------------------------
2172    Name: Matthias Blume
2173    Date: 2001/07/09 16:00:00 EDT
2174    Tag: blume-20010709-more-varargs
2175    Description:
2176    
2177    I changed the handling of varargs in ml-nlffigen again:
2178    The ellipsis ... will now simply be ignored (with an accompanying warning).
2179    
2180    The immediate effect is that you can actually call a varargs function
2181    from ML -- but you can't actually supply any arguments beyond the ones
2182    specified explicitly.  (For example, you can call printf with its format
2183    string, but you cannot pass additional arguments.)
2184    
2185    This behavior is only marginally more useful than the one before, but
2186    it has the advantage that a function or, more importantly, a function
2187    type never gets dropped on the floor, thus avoiding follow-up problems with
2188    other types that refer to the offending one.
2189    
2190    ----------------------------------------------------------------------
2191    Name: Matthias Blume
2192    Date: 2001/07/09 11:25:00 EDT
2193    Tag: blume-20010709-varargs
2194    Description:
2195    
2196    1. ckit-lib.cm now exports structure Error
2197    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
2198       with a warning accompanied by a source location.  Moreover, it
2199       merely skips the offending function or type and proceeds with the
2200       rest of its work.u  As a result, one can safely feed C code containing
2201       "..." to ml-nlffigen.
2202    3. There are some internal improvements to CM, providing slightly
2203       more general string substitutions in the tools subsystem.
2204    
2205    ----------------------------------------------------------------------
2206    Name: Matthias Blume
2207    Date: 2001/06/27 15:10:00 EDT
2208    Tag: blume-20010627-concur
2209    Description:
2210    
2211    Fixed a small bug in CM's handling of parallel compilation.
2212    (You could observe the bug by Control-C-interrupting an ordinary
2213    CMB.make or CM.stabilize and then attaching some compile servers.
2214    The result was that all of a sudden the previously interrupted
2215    compilation would continue on its own.  This was because of
2216    an over-optimization: CM did not bother to clean out certain queues
2217    when no servers were attached "anyway", resulting in the contents
2218    of these queues to grab control when new servers did get attached.)
2219    
2220    There is also another minor update to the CM manual.
2221    
2222    ----------------------------------------------------------------------
2223    Name: Matthias Blume
2224    Date: 2001/06/26 16:15:00 EDT
2225    Tag: blume-20010626-cmdoc
2226    Description:
2227    
2228    Minor typo fixed in CM manual (syntax diagram for libraries).
2229    
2230    ----------------------------------------------------------------------
2231    Name: Matthias Blume
2232    Date: 2001/06/25 22:55:00 EDT
2233    Tag: blume-20010625-x86pc
2234    Description:
2235    
2236    Fixed a nasty bug in the X86 assembly code that caused signal
2237    handlers to fail (crash) randomly.
2238    
2239    ----------------------------------------------------------------------
2240    Name: Matthias Blume
2241    Date: 2001/06/25 12:05:00 EDT
2242    Tag: blume-20010625-nlffigen
2243    Description:
2244    
2245    This update fixes a number of minor bugs in ml-nlffigen as reported by
2246    Nick Carter <nbc@andrew.cmu.edu>.
2247    
2248      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
2249      2. Default names for generated files are now derived from the name of
2250         the C file *without its directory*.  In particular, this causes generated
2251         files to be placed locally even if the C file is in some system directory.
2252      3. Default names for generated signatures and structures are also derived
2253         from the C file name without its directory.  This avoids silly things
2254         like "structure GL/GL".
2255         (Other silly names are still possible because ml-nlffigen does not do
2256          a thorough check of whether generated names are legal ML identifiers.
2257          When in doubt, use command line arguments to force particular names.)
2258    
2259    ----------------------------------------------------------------------
2260    Name: Matthias Blume
2261    Date: 2001/06/21 12:25:00 EDT
2262    Tag: blume-20010621-eXene
2263    Description:
2264    
2265    eXene now compiles and (sort of) works again.
2266    
2267    The library name (for version > 110.33) is $/eXene.cm.
2268    
2269    I also added an new example in src/eXene/examples/nbody.  See the
2270    README file there for details.
2271    
2272    ----------------------------------------------------------------------
2273    Name: Matthias Blume
2274    Date: 2001/06/20 16:40:00 EDT
2275    Tag: blume-20010620-cml
2276    Description:
2277    
2278    CML now compiles and works again.
2279    
2280    Libraries (for version > 110.33):
2281    
2282      $cml/cml.cm            Main CML library.
2283      $cml/basis.cm          CML's version of $/basis.cm.
2284      $cml/cml-internal.cm   Internal helper library.
2285      $cml/core-cml.cm       Internal helper library.
2286      $cml-lib/trace-cml.cm  Tracing facility.
2287      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
2288    
2289    The installer (config/install.sh) has been taught how to properly
2290    install this stuff.
2291    
2292    ----------------------------------------------------------------------
2293    Name: Matthias Blume
2294    Date: 2001/06/19 17:55:00 EDT
2295    Tag: blume-20010619-instantiate
2296    Description:
2297    
2298    This un-breaks the fix for bug 1432.
2299    (The bug was originally fixed in 110.9 but I broke it again some
2300    time after that.)
2301    
2302    ----------------------------------------------------------------------
2303    Name: Matthias Blume
2304    Date: 2001/06/19 17:25:00 EDT
2305    Tag: blume-20010619-signals
2306    Description:
2307    
2308    This should (hopefully) fix the long-standing signal handling bug.
2309    (The runtime system was constructing a continuation record with an
2310    incorrect descriptor which would cause the GC to drop data on the floor...)
2311    
2312    ----------------------------------------------------------------------
2313    Name: Matthias Blume
2314    Date: 2001/06/15 15:05:00 EDT
2315    Tag: blume-20010615-moresparc
2316    Description:
2317    
2318    Here is a short late-hour update related to Sparc c-calls:
2319    
2320     -- made handling of double-word arguments a bit smarter
2321    
2322     -- instruction selection phase tries to collapse certain clumsily
2323        constructed ML-Trees; typical example:
2324    
2325            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
2326    
2327        This currently has no further impact on SML/NJ since mlriscGen does
2328        not seem to generate such patterns in the first place, and c-calls
2329        (which did generate them in the beginning) has meanwhile been fixed
2330        so as to avoid them as well.
2331    
2332    ----------------------------------------------------------------------
2333    Name: Matthias Blume
2334    Date: 2001/06/15 15:05:00 EDT
2335    Tag: blume-20010615-sparc
2336    Description:
2337    
2338    The purpose of this update is to provide an implementation of NLFFI
2339    on Sparc machines.
