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 605, Fri Apr 7 14:06:42 2000 UTC revision 1148, Fri Mar 15 21:38:57 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: 2002/03/15 16:40:00 EST
18    Tag: blume-20020315-basis
19    Description:
20    
21    Provided (preliminary?) implementations for
22    
23      {String,Substring}.{concatWith,isSuffix,isSubstring}
24    
25    and
26    
27      Substring.full
28    
29    Those are in the Basis spec but they were missing in SML/NJ.
30    
31    ----------------------------------------------------------------------
32    Name: Matthias Blume
33    Date: 2002/03/14 21:30:00 EST
34    Tag: blume-20020314-controls
35    Description:
36    
37    Controls:
38    ---------
39    
40    1. Factored out the recently-added Controls : CONTROLS stuff and put
41       it into its own library $/controls-lib.cm.  The source tree for
42       this is under src/smlnj-lib/Controls.
43    
44    2. Changed the names of types and functions in this interface, so they
45       make a bit more "sense":
46    
47          module -> registry
48          'a registry -> 'a group
49    
50    3. The interface now deals in ref cells only.  The getter/setter interface
51       is (mostly) gone.
52    
53    4. Added a function that lets one register an already-existing ref cell.
54    
55    5. Made the corresponding modifications to the rest of the code so that
56       everything compiles again.
57    
58    6. Changed the implementation of Controls.MLRISC back to something closer
59       to the original.  In particular, this module (and therefore MLRISC)
60       does not depend on Controls.  There now is some link-time code in
61       int-sys.sml that registers the MLRISC controls with the Controls
62       module.
63    
64    CM:
65    ---
66    
67      * One can now specify the lambda-split aggressiveness in init.cmi.
68    
69    ----------------------------------------------------------------------
70    Name: Allen Leung
71    Date: 2002/03/13 17:30:00 EST
72    Tag: leunga-20020313-x86-fp-unary
73    Description:
74    
75    Bug fix for:
76    
77    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
78    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
79    > - fun f(x,(y,z)) = Real.~ y;
80    > [autoloading]
81    > [autoloading done]
82    >       fchsl   (%eax), 184(%esp)
83    > Error: MLRisc bug: X86MCEmitter.emitInstr
84    >
85    > uncaught exception Error
86    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
87    
88    The problem was that the code generator did not generate any fp registers
89    in this case, and the ra didn't know that it needed to run the X86FP phase to
90    translate the pseudo fp instruction.   This only happened with unary fp
91    operators in certain situations.
92    
93    ----------------------------------------------------------------------
94    Name: Matthias Blume
95    Date: 2002/03/13 14:00:00 EST
96    Tag: blume-20020313-overload-etc
97    Description:
98    
99    1. Added _overload as a synonym for overload for backward compatibility.
100       (Control.overloadKW must be true for either version to be accepted.)
101    
102    2. Fixed bug in install script that caused more things to be installed
103       than what was requested in config/targets.
104    
105    3. Made CM aware of the (_)overload construct so that autoloading
106       works.
107    
108    ----------------------------------------------------------------------
109    Name: Matthias Blume
110    Date: 2002/03/12 22:03:00 EST
111    Tag: blume-20020312-url
112    Description:
113    
114    Forgot to update BOOT and srcarchiveurl.
115    
116    ----------------------------------------------------------------------
117    Name: Matthias Blume
118    Date: 2002/03/12 17:30:00 EST
119    Tag: blume-20020312-version110392
120    Description:
121    
122    Yet another version number bump (because of small changes to the
123    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
124    
125    Changes:
126    
127      The new pid generation scheme described a few weeks ago was overly
128      complicated.  I implemented a new mechanism that is simpler and
129      provides a bit more "stability":  Once CM has seen a compilation
130      unit, it keeps its identity constant (as long as you do not delete
131      those crucial CM/GUID/* files).  This means that when you change
132      an interface, compile, then go back to the old interface, and
133      compile again, you arrive at the original pid.
134    
135      There now also is a mechanism that instructs CM to use the plain
136      environment hash as a module's pid (effectively making its GUID
137      the empty string).  For this, "noguid" must be specified as an
138      option to the .sml file in question within its .cm file.
139      This is most useful for code that is being generated by tools such
140      as ml-nlffigen (because during development programmers tend to
141      erase the tool's entire output directory tree including CM's cached
142      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
143      revert to the old, broken behavior of SML/NJ, but in specific cases
144      where there is no danger of interface confusion, its use is ok
145      (I think).
146    
147      ml-nlffigen by default generates "noguid" annotations.  They can be
148      turned off by specifying -guid in its command line.
149    
150    ----------------------------------------------------------------------
151    Name: Lal George
152    Date: 2002/03/12 12 14:42:36 EST
153    Tag: george-20020312-frequency-computation
154    Description:
155    
156    Integrated jump chaining and static block frequency into the
157    compiler. More details and numbers later.
158    
159    ----------------------------------------------------------------------
160    Name: Lal George
161    Date: 2002/03/11 11 22:38:53 EST
162    Tag: george-20020311-jump-chain-elim
163    Description:
164    
165    Tested the jump chain elimination on all architectures (except the
166    hppa).  This is on by default right now and is profitable for the
167    alpha and x86, however, it may not be profitable for the sparc and ppc
168    when compiling the compiler.
169    
170    The gc test will typically jump to a label at the end of the cluster,
171    where there is another jump to an external cluster containing the actual
172    code to invoke gc. This is to allow factoring of common gc invocation
173    sequences. That is to say, we generate:
174    
175            f:
176               testgc
177               ja   L1      % jump if above to L1
178    
179            L1:
180               jmp L2
181    
182    
183    After jump chain elimination the 'ja L1' instructions is converted to
184    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
185    up being implemented in their long form (if L2 is far away) using:
186    
187            jbe     L3      % jump if below or equal to L3
188            jmp     L2
189         L3:
190            ...