2340    
2341    Here are the changes in detail:
2342    
2343       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
2344       the Sparc implementation of the c-calls API.
2345       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
2346       for accessing the ML frame.  Thus, we have a real frame pointer and
2347       can freely modify %sp without need for an omit-frame-ptr phase.
2348       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
2349       this case.
2350       * ml-nlffigen has been taught to produce code for different architectures
2351       and calling conventions.
2352       * In a way similar to what was done in the x86 case, the Sparc
2353       backend uses its own specific extension to mltree.  (For example,
2354       it needs to be able to generate UNIMP instructions which are part
2355       of the calling convention.)
2356       * ml-nlffi-lib was reorganized to make it more modular (in particular,
2357       to make it easier to plug in new machine- and os-dependent parts).
2358    
2359    There are some other fairly unrelated bug fixes and cleanups as well:
2360    
2361       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
2362       that they properly share their libraries with existing SML/NJ libraries.
2363       * I fixed a minor cosmetic bug in CM, supressing certain spurious
2364       follow-up error messages.
2365       * Updates to CM/CMB documentation.
2366    
2367    TODO items:
2368    
2369       * MLRISC should use a different register as its asmTemp on the Sparc.
2370         (The current %o2 is a really bad choice because it is part of the
2371         calling conventions, so things might interfere in unexpected ways.)
2372    
2373    ----------------------------------------------------------------------
2374    Name: Matthias Blume
2375    Date: 2001/06/07
2376    Tag: blume-20010607-calls
2377    Description:
2378    
2379    A number of internal changes related to C calls and calling conventions:
2380    
2381    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
2382       number of bytes popped implicitly (by the callee).  In most cases
2383       this field is 0 but on x86/win32 it is some non-zero value.  This
2384       is information provided for the benefit of the "omit-frameptr" pass.
2385    2. The CALL instruction on the x86 carries a similar "pops" field.
2386       The instruction selection phase copies its value from the ML-Tree
2387       CALL statement.
2388    3. On all other architectures, the instruction selection phase checks
2389       whether "pops=0" and complains if not.
2390    4. The c-calls implementation for x86 now accepts two calling conventions:
2391       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
2392       up after the call and pops is set to 0.  For "stdcall", the caller
2393       does nothing, leaving the cleanup to the callee; pops is set to
2394       the number of bytes that were pushed onto the stack.
2395    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
2396       distinguish between "ccall" and "stdcall".
2397    6. The UNIMP instruction has been added to the supported Sparc instruction
2398       set. (This is needed for implementing the official C calling convention
2399       on this architecture.)
2400    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
2401       work with the latest CM.
2402    
2403    ----------------------------------------------------------------------
2404    Name: Matthias Blume
2405    Date: 2001/06/05 15:10:00 EDT
2406    Tag: blume-20010605-cm-index
2407    Description:
2408    
2409    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
2410    
2411    1. CM can now generate "index files".  These are human-readable files
2412       that list on a per-.cm-file basis each toplevel symbol defined or
2413       imported.  The location of the index file for
2414       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
2415       To enable index-file generation, set CM.Control.generate_index to true
2416       or export an environment-symbol: export CM_GENERATE_INDEX=true.
2417    
2418       The CM manual has been updated accordingly.
2419    
2420    2. I made some slight modifications to the c-calls API in MLRISC.
2421    
2422         a) There is now a callback to support saving/restoring of
2423            dedicated but caller-save registers around the actual call
2424            instruction.
2425         b) One can optionally specify a comment-annotation for the
2426            call instruction.
2427    
2428    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
2429       (For example, the comment annotation shows the C prototype of
2430        the function being called.)
2431    
2432    ----------------------------------------------------------------------
2433    Name: Matthias Blume
2434    Date: 2001/06/01 13:30:00 EDT
2435    Tag: blume-20010601-nlffi-cleanup
2436    Description:
2437    
2438    This is mostly a cleanup of MLFFI stuff:
2439    
2440       - some signature files have been put into a more exposed place
2441       - the ugly 'f type parameter is gone (simplifies types tremendously!)
2442       - ml-nlffigen changed accordingly
2443       - tutorial updated
2444    
2445    Other changes:
2446    
2447       - author's affiliation in CM manual(s) updated
2448       - some more recognized keywords added to Allen's sml.sty
2449    
2450    ----------------------------------------------------------------------
2451    Name: Matthias Blume
2452    Date: 2001/05/25 15:30:00 EDT
2453    Tag: blume-20010525-iptr
2454    Description:
2455    
2456      - put the official 110.33-README (as it appears on the ftp server) under
2457        CVS
2458      - fixed a small bug related to incomplete pointer types in
2459        ml-nlffigen
2460      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
2461        (it does not need the 'f type parameter)
2462    
2463    ----------------------------------------------------------------------
2464    Name: Matthias Blume
2465    Date: 2001/05/23 14:30:00 EDT
2466    Tag: Release_110_33
2467    Description:
2468    
2469    New version number (110.33).  New bootfiles.
2470    
2471    ----------------------------------------------------------------------
2472    Name: Matthias Blume
2473    Date: 2001/05/22 18:06:00 EDT
2474    Tag: blume-20010522-targets
2475    Description:
2476    
2477    Made install.sh use file config/targets.customized if it exists, falling
2478    back to config/targets if it doesn't.  This way one can have a customized
2479    version of the targets file without touching the "real thing", thus
2480    eliminating the constant fear of accidentally checking something bogus
2481    back into the CVS repository...  (File config/targets.customized must
2482    not be added to the repository!)
2483    
2484    ----------------------------------------------------------------------
2485    Name: Matthias Blume
2486    Date: 2001/05/22 16:30:00 EDT
2487    Tag: blume-20010522-minitut
2488    Description:
2489    
2490    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
2491       struct returns.
2492    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
2493       incomplete, preliminary documentation for NLFFI.
2494    
2495    ----------------------------------------------------------------------
2496    Name: Matthias Blume
2497    Date: 2001/05/14 11:30:00 EDT
2498    Tag: blume-20010514-script
2499    Description:
2500    
2501    Some bugs in install script fixed.
2502    
2503    In addition to that I also made a slight change to the NLFFI API:
2504    Functors generated by ml-nlffigen now take the dynamic library as a
2505    straight functor argument, not as a suspended one.  (The original
2506    functor code used to force the suspension right away anyway, so there
2507    was nothing gained by this complication of the interface.)
2508    
2509    ----------------------------------------------------------------------
2510    Name: Matthias Blume
2511    Date: 2001/05/11 14:35:00 EDT
2512    Tag: blume-20010511-ml-nlffi
2513    Description:
2514    
2515    I finally took the plunge and added my new FFI code to the main
2516    repository.  For x86-linux it is now ready for prime-time.
2517    
2518    There are two new subdirectories of "src":
2519    
2520      - ml-nlffi-lib:
2521           The utility library for programs using the FFI interface.
2522           Here is the implementation of $/c.cm and its associated low-level
2523           partners $/c-int.cm and $/memory.cm.
2524      - ml-nlffigen:
2525           A stand-alone program for generating ML glue code from C source
2526           code.
2527    
2528    Building ml-nlffigen requires $/ckit-lib.cm.
2529    
2530    The config/install.sh script has been updates to do the Right Thing
2531    (hopefully).
2532    
2533    Notice that the source tree for the C-Kit will not be put under "src"
2534    but directly under the installation root directory.  (This is the
2535    structure that currently exists on the CVS server when you check out
2536    module "sml".)  Fortunately, config/install.sh knows about this oddity.
2537    
2538    Bugs: No documentation yet.
2539    
2540    ----------------------------------------------------------------------
2541    Name: Matthias Blume
2542    Date: 2001/05/09 16:35:00 EDT
2543    Tag: blume-20010509-cpscontract
2544    Description:
2545    
2546    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
2547    wrapper/unwrapper elimination did not decrement usage counts and some
2548    dead variables got overlooked by the dead-up logic.)
2549    
2550    ----------------------------------------------------------------------
2551    Name: Lal George
2552    Date: 2001/05/08  17:26:09 EDT
2553    Tag: george-20010508-omit-frameptr
2554    Description:
2555    
2556    Changes to implement the omit-frame-pointer optimization to support
2557    raw C calls. For now, there is only support on the Intel x86, but
2558    other architectures will follow as more experience is gained with this.
2559    
2560    
2561    ----------------------------------------------------------------------
2562    Name: Matthias Blume
2563    Date: 2001/05/07 14:40:00 EDT
2564    Tag: blume-20010507-proxies
2565    Description:
2566    
2567    I made into "proxy libraries" all libraries that qualify for such a
2568    change.  (A qualifying library is a library that has another library or
2569    groups as its sole member and repeats that member's export list
2570    verbatim.  A proxy library avoids this repetition by omitting its export
2571    list, effectively inheriting the list that its (only) member exports.
2572    See the CM manual for more explanation.)
2573    The main effect is that explicit export lists for these libraries
2574    do not have to be kepts in sync, making maintenance a bit easier.
2575    
2576    I also added copyright notices to many .cm-files.
2577    
2578    Last but not least, I made a new set of bootfiles.
2579    
2580    ----------------------------------------------------------------------
2581    Name: Matthias Blume
2582    Date: 2001/05/04 17:00:00 EDT
2583    Tag: blume-20010504-cm-lsplit
2584    Description:
2585    
2586    0. John merged pending changes to $/smlnj-lib.cm
2587    
2588    1. Allen's previous change accidentally backed out of one of Lal's
2589       earlier changes.  I undid this mistake (re-introducing Lal's change).
2590    
2591    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
2592       within the compiler where applicable.  There is some code simplification
2593       because of that.
2594    
2595    3. The "split" phase (in FLINT) is now part of the default list of phases.
2596       Compiler.Control.LambdaSplitting.* can be used to globally control the
2597       lambda-splitting (cross-module-inlining) engine.  In addition to that,
2598       it can now also be controlled on a per-source basis: CM has been taught
2599       a new tool parameter applicable to ML source files.
2600    
2601       - To turn lambda-splitting off completely:
2602            local open Compiler.Control.LambdaSplitting in
2603                val _ = set Off
2604            end
2605       - To make "no lambda-splitting" the global default (but allow per-source
2606         overriding); this is the initial setting:
2607            local open Compiler.Control.LambdaSplitting in
2608                val _ = set (Default NONE)
2609            end
2610       - To make "lambda-splitting with aggressiveness a" the global default
2611         (and allow per-source overriding):
2612            local open Compiler.Control.LambdaSplitting in
2613                val _ = set (Default (SOME a))
2614            end
2615    
2616       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
2617         write (in the respective .cm-file):
2618            a.sml (lambdasplitting:off)
2619       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
2620            a.sml (lambdasplitting:on)
2621       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
2622         <a> is a decimal non-negative integer):
2623            a.sml (lambdasplitting:<a>)
2624       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
2625            a.sml (lambdasplitting:infinity)
2626       - To use the global default for a.sml:
2627            a.sml (lambdasplitting:default)
2628         or simply
2629            a.sml
2630    
2631    ----------------------------------------------------------------------
2632    Name: Allen Leung
2633    Date: 2001/05/04 01:57:00 EDT
2634    Tag: leunga-20010504-sync
2635    Description:
2636    
2637      MLRISC features.
2638    
2639      1. Fix to CMPXCHG instructions.
2640      2. Changed RA interface to allow annotations in callbacks.
2641      3. Added a new method to the stream interface to allow annotations updates.
2642    
2643    ----------------------------------------------------------------------
2644    Name: Matthias Blume
2645    Date: 2001/05/01 11:45:00 EDT
2646    Tag: blume-20010501-pcedittmp
2647    Description:
2648    
2649    Changed install.sh to use the current working directory instead of
2650    /usr/tmp for a temporary file (pcedittmp).  The previous choice
2651    of /usr/tmp caused trouble with MacOS X because of file premission
2652    problems.
2653    
2654    ----------------------------------------------------------------------
2655    Name: Matthias Blume
2656    Date: 2001/04/20 11:10:00 EDT
2657    Tag: blume-20010420-inMLflag
2658    Description:
2659    
2660     - added vp_limitPtrMask to vproc-state.h
2661       (for use by the raw-C-calls mechanism to implement proper interrupt
2662        handling)
2663     - made the ML compiler aware of various data-structure offsets so it
2664       can generate code for accessing the vp_inML flag and vp_limitPtrMask
2665     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
2666       raw C-calls
2667    
2668    ----------------------------------------------------------------------
2669    Name: Lal George
2670    Date: 2001/04/20 09:15:28 EDT
2671    Tag: george-20010420-macosX
2672    Description:
2673    
2674     - Changes to port to Mac OS X; Darwin.
2675    
2676     - In the process I found that sqrt was broken on the PPC, because the
2677       fsqrt instruction is not implemented.