191    
192    
193    For large compilation units L2  may be far away.
194    
195    
196    ----------------------------------------------------------------------
197    Name: Matthias Blume
198    Date: 2002/03/11 13:30:00 EST
199    Tag: blume-20020311-mltreeeval
200    Description:
201    
202    A functor parameter was missing.
203    
204    ----------------------------------------------------------------------
205    Name: Allen Leung
206    Date: 2002/03/11 10:30:00 EST
207    Tag: leunga-20020311-runtime-string0
208    Description:
209    
210       The representation of the empty string now points to a
211    legal null terminated C string instead of unit.  It is now possible
212    to convert an ML string into C string with InlineT.CharVector.getData.
213    This compiles into one single machine instruction.
214    
215    ----------------------------------------------------------------------
216    Name: Allen Leung
217    Date: 2002/03/10 23:55:00 EST
218    Tag: leunga-20020310-x86-call
219    Description:
220    
221       Added machine generation for CALL instruction (relative displacement mode)
222    
223    ----------------------------------------------------------------------
224    Name: Matthias Blume
225    Date: 2002/03/08 16:05:00
226    Tag: blume-20020308-entrypoints
227    Description:
228    
229    Version number bumped to 110.39.1.  NEW BOOTFILES!
230    
231    Entrypoints: non-zero offset into a code object where execution should begin.
232    
233    - Added the notion of an entrypoint to CodeObj.
234    - Added reading/writing of entrypoint info to Binfile.
235    - Made runtime system bootloader aware of entrypoints.
236    - Use the address of the label of the first function given to mlriscGen
237      as the entrypoint.  This address is currently always 0, but it will
238      not be 0 once we turn on block placement.
239    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
240      with entry points) from mlriscGen.
241    
242    ----------------------------------------------------------------------
243    Name: Allen Leung
244    Date: 2002/03/07 20:45:00 EST
245    Tag: leunga-20020307-x86-cmov
246    Description:
247    
248       Bug fixes for CMOVcc on x86.
249    
250       1. Added machine code generation for CMOVcc
251       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
252       3. CMOVcc cannot have an immediate operand as argument.
253    
254    ----------------------------------------------------------------------
255    Name: Matthias Blume
256    Date: 2002/03/07 16:15:00 EST
257    Tag: blume-20020307-controls
258    Description:
259    
260    This is a very large but mostly boring patch which makes (almost)
261    every tuneable compiler knob (i.e., pretty much everything under
262    Control.* plus a few other things) configurable via both the command
263    line and environment variables in the style CM did its configuration
264    until now.
265    
266    Try starting sml with '-h' (or, if you are brave, '-H')
267    
268    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
269    implements the underlying generic mechanism.
270    
271    The interface to some of the existing such facilities has changed somewhat.
272    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
273    (The getFoo interface is still there for backward-compatibility, but its
274    use is deprecated.)
275    
276    The ml-build script passes -Cxxx=yyy command-line arguments through so
277    that one can now twiddle the compiler settings when using this "batch"
278    compiler.
279    
280    TODO items:
281    
282    We should go through and throw out all controls that are no longer
283    connected to anything.  Moreover, we should go through and provide
284    meaningful (and correct!) documentation strings for those controls
285    that still are connected.
286    
287    Currently, multiple calls to Controls.new are accepted (only the first
288    has any effect).  Eventually we should make sure that every control
289    is being made (via Controls.new) exactly once.  Future access can then
290    be done using Controls.acc.
291    
292    Finally, it would probably be a good idea to use the getter-setter
293    interface to controls rather than ref cells.  For the time being, both
294    styles are provided by the Controls module, but getter-setter pairs are
295    better if thread-safety is of any concern because they can be wrapped.
296    
297    *****************************************
298    
299    One bug fix: The function blockPlacement in three of the MLRISC
300    backpatch files used to be hard-wired to one of two possibilities at
301    link time (according to the value of the placementFlag).  But (I
302    think) it should rather sense the flag every time.
303    
304    *****************************************
305    
306    Other assorted changes (by other people who did not supply a HISTORY entry):
307    
308    1. the cross-module inliner now works much better (Monnier)
309    2. representation of weights, frequencies, and probabilities in MLRISC
310       changed in preparation of using those for weighted block placement
311       (Reppy, George)
312    
313    ----------------------------------------------------------------------
314    Name: Lal George
315    Date: 2002/03/07 14:44:24 EST 2002
316    Tag: george-20020307-weighted-block-placement
317    
318    Tested the weighted block placement optimization on all architectures
319    (except the hppa) using AMPL to generate the block and edge frequencies.
320    Changes were required in the machine properties to correctly
321    categorize trap instructions. There is an MLRISC flag
322    "weighted-block-placement" that can be used to enable weighted block
323    placement, but this will be ineffective without block/edge
324    frequencies (coming soon).
325    
326    
327    ----------------------------------------------------------------------
328    Name: Lal George
329    Date: 2002/03/05 17:24:48 EST
330    Tag: george-20020305-linkage-cluster
331    
332    In order to support the block placement optimization, a new cluster
333    is generated as the very first cluster (called the linkage cluster).
334    It contains a single jump to the 'real' entry point for the compilation
335    unit. Block placement has no effect on the linkage cluster itself, but
336    all the other clusters  have full freedom in the manner in which they
337    reorder blocks or functions.