2678    
2679    ----------------------------------------------------------------------
2680    Name: Matthias Blume
2681    Date: 2001/04/18 12:45:00 EDT
2682    Tag: blume-20010418-ccalls
2683    Description:
2684    
2685     - fixed two off-by-4 errors in the x86-specific c-calls implementation
2686       (this bug prevented structure arguments containing pointers from being
2687        passed correctly)
2688     - changed the raw-C-call code in mlriscGen.sml in such a way that
2689       structure arguments are represented as a pointer to the beginning
2690       of the structure (instead of having a series of synthesized arguments,
2691       one for each structure member)
2692    
2693     - made makeml script's verbosity level configurable via environment
2694       variable (MAKEML_VERBOSITY)
2695    
2696     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
2697       in rawmem-x86.sml; we are now using the real thing
2698    
2699    ----------------------------------------------------------------------
2700    Name: Matthias Blume
2701    Date: 2001/03/22 16:25:00 EST
2702    Tag: blume-20010322-bootfiles
2703    Description:
2704    
2705    Created a new set of bootfiles (for your automatic installation convenience).
2706    
2707    ----------------------------------------------------------------------
2708    Name: Matthias Blume
2709    Date: 2001/03/22 15:10:00 EST
2710    Tag: blume-20010322-rawmem-parcm
2711    Description:
2712    
2713    1. All "raw memory access" primitives for the new FFI are implemented now
2714       (at least on the x86).
2715    2. Some further cleanup of CM's parallel make mechanism.
2716    
2717    ----------------------------------------------------------------------
2718    Name: Matthias Blume
2719    Date: 2001/03/19 17:53:00 EST
2720    Tag: blume-20010319-parallel
2721    Description:
2722    
2723    Parallel make (using compile servers) now works again.
2724    
2725    To this end, CM.stabilize and CMB.make have been modified to work in
2726    two passes when compile servers are attached:
2727       1. Compile everything, do not perform stabilization; this pass
2728          uses compile servers
2729       2. Stabilize everything; this pass does not use compile servers
2730    If there are no compile servers, the two passes are combined into one
2731    (as before).  Splitting the passes increases the inherent parallelism
2732    in the dependency graph because the entire graph including all
2733    libraries is available at the same time.  This, in turn, improves
2734    server utilization.  The downside is that the master process will
2735    have to do some extra work after compilation is done (because for
2736    technical reasons it must re-read all the binfiles during stabilization).
2737    
2738    ----------------------------------------------------------------------
2739    Name: Matthias Blume
2740    Date: 2001/03/16 12:22:00 EST
2741    Tag: blume-20010316-bootfiles
2742    Description:
2743    
2744    Created a new set of bootfiles (for your automatic installation convenience).
2745    
2746    ----------------------------------------------------------------------
2747    Name: Matthias Blume
2748    Date: 2001/03/16 11:00:00 EST
2749    Tag: blume-20010316-MLTREE-fixup
2750    Description:
2751    
2752    This is a minor fixup for an (untagged) earlier commit by Allen.
2753    (A file was missing).
2754    
2755    ----------------------------------------------------------------------
2756    Name: Allen Leung
2757    Date: Mon Mar  5 18:54:57 EST 2001
2758    Tag: leunga-20010305-cut-support
2759    
2760    1. New support for alternative control-flow in MLTREE.
2761       Currently we support
2762    
2763          FLOW_TO(CALL ...., [k1,...,kn])
2764    
2765       This is needed for 'cuts to' in C-- and try/handle-like constructs
2766       in Moby
2767    
2768       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
2769    
2770    2. Register Allocator
2771    
2772       Changes in interface [from Fermin, John]
2773    
2774    3. Alpha 8-bit SLL support [Fermin]
2775    
2776    4. All architectures
2777    
2778       A new module (ClusterExpandCopies) for expanding parallel copies.
2779    
2780    ----------------------------------------------------------------------
2781    Name: Allen Leung
2782    Date: 2001/02/27 23:07:00 EST
2783    Tag: leunga-20010227-minor-stuff
2784    
2785    1. Alpha bug fix for CMOVNE
2786    2. Handle mltree COND(..,FCMP ...,...)
2787    3. Bug fix in simplifier
2788    
2789    ----------------------------------------------------------------------
2790    Name: Matthias Blume
2791    Date: 2001/01/30 17:50:00 EST
2792    Tag: blume-20010130-sync
2793    Description:
2794    
2795    This is just a minor update to sync my devel branch with the main brach.
2796    The only visible change is the addition of some README files.
2797    
2798    ----------------------------------------------------------------------
2799    Name: Matthias Blume
2800    Date: 2001/01/12 23:30:00 JST
2801    Tag: blume-20010112-bootfiles
2802    Description:
2803    
2804    Made a new set of bootfiles that goes with the current state of the
2805    repository.
2806    
2807    ----------------------------------------------------------------------
2808    Name: Matthias Blume
2809    Date: 2001/01/12 21:20:00 JST
2810    Tag: blume-20010112-sync
2811    Description:
2812    
2813    I am just flushing out some minor changes that had accumulated in
2814    my private branch in order to sync with the main tree.  (This is
2815    mainly because I had CVS trouble when trying to merge _into_ my
2816    private branch.)
2817    
2818    Most people should be completely unaffected by this.
2819    
2820    ----------------------------------------------------------------------
2821    Name: Allen Leung
2822    Date: Thu Jan 11 21:03:00 EST 2001
2823    Tag: leunga-20010111-labexp=mltree
2824    Description:
2825    
2826    1.  Removed the type LabelExp and replace it by MLTree.
2827    2.  Rewritten mltree-simplify with the pattern matcher tool.
2828    3.  There were some bugs in alpha code generator which would break
2829        64-bit code generation.
2830    4.  Redo the tools to generate code with the
2831    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
2832        are now generated by perl scripts.
2833    
2834    ----------------------------------------------------------------------
2835    Name: Matthias Blume
2836    Date: 2001/01/10 21:55:00 JST
2837    Tag: blume-20010110-rcc
2838    Description:
2839    
2840    The RCC stuff now seems to work (but only on the x86).
2841    This required hacking of the c-calls interface (and -implementation) in
2842    MLRISC.
2843    
2844    Normal compiler users should be unaffected.
2845    
2846    ----------------------------------------------------------------------
2847    Name: Matthias Blume
2848    Date: 2001/01/09 01:20:00 JST
2849    Tag: blume-20010109-rcc
2850    Description:
2851    
2852    This is a fairly big patch, flushing out a large number of pending
2853    changes that I made to my development copy over the last couple of days.
2854    
2855    Of practical relevance at this moment is a workaround for a pickling
2856    bug that Allen ran into the other day.  The cause of the bug itself is
2857    still unknown and it might be hard to fix it properly, but the
2858    workaround has some merits of its own (namely somewhat reducing pickling
2859    overhead for certain libraries).  Therefore, I think this solution should
2860    be satisfactory at this time.
2861    
2862    The rest of the changes (i.e., the vast majority) has to do with my
2863    ongoing efforts of providing direct support for C function calls from
2864    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
2865    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
2866    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
2867    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
2868    handle RCC.
2869    
2870    The last part (the changes to mlriscGen.sml) are still known to be
2871    wrong on the x86 and not implemented on all other architectures.  But
2872    the infrastructure is in place. I had to change a few functor
2873    signatures in the backend to be able to route the CCalls interface
2874    from MLRISC there, and I had to specialize the mltree type (on the
2875    x86) to include the necessary extensions. (The extensions themselves
2876    were already there and redy to go in MLRISC/x86).
2877    
2878    Everything should be very happy as soon as someone helps me with
2879    mlriscGen.sml...
2880    
2881    In any case, nothing of this should matter to anyone as long as the
2882    new primop is not being used (which is going to be the case unless you
2883    find it where I hid it :). The rest of the compiler is completely
2884    unaffected.
2885    
2886    ----------------------------------------------------------------------
2887    Name: Matthias Blume
2888    Date: 2001/01/05 00:30:00 JST
2889    Tag: blume-20010105-primops
2890    Description:
2891    
2892    Added some experimental support for work that I am doing right now.
2893    These changes mostly concern added primops, but there is also a new
2894    experimental C library in the runtime system (but currently not enabled
2895    anywhere except on Linux/X86).
2896    
2897    In the course of adding primops (and playing with them), I discovered that
2898    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
2899    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
2900    To recover, I made type information in INL_PRIM mandatory and changed
2901    prim.sml as well as built-in.sml accordingly.  The InLine structure now
2902    has complete, correct type information (i.e., no bottom types).
2903    
2904    Since all these changes mean that we need new binfiles, I also bumped the
2905    version number to 110.32.1.
2906    
2907    ----------------------------------------------------------------------
2908    Name: Matthias Blume
2909    Date: 2000/12/30 22:10:00 JST
2910    Tag: blume-20001230-various
2911    Description:
2912    
2913    Added proxy libraries for MLRISC and let MLRISC libraries refer
2914    to each other using path anchors.  (See CM manual for explanation.)
2915    
2916    Updated CM documentation.
2917    
2918    Fixed some bugs in CM.
2919    
2920    Implemented "proxy" libraries (= syntactic sugar for CM).
2921    
2922    Added "-quiet" option to makeml and changed runtime system accordingly.
2923    
2924    Added cleanup handler for exportML to reset timers and compiler stats.
2925    
2926    ----------------------------------------------------------------------
2927    Name: Lal George
2928    Date: 2000/12/22 22:22:58 EST 2000
2929    Tag: Release_110_32
2930    Description:
2931    
2932            Infinite precision used throughout MLRISC.
2933            see MLRISC/mltree/machine-int.sig
2934    
2935    ----------------------------------------------------------------------
2936    Name: Matthias Blume
2937    Date: 2000/12/22 23:16:00 JST
2938    Tag: blume-20001222-warn
2939    Description:
2940    
2941    Corrected wording and formatting of some CM warning message which I
2942    broke in my previous patch.
2943    
2944    ----------------------------------------------------------------------
2945    Name: Matthias Blume
2946    Date: 2000/12/22 21:20:00 JST
2947    Tag: blume-20001222-anchorenv
2948    Description:
2949    
2950    Fixed CM's handling of anchor environments in connection with CMB.make.
2951    
2952    ----------------------------------------------------------------------
2953    Name: Matthias Blume
2954    Date: 2000/12/22 13:15:00 JST
2955    Tag: blume-20001222-cleanup
2956    Description:
2957    
2958    Removed src/cm/ffi which does not (and did not) belong here.
2959    
2960    ----------------------------------------------------------------------
2961    Name: Matthias Blume
2962    Date: 2000/12/21 23:55:00 JST
2963    Tag: blume-20001221-exn
2964    Description:
2965    
2966    Probably most important: CM no longer silently swallows all exceptions
2967    in the compiler.
2968    Plus: some other minor CM changes.  For example, CM now reports some
2969    sizes for generated binfiles (code, data, envpickle, lambdapickle).
2970    
2971    ----------------------------------------------------------------------
2972    Name: Matthias Blume
2973    Date: 2000/12/15 00:01:05 JST
2974    Tag: blume-20001215-dirtool
2975    Description:
2976    
2977    - "dir" tool added.
2978    - improvements and cleanup to Tools structure
2979    - documentation updates
2980    
2981    ----------------------------------------------------------------------
2982    Name: Allen Leung
2983    Date: Thu Dec 14 03:45:24 EST 2000
2984    Description:
2985    Tag:  leunga-20001214-int-inf
2986    Description:
2987    
2988       In IntInf, added these standard functions, which are missing from our
2989    implementation:
2990    
2991        andb : int * int -> int
2992        xorb : int * int -> int
2993        orb  : int * int -> int
2994        notb : int -> int
2995         <<   : int * word -> int
2996        ~>>  : int * word -> int
2997    
2998       Not tested, I hope they are correct.
2999    
3000    ----------------------------------------------------------------------
3001    Name: Allen Leung
3002    Date: Fri Dec  8 19:23:26 EST 2000
3003    Description:
3004    Tag:  leunga-20001208-nowhere
3005    Description:
3006    
3007      Slight improvements to the 'nowhere' tool to handle OR-patterns,
3008    to generate better error messages etc.  Plus a brief manual.
3009    
3010    ----------------------------------------------------------------------
3011    Name: Lal George
3012    Date: 2000/12/08 09:54:02 EST 2000
3013    Tag: Release_110_31
3014    Description:
3015    
3016    - Version 110.31
3017    ----------------------------------------------------------------------
3018    Name: Allen Leung
3019    Date: Thu Dec  7 22:01:04 EST 2000
3020    Tag:  leunga-20001207-cell-monster-hack
3021    Description:
3022    
3023    Major MLRISC internal changes.  Affect all clients.
3024    Summary:
3025    
3026    1.  Type CELLS.cell = int is now replaced by a datatype.
3027        As a result, the old regmap is now gone.  Almost all interfaces
3028        in MLRISC change as a consequence.
3029    
3030    2.  A new brand version of machine description tool (v3.0) that generates
3031        modules expecting the new interface.  The old version is removed.