338    
339    On the x86 the typical linkage code that is generated is:
340       ----------------------
341            .align 2
342       L0:
343            addl    $L1-L0, 72(%esp)
344            jmp     L1
345    
346    
347            .align  2
348       L1:
349       ----------------------
350    
351    72(%esp) is the memory location for the stdlink register. This
352    must contain the address of the CPS function being called. In the
353    above example, it contains the address of  L0; before
354    calling L1 (the real entry point for the compilation unit), it
355    must contain the address for L1, and hence
356    
357            addl $L1-L0, 72(%esp)
358    
359    I have tested this on all architectures except the hppa.The increase
360    in code size is of course negligible
361    
362    ----------------------------------------------------------------------
363    Name: Allen Leung
364    Date: 2002/03/03 13:20:00 EST
365    Tag: leunga-20020303-mlrisc-tools
366    
367      Added #[ ... ] expressions to mlrisc tools
368    
369    ----------------------------------------------------------------------
370    Name: Matthias Blume
371    Date: 2002/02/27 12:29:00 EST
372    Tag: blume-20020227-cdebug
373    Description:
374    
375    - made types in structure C and C_Debug to be equal
376    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
377    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
378    
379    ----------------------------------------------------------------------
380    Name: Matthias Blume
381    Date: 2002/02/26 12:00:00 EST
382    Tag: blume-20020226-ffi
383    Description:
384    
385    1. Fixed a minor bug in CM's "noweb" tool:
386       If numbering is turned off, then truly don't number (i.e., do not
387       supply the -L option to noweb).  The previous behavior was to supply
388       -L'' -- which caused noweb to use the "default" line numbering scheme.
389       Thanks to Chris Richards for pointing this out (and supplying the fix).
390    
391    2. Once again, I reworked some aspects of the FFI:
392    
393       A. The incomplete/complete type business:
394    
395       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
396         gone!
397       - ML types representing an incomplete type are now *equal* to
398         ML types representing their corresponding complete types (just like
399         in C).  This is still safe because ml-nlffigen will not generate
400         RTTI for incomplete types, nor will it generate functions that
401         require access to such RTTI.   But when ML code generated from both
402         incomplete and complete versions of the C type meet, the ML types
403         are trivially interoperable.
404    
405         NOTE:  These changes restore the full generality of the translation
406         (which was previously lost when I eliminated functorization)!
407    
408       B. Enum types:
409    
410       - Structure C now has a type constructor "enum" that is similar to
411         how the "su" constructor works.  However, "enum" is not a phantom
412         type because each "T enum" has values (and is isomorphic to
413         MLRep.Signed.int).
414       - There are generic access operations for enum objects (using
415         MLRep.Signed.int).
416       - ml-nlffigen will generate a structure E_foo for each "enum foo".
417         * The structure contains the definition of type "mlrep" (the ML-side
418         representation type of the enum).  Normally, mlrep is the same
419         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
420         then mlrep will be defined as a datatype -- thus facilitating
421         pattern matching on mlrep values.
422         ("-ec" will be suppressed if there are duplicate values in an
423          enumeration.)
424         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
425         will be generated for each C enum constant xxx.
426         * Conversion functions m2i and i2m convert between mlrep and
427         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
428         * Coversion functions c and ml convert between mlrep and "tag enum".
429         * Access functions (get/set) fetch and store mlrep values.
430       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
431         enumerations are merged into one single enumeration represented by
432         structure E_'.
433    
434    ----------------------------------------------------------------------
435    Name: Allen Leung
436    Date: 2002/02/25 04:45:00 EST
437    Tag: leunga-20020225-cps-spill
438    
439    This is a new implementation of the CPS spill phase.
440    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
441    In case of problems, replace it with the old file spill.sml
442    
443    The current compiler runs into some serious performance problems when
444    constructing a large record.  This can happen when we try to compile a
445    structure with many items.  Even a very simple structure like the following
446    makes the compiler slow down.
447    
448        structure Foo = struct
449           val x_1 = 0w1 : Word32.int
450           val x_2 = 0w2 : Word32.int
451           val x_3 = 0w3 : Word32.int
452           ...
453           val x_N = 0wN : Word32.int
454        end
455    
456    The following table shows the compile time, from N=1000 to N=4000,
457    with the old compiler:
458    
459    N
460    1000   CPS 100 spill                           0.04u  0.00s  0.00g
461           MLRISC ra                               0.06u  0.00s  0.05g
462              (spills = 0 reloads = 0)
463           TOTAL                                   0.63u  0.07s  0.21g
464    
465    1100   CPS 100 spill                           8.25u  0.32s  0.64g
466           MLRISC ra                               5.68u  0.59s  3.93g
467              (spills = 0 reloads = 0)
468           TOTAL                                   14.71u  0.99s  4.81g
469    
470    1500   CPS 100 spill                           58.55u  2.34s  1.74g
471           MLRISC ra                               5.54u  0.65s  3.91g
472              (spills = 543 reloads = 1082)
473           TOTAL                                   65.40u  3.13s  6.00g
474    
475    2000   CPS 100 spill                           126.69u  4.84s  3.08g
476           MLRISC ra                               0.80u  0.10s  0.55g
477              (spills = 42 reloads = 84)
478           TOTAL                                   129.42u  5.10s  4.13g
479    
480    3000   CPS 100 spill                           675.59u  19.03s  11.64g
481           MLRISC ra                               2.69u  0.27s  1.38g
482              (spills = 62 reloads = 124)
483           TOTAL                                   682.48u  19.61s  13.99g
484    
485    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
486           MLRISC ra                               4.96u  0.27s  2.72g
487              (spills = 85 reloads = 170)
488           TOTAL                                   2375.26u  57.21s  48.00g
489    
490    As you can see the old cps spill module suffers from some serious
491    performance problem.  But since I cannot decipher the old code fully,
492    instead of patching the problems up, I'm reimplementing it
493    with a different algorithm.  The new code is more modular,
494    smaller when compiled, and substantially faster
495    (O(n log n) time and O(n) space).  Timing of the new spill module:
496    
497    4000  CPS 100 spill                           0.02u  0.00s  0.00g
498          MLRISC ra                               0.25u  0.02s  0.15g
499             (spills=1 reloads=3)
500          TOTAL                                   7.74u  0.34s  1.62g
501    
502    Implementation details:
503    
504    As far as I can tell, the purpose of the CPS spill module is to make sure the
505    number of live variables at any program point (the bandwidth)
506    does not exceed a certain limit, which is determined by the
507    size of the spill area.
508    
509    When the bandwidth is too large, we decrease the register pressure by
510    packing live variables into spill records.  How we achieve this is
511    completely different than what we did in the old code.
512    
513    First, there is something about the MLRiscGen code generator
514    that we should be aware of:
515    
516    o MLRiscGen performs code motion!
517    
518       In particular, it will move floating point computations and
519       address computations involving only the heap pointer to
520       their use sites (if there is only a single use).
521       What this means is that if we have a CPS record construction
522       statement
523    
524           RECORD(k,vl,w,e)
525    
526       we should never count the new record address w as live if w
527       has only one use (which is often the case).
528    
529       We should do something similar to floating point, but the transformation
530       there is much more complex, so I won't deal with that.
531    
532    Secondly, there are now two new cps primops at our disposal:
533    
534     1. rawrecord of record_kind option
535        This pure operator allocates some uninitialized storage from the heap.
536        There are two forms:
537    
538         rawrecord NONE [INT n]  allocates a tagless record of length n
539         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
540                                     and initializes the tag.
541    
542     2. rawupdate of cty
543          rawupdate cty (v,i,x)
544          Assigns to x to the ith component of record v.
545          The storelist is not updated.
546    
547    We use these new primops for both spilling and increment record construction.
548    
549     1. Spilling.
550    
551        This is implemented with a linear scan algorithm (but generalized
552        to trees).  The algorithm will create a single spill record at the
553        beginning of the cps function and use rawupdate to spill to it,
554        and SELECT or SELp to reload from it.  So both spills and reloads
555        are fine-grain operations.  In contrast, in the old algorithm
556        "spills" have to be bundled together in records.
557    
558        Ideally, we should sink the spill record construction to where
559        it is needed.  We can even split the spill record into multiple ones
560        at the places where they are needed.  But CPS is not a good
561        representation for global code motion, so I'll keep it simple and
562        am not attempting this.
563    
564     2. Incremental record construction (aka record splitting).
565    
566        Long records with many component values which are simulatenously live
567        (recall that single use record addresses are not considered to
568         be live) are constructed with rawrecord and rawupdate.
569        We allocate space on the heap with rawrecord first, then gradually
570        fill it in with rawupdate.  This is the technique suggested to me
571        by Matthias.
572    
573        Some restrictions on when this is applicable:
574        1. It is not a VECTOR record.  The code generator currently does not handle
575           this case. VECTOR record uses double indirection like arrays.
576        2. All the record component values are defined in the same "basic block"
577           as the record constructor.  This is to prevent speculative
578           record construction.
579    
580    ----------------------------------------------------------------------
581    Name: Allen Leung
582    Date: 2002/02/22 01:02:00 EST
583    Tag: leunga-20020222-mlrisc-tools
584    
585    Minor bug fixes in the parser and rewriter
586    
587    ----------------------------------------------------------------------
588    Name: Allen Leung
589    Date: 2002/02/21 20:20:00 EST
590    Tag: leunga-20020221-peephole
591    
592    Regenerated the peephole files.  Some contained typos in the specification
593    and some didn't compile because of pretty printing bugs in the old version
594    of 'nowhere'.
595    
596    ----------------------------------------------------------------------
597    Name: Allen Leung
598    Date: 2002/02/19 20:20:00 EST
599    Tag: leunga-20020219-mlrisc-tools
600    Description:
601    
602       Minor bug fixes to the mlrisc-tools library:
603    
604       1.  Fixed up parsing colon suffixed keywords
605       2.  Added the ability to shut the error messages up
606       3.  Reimplemented the pretty printer and fixed up/improved
607           the pretty printing of handle and -> types.
608       4.  Fixed up generation of literal symbols in the nowhere tool.
609       5.  Added some SML keywords to to sml.sty
610    
611    ----------------------------------------------------------------------
612    Name: Matthias Blume
613    Date: 2002/02/19 16:20:00 EST
614    Tag: blume-20020219-cmffi
615    Description:
616    
617    A wild mix of changes, some minor, some major:
618    
619    * All C FFI-related libraries are now anchored under $c:
620        $/c.cm      --> $c/c.cm
621        $/c-int.cm  --> $c/internals/c-int.cm
622        $/memory.cm --> $c/memory/memory.cm
623    
624    * "make" tool (in CM) now treats its argument pathname slightly
625      differently:
626        1. If the native expansion is an absolute name, then before invoking
627           the "make" command on it, CM will apply OS.Path.mkRelative
628           (with relativeTo = OS.FileSys.getDir()) to it.
629        2. The argument will be passed through to subsequent phases of CM
630           processing without "going native".  In particular, if the argument
631           was an anchored path, then "make" will not lose track of that anchor.
632    
633    * Compiler backends now "know" their respective C calling conventions
634      instead of having to be told about it by ml-nlffigen.  This relieves
635      ml-nlffigen from one of its burdens.
636    
637    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
638    
639    * Export C_DEBUG and C_Debug from $c/c.cm.
640    
641    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
642      subtyping relationship between incomplete pointers and their complete
643      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
644      with the convention of instantiating 'o with ('t, 'c) obj whenever
645      the pointer target type is complete.  In the incomplete case, 'o
646      will be instantiated with some "'c iobj" -- a type obtained by
647      using one of the functors PointerToIncompleteType or PointerToCompleteType.
648    
649      Operations that work on both incomplete and complete pointer types are
650      typed as taking an 'o ptr while operations that require the target to
651      be known are typed as taking some ('t, 'c) obj ptr.