3032    
3033    3.  The RA interface has been further abstracted into two new functors.
3034        RISC_RA and X86RA.  These functors have much simpler interfaces.
3035        [See also directory MLRISC/demo.]
3036    
3037    4.  Some other new source->source code generation tools are available:
3038    
3039        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
3040        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
3041           I use this tool to generate the peephole optimizers---with the new
3042           cell type changes, peephole rules are becoming difficult to write
3043           without conditional pattern matching.
3044    
3045    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
3046        cover the entire MLRISC source tree so many things broke.
3047    
3048    6.  CM files have been moved to the subdirectory MLRISC/cm.
3049        They are moved because there are a lot of them and they clutter up the
3050        root dir.
3051    
3052    7.  More detailed documentation to come...
3053    
3054        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
3055        a makeml -rebuild first.  This is because of other other
3056        changes that Matthias has made (see below).
3057    
3058    
3059    ----------------------------------------------------------------------
3060    Name: Matthias Blume
3061    Date: 2000/11/30 23:12:00 JST
3062    Tag: blume-20001130-filereorg
3063    Description:
3064    
3065    Some manual updates and some file reorganizations in CM.
3066    
3067    ----------------------------------------------------------------------
3068    Name: Matthias Blume
3069    Date: 2000/11/24 17:45:00 JST
3070    Tag: blume-20001124-link
3071    Description:
3072    
3073    Drastically improved link traversal code for the case that the dynamic
3074    value was already loaded at bootstrap time.  As a result, CM and CMB
3075    now both load blazingly fast -- even on a very slow machine.  Also,
3076    memory consumption has been further reduced by this.
3077    
3078    Warning: The format of the PIDMAP file has changed.  THerefore, to
3079    bootstrap you have to do this:
3080    
3081    1. Run CMB.make
3082    2. Make a symbolic link for the boot directory:
3083         ln -s sml.boot.ARCH-OS xxx
3084    3. "Rebuild" the boot directory:
3085         ./makeml -boot xxx -rebuild sml ; rm xxx
3086    4. Boot normally:
3087          ./makeml
3088    
3089    ----------------------------------------------------------------------
3090    Name: Matthias Blume
3091    Date: 2000/11/21 21:20:00 JST
3092    Tag: blume-20001121-tools
3093    Description:
3094    
3095    Continued hacking on autoloading problem -- with success this time.
3096    Also changed tool-plugin mechanism.  See new CM manual.
3097    
3098    ----------------------------------------------------------------------
3099    Name: Matthias Blume
3100    Date: 2000/11/19 14:30:00 JST
3101    Tag:  blume-20001119-autoload
3102    Description:
3103    
3104    Some hacking to make autoloading faster.  Success for CMB, no success
3105    so far for CM.  There is a reduced structure CM' that autoloads faster.
3106    (This is a temporary, non-documented hack to be eliminated again when
3107    the general problem is solved.)
3108    
3109    ----------------------------------------------------------------------
3110    Name: Matthias Blume
3111    Date: 2000/11/17 14:10:00 JST
3112    Tag: blume-20001117-pickle-lib
3113    Description:
3114    
3115    1. Eliminated comp-lib.cm
3116    2. Made pickle-lib.cm
3117    3. Eliminated all uses of intset.sml (from comp-lib.cm)
3118    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
3119       equivalent constructs from smlnj-lib.cm (INtHashTable).
3120    5. Point 4. also goes for those uses of intmap.* in MLRISC.
3121       Duplicated intmap modules thrown out.
3122    6. Hunted down all duplicated SCC code and replaced it with
3123       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
3124    7. Rewrote Feedback module.
3125    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
3126       should be thrown out and equivalent modules from smlnj-lib.cm
3127       should be used (IntRedBlackSet, IntListSet, ...).
3128    
3129    Confirmed that compiler compiles to fixpoint.
3130    
3131    ----------------------------------------------------------------------
3132    Name: Allen Leung
3133    Date: 2000/11/10 18:00:00
3134    Tag: leunga-20001110-new-x86-fp
3135    
3136    A new x86 floating point code generator has been added.
3137    By default this is turned off.  To turn this on, do:
3138    
3139        CM.autoload "$smlnj/compiler.cm";
3140        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
3141    
3142    Changes:
3143    
3144    1.  Changed FTAN to FPTAN so that the assembly output is correct.
3145    2.  Changed the extension callback for FTANGENT to generate:
3146    
3147              fptan
3148              fstp  %st(0)
3149        instead of
3150              fptan
3151              fstpl ftempmem
3152    
3153    3.  Numerous assembly fixes for x86.
3154    
3155    5.  Cleaned up the machine code output module x86/x86MC.sml and added
3156        support for a whole bunch of instructions and addressing modes:
3157    
3158          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
3159          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
3160          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
3161          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
3162          fxch %st(n)
3163          fld %st(n)
3164          fst %st(n)
3165          fst mem
3166          fstp %st(n)
3167          fucom %st(n)
3168          fucomp %st(n)
3169    
3170        All these are now generated when the fast fp mode is turned on.
3171    
3172    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
3173    
3174    ----------------------------------------------------------------------
3175    Name: Matthias Blume
3176    Date: 2000/11/09 11:20:00 JST
3177    Tag: blume-20001109-scc
3178    Description:
3179    
3180    Eliminated some code duplication:
3181    
3182    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
3183       (Otherwise the functor is useless.)
3184    2. Used GraphSCCFn where SCCUtilFun was used previously.
3185    3. Got rid of SCCUtilFun (in comp-lib.cm).
3186    
3187    ----------------------------------------------------------------------
3188    Name: Lal George
3189    Date: 2000/11/06 09:02:21 EST 2000
3190    Tag: Release_110_30
3191    Description:
3192    
3193    - Version 110.30
3194    ----------------------------------------------------------------------
3195    Name: Matthias Blume
3196    Date: 2000/11/04 14:45:00
3197    Tag: blume-20001104-mlbuild
3198    Description:
3199    
3200    - Made ml-build faster on startup.
3201    - Documentation fixes.
3202    
3203    ----------------------------------------------------------------------
3204    Name: Matthias Blume
3205    Date: 2000/11/02 17:00:00 JST
3206    Tag: blume-20001102-condcomp
3207    Description:
3208    
3209    - Small tweaks to pickler -- new BOOTFILES!
3210    - Version bumped to 110.29.2.
3211    - Added conditional compilation facility to init.cmi (see comment there).