652    
653      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
654      where void is an eqtype without any values.  This makes it possible
655      to work on voidptr values using functions meant to operate on light
656      incomplete pointers.
657    
658    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
659      been vastly simplified.
660    
661    ----------------------------------------------------------------------
662    Name: Matthias Blume
663    Date: 2002/02/19 10:48:00 EST
664    Tag: blume-20020219-pqfix
665    Description:
666    
667    Applied Chris Okasaki's bug fix for priority queues.
668    
669    ----------------------------------------------------------------------
670    Name: Matthias Blume
671    Date: 2002/02/15 17:05:00
672    Tag: Release_110_39
673    Description:
674    
675    Last-minute retagging is becoming a tradition... :-(
676    
677    This is the working release 110.39.
678    
679    ----------------------------------------------------------------------
680    Name: Matthias Blume
681    Date: 2002/02/15 16:00:00 EST
682    Tag: Release_110_39-orig
683    Description:
684    
685    Working release 110.39.  New bootfiles.
686    
687    (Update: There was a small bug in the installer so it wouldn't work
688    with all shells.  So I retagged. -Matthias)
689    
690    ----------------------------------------------------------------------
691    Name: Matthias Blume
692    Date: 2002/02/15 14:17:00 EST
693    Tag: blume-20020215-showbindings
694    Description:
695    
696    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
697    the latter can be useful for exploring what bindings are available at
698    the interactive prompt.  (The first function returns only the list
699    of symbols that are really bound, the second prints those but also the
700    ones that CM's autoloading mechanism knows about.)
701    
702    ----------------------------------------------------------------------
703    Name: Matthias Blume
704    Date: 2002/02/15 12:08:00 EST
705    Tag: blume-20020215-iptrs
706    Description:
707    
708    Two improvements to ml-nlffigen:
709    
710      1. Write files only if they do not exist or if their current contents
711         do not coincide with what's being written.  (That is, avoid messing
712         with the time stamps unless absolutely necessary.)
713    
714      2. Implement a "repository" mechanism for generated files related
715         to "incomplete pointer types".   See the README file for details.
716    
717    ----------------------------------------------------------------------
718    Name: Matthias Blume
719    Date: 2002/02/14 11:50:00 EST
720    Tag: blume-20020214-quote
721    Description:
722    
723    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
724    because of the new and improved tag generation scheme.  (Thanks to Allen
725    Leung for pointing it out.)
726    
727    ----------------------------------------------------------------------
728    Name: Lal George
729    Date: 2002/02/14 09:55:27 EST 2002
730    Tag: george-20020214-isabelle-bug
731    Description:
732    
733    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
734    of Isabelle on the x86.
735    
736    From Allen:
737    -----------
738     I've found the problem:
739    
740         in ra-core.sml, I use the counter "blocked" to keep track of the
741         true number of elements in the freeze queue.  When the counter goes
742         to zero, I skip examining the queue.  But I've messed up the
743         bookkeeping in combine():
744    
745             else ();
746             case !ucol of
747               PSEUDO => (if !cntv > 0 then
748                     (if !cntu > 0 then blocked := !blocked - 1 else ();
749                                        ^^^^^^^^^^^^^^^^^^^^^^^
750                      moveu := mergeMoveList(!movev, !moveu)
751                     )
752                  else ();
753    
754         combine() is called to coalesce two nodes u and v.
755         I think I was thinking that if the move counts of u and v are both
756         greater than zero then after they are coalesced then one node is
757         removed from the freeze queue.  Apparently I was thinking that
758         both u and v are of low degree, but that's clearly not necessarily true.
759    
760    
761    02/12/2002:
762        Here's the patch.  HOL now compiles.
763    
764        I don't know how this impact on performance (compile
765        time or runtime).  This bug caused the RA (especially on the x86)
766        to go thru the potential spill phase when there are still nodes on the
767        freeze queue.
768    
769    
770    
771    
772    ----------------------------------------------------------------------
773    Name: Matthias Blume
774    Date: 2002/02/13 22:40:00 EST
775    Tag: blume-20020213-fptr-rtti
776    Description:
777    
778    Fixed a bug in ml-nlffigen that was introduced with one of the previous
779    updates.
780    
781    ----------------------------------------------------------------------
782    Name: Matthias Blume
783    Date: 2002/02/13 16:41:00 EST
784    Tag: blume-20020213-cmlpq
785    Description:
786    
787    Added new priority queue export symbols (which have just been added to
788    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
789    not compile and the installer would choke.)
790    
791    ----------------------------------------------------------------------
792    Name: Matthias Blume
793    Date: 2002/02/13 16:15:00 EST
794    Tag: blume-20020213-various
795    Description:
796    
797    1. More tweaks to ml-nlffigen:
798    
799       - better internal datastructures (resulting in slight speedup)
800       - "-match" option requires exact match
801       - "localized" gensym counters (untagged structs/unions nested within
802         other structs/unions or within typedefs get a fresh counter; their
803         tag will be prefixed by a concatenation of their parents' tags)
804       - bug fixes (related to calculation of transitive closure of types
805         to be included in the output)
806    
807    2. Minor Basis updates:
808    
809       - added implementations for List.collate and Option.app
810    
811    ----------------------------------------------------------------------
812    Name: Matthias Blume
813    Date: 2002/02/11 15:55:00 EST
814    Tag: blume-20020211-gensym
815    Description:
816    
817    Added a "-gensym" option to command line of ml-nlffigen.  This can be
818    used to specify a "stem" -- a string that is inserted in all "gensym'd"
819    names (ML structure names that correspond to unnamed C structs, unions,
820    and enums), so that separate runs of ml-nlffigen do not clash.
821    
822    ----------------------------------------------------------------------
823    Name: Matthias Blume
824    Date: 2002/02/11 12:05:00 EST
825    Tag: blume-20020211-gensml
826    Description:
827    
828    A quick fix for a problem with GenSML (in the pgraph-util library):
829    Make generation of toplevel "local" optional.  (Strictly speaking,
830    signature definitions within "local" are not legal SML.)
831    
832    Other than that: updates to INSTALL and cm/TODO.
833    
834    ----------------------------------------------------------------------
835    Name: Matthias Blume
836    Date: 2002/02/08 15:00:00 EST
837    Tag: blume-20020208-uniquepid
838    Description:
839    
840    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
841    
842    1. The installer (config/install.sh) has gotten smarter:
843    
844         - Configuration options are a bit easier to specify now
845           (in config/targets).
846         - Bug in recognizing .tar.bz2 files fixed.
847         - Installer automatically resolves dependencies between
848           configuration options (e.g., if you ask for eXene, you will
849           also get cml -- regardless whether you asked for it or not).
850         - Installer can run in "quieter mode" by setting the environment
851           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
852           "completely silent", though.
853         - Build HashCons library as part of smlnj-lib.
854    
855    2. A new scheme for assigning persistent identifiers to compilation
856       units (and, by extension, to types etc.) has been put into place.
857       This fixes a long-standing bug where types and even dynamic values
858       can get internally confused, thereby compromising type safety
859       (abstraction) and dynamic correctness.  See
860    
861         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
862    
863       for an example of how things could go wrong until now.
864    
865       The downside of the new scheme is that pids are not quite as
866       persistent as they used to be: CM will generate a fresh pid
867       for every compilation unit that it thinks it sees for the first
868       time.  That means that if you compile starting from a clean, fresh
869       source tree at two different times, you end up with different
870       binaries.
871    
872       Cutoff recompilation, however, has not been compromised because
873       CM keeps pid information in special caches between runs.
874    
875    ----------------------------------------------------------------------
876    Name: Lal George
877    Date: 2002/02/07 15:34:13 EST 2002
878    Tag: <none>
879    Description:
880    
881    Compilers that generate assembly code may produce  global labels
882    whose value is resolved at link time. The various peephole optimization
883    modules did not take this in account.
884    
885    TODO. The Labels.addrOf function should really return an option
886    type so that clients are forced to deal with this issue, rather
887    than an exception being raised.
888    
889    ----------------------------------------------------------------------
890    Name: Lal George
891    Date: 2002/02/06 13:55:02 EST
892    Tag: george-20020206-ra-breakup
893    Description:
894    
895    1. A bug fix from Allen.
896    
897        A typo causes extra fstp %st(0)'s to be generated at compensation
898        edges, which might cause stack underflow traps at runtime.  This
899        occurs in fft where there are extraneous fstps right before the 'into'
900        trap instruction (in this case they are harmless since none of the
901        integers overflow.)
902    
903    2. Pulled out various utility modules that were embedded in the modules
904       of the register allocator. I need these modules for other purposes, but
905       they are not complete enough to put into a library (just yet).
906    ----------------------------------------------------------------------
907    Name: Matthias Blume
908    Date: 2002/01/31 16:05:00 EST
909    Tag: blume-20020131-sparc-ccalls
910    Description:
911    
912    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
913       of extra stack space by mistake.  Fixed.
914    
915    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
916    
917    ----------------------------------------------------------------------
918    Name: Allen Leung
919    Date: 2002/01/30
920    Tag: leunga-20020130-nowhere-bug-fix
921    Description:
922    
923       MLRISC bug fixes:
924       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
925       2. MachineInt.fromString was negating its value.
926    
927    ----------------------------------------------------------------------
928    Name: Matthias Blume
929    Date: 2002/01/29
930    Tag: blume-20020129-INSTALL
931    Description:
932    
933    - Added somewhat detailed installation instructions (file INSTALL).
934    - Fixed curl-detection bug in config/install.sh.
935    - It is now possible to select the URL getter using the URLGETTER
936      environment variable:
937    
938          not set / "unknown"      --> automatic detection (script tries wget,
939                                       curl, and lynx)
940          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
941                                       how to properly invoke them)
942          other                    --> use $URLGETTER directly, it must take
943                                       precisely two command-line arguments
944                                       (source URL and destination file name)
945    
946    ----------------------------------------------------------------------
947    Name: Matthias Blume
948    Date: 2002/01/28
949    Tag: blume-20020128-sparc-ccalls
950    Description:
951    
952    - Fixed problem with calculation of "used" registers in sparc-c-calls.
953    - Make use of the allocParam argument in sparc-c-calls.
954    
955    ----------------------------------------------------------------------
956    Name: Matthias Blume
957    Date: 2002/01/28
958    Tag: blume-20020128-allocParam
959    Description:
960    
961    John Reppy:  Changes c-calls API to accept client-callback for
962    allocating extra stack space.
963    me: Corresponding changes to mlriscGen (using a dummy argument that
964        does not change the current behavior).
965    
966    ----------------------------------------------------------------------
967    Name: Matthias Blume
968    Date: 2002/01/28 12:00:00
969    Tag: Release_110_38
970    Description:
971    
972    This time for real!!!
973    
974    ----------------------------------------------------------------------
975    Name: Matthias Blume
976    Date: 2002/01/28 10:56:00 EST
977    Tag: blume-20020128-retraction
978    Description:
979    
980    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
981       with blume-Release_110_38-retracted.)
982    
983    1. Fixed a problem with incorrect rounding modes in real64.sml.
984       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
985    
986    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
987       and enums fixed.  The naming of corresponding ML identifiers should
988       now be consistent again.
989    
990    ----------------------------------------------------------------------
991    Name: Allen Leung
992    Date: 2002/01/27
993    Tag: leunga-20020127-nowhere
994    Description:
995    
996       Added a target called nowhere in the configuration scripts.
997       Enabling this will build the MLRISC 'nowhere' tool (for translating
998       programs with where-clauses into legal SML code) during installation.