3212    ----------------------------------------------------------------------
3213    Name: Allen Leung
3214    Date: 2000/10/23 19:31:00
3215    Tag: leunga-20001023-demo-ra
3216    
3217    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
3218    2. Test programs for the graph library updated
3219    3. Some new MLRISC demo programs added
3220    
3221    ----------------------------------------------------------------------
3222    Name: Matthias Blume
3223    Date: 2000/08/31 22:15:00 JST
3224    Tag: blume-20001017-errmsg
3225    Description:
3226    
3227    More error message grief: Where there used to be no messages, there
3228    now were some that had bogus error regions.  Fixed.
3229    
3230    ----------------------------------------------------------------------
3231    Name: Matthias Blume
3232    Date: 2000/08/31 17:30:00 JST
3233    Tag: blume-20001017-v110p29p1
3234    Description:
3235    
3236    I made a version 110.29.1 with new bootfiles.
3237    
3238    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
3239              CM documentation changes and a small bugfix in CM's error reporting.
3240    
3241    ----------------------------------------------------------------------
3242    Name: Lal George
3243    Date: 2000/09/27 14:42:35 EDT
3244    Tag: george-20000927-nodestatus
3245    Description:
3246    
3247    Changed the type of the nodestatus, so that:
3248    
3249            SPILLED(~1)             is now SPILLED
3250            SPILLED(m) where m>=0   is now MEMREG(m)
3251            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
3252    
3253    ----------------------------------------------------------------------
3254    Name: Matthias Blume
3255    Date: 2000/09/07 14:45:00 JST
3256    Tag: blume-20000907-cmerrmsg
3257    Description:
3258    
3259    Small tweak to CM to avoid getting ML syntax error messages twice.
3260    
3261    ----------------------------------------------------------------------
3262    Name: Matthias Blume
3263    Date: 2000/08/31 18:00:00 JST
3264    Tag: blume-20000831-cvsbootfiles
3265    Description:
3266    
3267    New URL for boot files (because the 110.29 files on the BL server do
3268    now work correctly with my updated install scripts for yacc and lex).
3269    
3270    ----------------------------------------------------------------------
3271    Name: Matthias Blume
3272    Date: 2000/08/08 12:33:00 JST
3273    Tag: blume-20000808-manual
3274    Description:
3275    
3276    Tiny update to CM manual.
3277    
3278    ----------------------------------------------------------------------
3279    Name: Allen Leung
3280    Date: 2000/08/7 19:31:00
3281    Tag: leunga-20000807-a-whole-bunch-of-stuff
3282    
3283      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
3284    and Mobdy.
3285    
3286    1.  x86
3287    
3288       a.  Fixes to peephole module by John and Dan.
3289       b.  Assembly fix to SETcc by Allen.
3290       c.  Fix to c-call by John.
3291       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
3292       e.  Instruction selection optimization to SETcc as suggested by John.
3293    
3294           For example,
3295    
3296            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
3297    
3298           should generate:
3299    
3300            MOVL a, x
3301            SUBL b, x
3302            SHRL 31, x
3303    
3304    2.  IR stuff
3305    
3306         A bunch of new DJ-graph related algorithms added.  These
3307         speed up SSA construction.
3308    
3309    3.  SSA + Scheduling
3310    
3311         Added code for SSA and scheduling to the repository
3312    
3313    ----------------------------------------------------------------------
3314    Name: Lal George
3315    Date: 2000/07/27 11:53:14 EDT
3316    
3317    Tag: lal-20000727-linux-ppc
3318    Description:
3319    
3320     Made changes to support Linux PPC.
3321     p.s. I have confirmation that the 110.29 boot files work fine.
3322    
3323    ----------------------------------------------------------------------
3324    Name: Matthias Blume
3325    Date: 2000/07/27 17:40:00 JST
3326    Tag: blume-20000727-scripts
3327    Description:
3328    
3329    !!!! WARNING !!!!
3330    You must recompile the runtime system!
3331    !!!! WARNING !!!!
3332    
3333    This is basically another round of script-enhancements:
3334    
3335    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
3336       and undefine CM preprocessor symbols.
3337    
3338    2. ml-build avoids generating a new heap image if it finds that the
3339       existing one is still ok.  (The condition is that no ML file had to
3340       be recompiled and all ML files are found to be older that the heap
3341       file.)
3342    
3343       To make this work smoothly, I also hacked the runtime system as
3344       well as SMLofNJ.SysInfo to get access to the heap image suffix
3345       (.sparc-solaris, ...) that is currently being used.
3346    
3347       Moreover, the signature of CM.mk_standalone has changed.  See the
3348       CM manual.
3349    
3350    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
3351       CM manual for details.)
3352    
3353    4. More CM manual updates:
3354        - all of the above has been documented.
3355        - there is now a section describing the (CM-related) command line
3356          arguments that are accepted by the "sml" command
3357    
3358    ----------------------------------------------------------------------
3359    Name: Matthias Blume
3360    Date: 2000/07/25 16:20:00 JST
3361    Tag: blume-20000725-makedepend
3362    Description:
3363    
3364    Added a script called ml-makedepend.  This can be used in makefiles
3365    for Unix' make in a way very similar to the "makedepend" command for
3366    C.
3367    
3368    The script internally uses function CM.sources.
3369    
3370    Synopsis:
3371    
3372        ml-makedepend [-f makefile] cmfile targetname
3373    
3374    The default for the makefile is "makefile" (or "Makefile" should
3375    "makefile" not exist).
3376    
3377    ml-makedepend adds a cmfile/targetname-specific section to this
3378    makefile (after removing the previous version of this section).  The
3379    section contains a single dependency specification with targetname on
3380    the LHS (targetname is an arbitrary name), and a list of files derived
3381    from the cmfile on the RHS.  Some of the files on the RHS are
3382    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
3383    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
3384    path names.  The makefile writer is responsible for making sure that
3385    these variables have correct at the time "make" is invoked.
3386    
3387    ----------------------------------------------------------------------
3388    Name: Matthias Blume
3389    Date: 2000/07/22 23:30:00 JST
3390    Tag: blume-20000722-urlupdate
3391    Description:
3392    
3393    Changed BOOT and config/srcarchiveurl to point to BL server:
3394    
3395        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
3396    
3397    ----------------------------------------------------------------------
3398    Name: Matthias Blume
3399    Date: 2000/07/18 18:00:00 JST
3400    Tag: blume-20000718-Version_110_29
3401    Description:
3402    
3403    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
3404    
3405    2. Updated config/version to 110.29
3406    
3407    3. Updated config/srcarchiveurl
3408    
3409    3. New boot files!