999    
1000    ----------------------------------------------------------------------
1001    Name: Matthias Blume
1002    Date: 2002/01/25 21:27:00 EST
1003    Tag: blume-Release_110_38-retracted
1004    Description:
1005    
1006    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
1007    
1008    README will be added later.
1009    
1010    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
1011    (Release_110_38) removed.  Reason: Last-minute bug fixes.
1012    
1013    ----------------------------------------------------------------------
1014    Name: Matthias Blume
1015    Date: 2002/01/25
1016    Tag: blume-20020125-ffi
1017    Description:
1018    
1019    A large number of tweaks and improvements to ml-nlffi-lib and
1020    ml-nlffigen:
1021    
1022       - ML represenation types have been streamlined
1023       - getter and setter functions work with concrete values, not abstract
1024         ones where possible
1025       - ml-nlffigen command line more flexible (see README file there)
1026       - some bugs have been fixed (hopefully)
1027    
1028    ----------------------------------------------------------------------
1029    Name: Lal George
1030    Date: 2002/01/24
1031    Tag: george-20020124-risc-ra-interface
1032    Description:
1033    
1034       There is a dramatic simplification in the interface to the
1035       register allocator for RISC architectures as a result of making
1036       parallel copy instructions explicit.
1037    
1038    ----------------------------------------------------------------------
1039    Name: Matthias Blume
1040    Date: 2002/01/22
1041    Tag: blume-20020122-x86-ccalls
1042    Description:
1043    
1044    Bug fix for c-calls on x86 (having to do with how char- and
1045    short-arguments are being handled).
1046    
1047    ----------------------------------------------------------------------
1048    Name: Matthias Blume
1049    Date: 2002/01/21
1050    Tag: blume-20020121-ff
1051    Description:
1052    
1053    Another day of fiddling with the FFI...
1054    
1055    1. Bug fix/workaround:  CKIT does not complain about negative array
1056       dimensions, so ml-nlffigen has to guard itself against this possibility.
1057       (Otherwise a negative dimension would send it into an infinite loop.)
1058    
1059    2. Some of the abstract types (light objects, light pointers, most "base"
1060       types) in structure C are now eqtypes.
1061    
1062    3. Added constructors and test functions for NULL function pointers.
1063    
1064    ----------------------------------------------------------------------
1065    Name: Matthias Blume
1066    Date: 2002/01/18
1067    Tag: blume-20020118-ready-for-new-release
1068    Description:
1069    
1070    Made config/srcarchiveurl point to a new place.  (Will provide boot
1071    files shortly.)
1072    
1073    Maybe we christen this to be 110.38?
1074    
1075    ----------------------------------------------------------------------
1076    Name: Matthias Blume
1077    Date: 2002/01/18
1078    Tag: blume-20020118-more-ffifiddle
1079    Description:
1080    
1081    Today's FFI fiddling:
1082    
1083      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
1084        C.Set.* that use concrete (MLRep.*) arguments and results instead
1085        of abstract ones.
1086    
1087      - Provided word-style bit operations etc. for "int" representation
1088        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
1089        and Long.
1090    
1091    ----------------------------------------------------------------------
1092    Name: Matthias Blume
1093    Date: 2002/01/18
1094    Tag: blume-20020118-use-x86-fp
1095    Description:
1096    
1097    Now that x86-fast-fp seems to be working, I turned it back on again
1098    by default.  (Seems to work fine now, even with the FFI.)
1099    
1100    Other than that, I added some documentation about the FFI to
1101    src/ml-nlffigen/README and updated the FFI test examples in
1102    src/ml-nlffi-lib/Tests/*.
1103    
1104    ----------------------------------------------------------------------
1105    Name: Allen Leung
1106    Date: 2002/01/17
1107    Tag: leunga-20020117-x86-fast-fp-call
1108    Description:
1109    
1110       1. Fixed a problem with handling return fp values when x86's fast fp
1111          mode is turned on.
1112    
1113       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
1114          of %f32.
1115    
1116       3. Added a constructor INT32lit to the ast of MLRISC tools.
1117    
1118    ----------------------------------------------------------------------
1119    Name: Matthias Blume
1120    Date: 2002/01/16
1121    Tag: blume-20020116-ffifiddle
1122    Description:
1123    
1124    More fiddling with the FFI interface:
1125    
1126     - Make constness 'c instead of rw wherever possible.  This eliminates
1127       the need for certain explicit coercions.  (However, due to ML's
1128       value polymorphism, there will still be many cases where explicit
1129       coercions are necessary.  Phantom types are not the whole answer
1130       to modeling a subtyping relationship in ML.)
1131    
1132     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
1133    
1134     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
1135       clumsy workaround for varargs problem.)
1136    
1137    ----------------------------------------------------------------------
1138    Name: Lal George
1139    Date: 2002/01/15
1140    Tag: <none>
1141    Description:
1142    
1143    1. Since COPY instructions are no longer native to the architecture,
1144       a generic functor can be used to implement the expandCopies function.
1145    
1146    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
1147       TEXT segment.
1148    
1149    ----------------------------------------------------------------------
1150    Name: Matthias Blume
1151    Date: 2002/01/15
1152    Tag: blume-20020115-ffiupdates
1153    Description:
1154    
1155    1. Fix for bug resulting in single-precision float values being returned
1156       incorrectly from FFI calls.