3410       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
3411    
3412    ----------------------------------------------------------------------
3413    Name: Matthias Blume
3414    Date: 2000/07/11 13:58:00 JST
3415    Tag: blume-20000711-doctypo
3416    Description:
3417    
3418    Fixed a few typos in CM manual.
3419    
3420    ----------------------------------------------------------------------
3421    Name: Allen Leung
3422    Date: 2000/06/15 00:38:00
3423    Tag: leunga-20000704-sparc-x86
3424    
3425    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
3426    2. fix to x86 RET bug [found by Dan Grossman]
3427    3. sparc assembly bug fix for ticc instructions [found by Fermin]
3428    
3429       Affects c-- and moby only
3430    
3431    ----------------------------------------------------------------------
3432    Name: Matthias Blume
3433    Date: 2000/07/04 15:26:00
3434    Tag: blume-20000704-trigger
3435    Description:
3436    
3437    1. Improvements to CM manual.
3438    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
3439       of getting a back-trace.  The function, when called, raises an
3440       internal exception which explicitly carries the full back-trace history,
3441       so it is unaffected by any intervening handle-raise pairs ("trivial"
3442       or not).  The interactive loop will print that history once it arrives
3443       at top level.
3444       Short of having all exceptions implicitly carry the full history, the
3445       recommended way of using this facility is:
3446         - compile your program with instrumentation "on"
3447         - run it, when it raises an exception, look at the history
3448         - if the history is "cut off" because of some handler, go and modify
3449           your program so that it explicitly calls BTrace.trigger
3450         - recompile (still instrumented), and rerun; look at the full history
3451    
3452    ----------------------------------------------------------------------
3453    Name: Matthias Blume
3454    Date: 2000/07/03 15:36:00 JST
3455    Tag: blume-20000702-manual
3456    Description:
3457    
3458    Small corrections and updates to CM manual.
3459    
3460    ----------------------------------------------------------------------
3461    Name: Matthias Blume
3462    Date: 2000/06/29 16:04:00 JST
3463    Tag: blume-20000629-yacctool
3464    Description:
3465    
3466    Changes:
3467    
3468    1. Class "mlyacc" now takes separate arguments to pass options to
3469       generated .sml- and .sig-files independently.
3470    2. Corresponding CM manual updates.
3471    3. BTrace module now also reports call sites.  (However, for loop clusters
3472       it only shows from where the cluster was entered.)  There are associated
3473       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
3474    
3475    ----------------------------------------------------------------------
3476    Name: Matthias Blume
3477    Date: 2000/06/27 16:51:00 JST
3478    Tag: blume-20000627-noweb
3479    Description:
3480    
3481    Changes:
3482    
3483     1. Implemented "subdir" and "witness" options for noweb tool.
3484        This caused some slight internal changes in CM's tool implementation.
3485     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
3486        some remaining issues from earlier path anchor changes.
3487     3. Updated CM manual accordingly.
3488    
3489     4. Changed implementation of back-tracing so that I now consider it
3490        ready for prime-time.
3491    
3492        In particular, you don't have to explicitly trigger the back-trace
3493        anymore.  Instead, if you are running BTrace-instrumented code and
3494        there is an uncaught exception (regardless of whether or not it was
3495        raised in instrumented code), the top-level evalloop will print
3496        the back-trace.
3497    
3498        Features:
3499    
3500          - Instrumented and uninstrumented code work together seemlessly.
3501            (Of course, uninstrumented code is never mentioned in actual
3502             back-traces.)
3503    
3504          - Asymptotic time- and space-complexity of instrumented code is
3505            equal to that of uninstrumented code.  (This means that
3506            tail-recursion is preserved by the instrumentation phase.)
3507    
3508          - Modules whose code has been instrumented in different sessions
3509            work together without problem.
3510    
3511          - There is no penalty whatsoever on uninstrumented code.
3512    
3513          - There is no penalty on "raise" expressions, even in
3514            instrumented code.
3515    
3516        A potential bug (or perhaps it is a feature, too):
3517    
3518          A back-trace reaches no further than the outermost instrumented
3519          non-trivial "raise".  Here, a "trivial" raise is one that is the
3520          sole RHS of a "handle" rule.  Thus, back-traces reach trough
3521    
3522               <exp> handle e => raise e
3523    
3524          and even
3525    
3526               <exp> handle Foo => raise Bar
3527    
3528          and, of course, through
3529    
3530               <exp> handle Foo => ...
3531    
3532         if the exception was not Foo.
3533    
3534         Back-traces always reach right through any un-instrumented code
3535         including any of its "handle" expressions, trivial or not.
3536    
3537       To try this out, do the following:
3538    
3539         - Erase all existing binfiles for your program.
3540           (You may keep binfiles for those modules where you think you
3541            definitely don't need back-tracing.)
3542         - Turn on back-trace instrumentation:
3543              SMLofNJ.Internals.BTrace.mode (SOME true);
3544         - Recompile your program.  (I.e., run "CM.make" or "use".)
3545         - You may now turn instrumentation off again (if you want):
3546              SMLofNJ.Internals.BTrace.mode (SOME false);
3547         - Run your program as usual.  If it raises an exception that
3548           reaches the interactive toplevel, then a back-trace will
3549           automatically be printed.  After that, the toplevel loop
3550           will print the exception history as usual.
3551    
3552    ----------------------------------------------------------------------
3553    Name: Matthias Blume
3554    Date: 2000/06/26 09:56:46 JST
3555    Tag: blume-20000626-setup
3556    Description:
3557    
3558    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
3559          ML code before and after compiling a file (e.g., to set compiler
3560          flags)
3561    
3562    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
3563              - associated changes to btrace.sml (BTrace instrumentation pass)
3564              - cleaner implementation of btimp.sml (BTrace tracing and report
3565                module)
3566    
3567    CM manual: * new path encoding documented
3568               * description of setup-parameter to "sml" added
3569    
3570    The biggest user-visible change to back-tracing is that it is no
3571    longer necessary to compile all traced modules within the same
3572    session.  (This was a real limitation.)
3573    
3574  ----------------------------------------------------------------------  ----------------------------------------------------------------------
3575  Name: Matthias Blume  Name: Matthias Blume
3576  Date: 2000/06/24 12:40:00 JST  Date: 2000/06/24 12:40:00 JST
# Line 1166  Line 4726 
4726       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
4727       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
4728       Please, don't be too harsh on me for messing with this code a bit more       Please, don't be too harsh on me for messing with this code a bit more
4729       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
4730       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
4731       Dave's hands. :)       Dave's hands. :)
4732    

Legend:
Removed from v.676  
changed lines
  Added in v.1175

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