1157    
1158    2. Small modifications to C FFI API:
1159    
1160        - memory-allocation routines return straight objects (no options)
1161          and raise an exception in out-of-memory situations
1162        - unsafe extensions to cast between function pointers and pointers
1163          from/to ints
1164        - added structure C_Debug as an alternative to structure C where
1165          pointer-dereferencing (|*| and |*!) always check for null-pointers
1166        - added open_lib' to DynLinkage;  open_lib' works like open_lib
1167          but also takes a (possibly empty) list of existing library handles
1168          that the current library depends on
1169    
1170    ----------------------------------------------------------------------
1171    Name: Matthias Blume
1172    Date: 2002/01/10
1173    Tag: blume-20020110-newffigen
1174    Description:
1175    
1176    1. Updates to portable graph code.
1177    
1178    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
1179       more scalable now so that even huge interfaces such as the one
1180       for GTK compile in finite time and space. :-)
1181       See src/ml-nlffigen/README for details on what's new.
1182    
1183    ----------------------------------------------------------------------
1184    Name: Lal George
1185    Date: 2001/01/09 14:31:35 EST 2002
1186    Tag: george-20011206-rm-native-copy
1187    Description:
1188    
1189            Removed the native COPY and FCOPY instructions
1190            from all the architectures and replaced it with the
1191            explicit COPY instruction from the previous commit.
1192    
1193            It is now possible to simplify many of the optimizations
1194            modules that manipulate copies. This has not been
1195            done in this change.
1196    
1197    ----------------------------------------------------------------------
1198    Name: Lal George
1199    Date: 2001/12/06 16:50:13 EST 2001
1200    Tag: george-20011206-mlrisc-instruction
1201    Description:
1202    
1203    Changed the representation of instructions from being fully abstract
1204    to being partially concrete. That is to say:
1205    
1206      from
1207            type instruction
1208    
1209      to
1210            type instr                              (* machine instruction *)
1211    
1212            datatype instruction =
1213                LIVE of {regs: C.cellset, spilled: C.cellset}
1214              | KILL of {regs: C.cellset, spilled: C.cellset}
1215              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
1216              | ANNOTATION of {i: instruction, a: Annotations.annotation}
1217              | INSTR of instr
1218    
1219    This makes the handling of certain special instructions that appear on
1220    all architectures easier and uniform.
1221    
1222    LIVE and KILL say that a list of registers are live or killed at the
1223    program point where they appear. No spill code is generated when an
1224    element of the 'regs' field is spilled, but the register is moved to
1225    the 'spilled' (which is present, more for debugging than anything else).
1226    
1227    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
1228    We used to generate:
1229    
1230            DEFFREG f1
1231            f1 := f2 + f3
1232            trapb
1233    
1234    but now generate:
1235    
1236            f1 := f2 + f3
1237            trapb
1238            LIVE {regs=[f1,f2,f3], spilled=[]}
1239    
1240    Furthermore, the DEFFREG (hack) required that all floating point instruction
1241    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
1242    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
1243    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
1244    this architecture).
1245    
1246    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
1247    all the architectures. This will result in further simplification of the
1248    register allocator that must be aware of them for coalescing purposes, and
1249    will also simplify certain aspects of the machine description that provides
1250    callbacks related to parallel copies.
1251    
1252    ANNOTATION should be obvious, and now INSTR represents the honest to God
1253    machine instruction set!
1254    
1255    The <arch>/instructions/<arch>Instr.sml files define certain utility
1256    functions for making porting easier -- essentially converting upper case
1257    to lower case. All machine instructions (of type instr) are in upper case,
1258    and the lower case form generates an MLRISC instruction. For example on
1259    the alpha we have:
1260    
1261      datatype instr =
1262         LDA of {r:cell, b:cell, d:operand}
1263       | ...
1264    
1265      val lda : {r:cell, b:cell, d:operand} -> instruction
1266        ...
1267    
1268    where lda is just (INSTR o LDA), etc.
1269    
1270    ----------------------------------------------------------------------
1271    Name: Matthias Blume
1272    Date: 2001/11/22 21:40:00 EST
1273    Tag: Release_110_37
1274    Description:
1275    
1276    Release 110.37.  This time for real.
1277    
1278    ----------------------------------------------------------------------
1279    Name: Matthias Blume
1280    Date: 2001/11/21 16:35:00 EST
1281    Tag: blume-20011121-foot-in-mouth
1282    Description:
1283    
1284    Removed the "Release_110_37" tag because of a serious bug.
1285    This will be re-tagged once the bug is fixed.
1286    
1287    ----------------------------------------------------------------------
1288    Name: Matthias Blume
1289    Date: 2001/11/21 16:14:00 EST
1290    Tag: blume-20011121-forgottenfile
1291    Description:
1292    
1293    Forgot to add a file.  (Just a .tex-file -- part of
1294    the CM manual source.)
1295    
1296    ----------------------------------------------------------------------
1297    Name: Matthias Blume
1298    Date: 2001/11/21 16:10:00 EST
1299    Tag: blume-20011121-invalid_110_37
1300    Description:
1301    
1302    Note: I removed the original tag "Release_110_37" from this commit
1303          because we found a serious bug in all non-x86 backends.
1304          - Matthias
1305    
1306    1. Modifications to the SML/NJ code generator and to the runtime system
1307       so that code object name strings are directly inserted into code
1308       objects at code generation time.  The only business the runtime system
1309       has with this is now to read the name strings on occasions.
1310       (The encoding of the name string has also changed somewhat.)
1311    
1312    2. CM now implements a simple "set calculus" for specifying export lists.
1313       In particular, it is now possible to refer to the export lists of
1314       other libraries/groups/sources and form unions as well as differences.
1315       See the latest CM manual for details.
1316    
1317    3. An separate notion of "proxy" libraries has again be eliminated from
1318       CM's model.  (Proxy libraries are now simply a special case of using
1319       the export list calculus.)
1320    
1321    4. Some of the existing libraries now take advantage of the new set
1322       calculus.
1323       (Notice that not all libraries have been converted because some
1324       of the existing .cm-files are supposed to be backward compatible
1325       with 110.0.x.)
1326    
1327    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
1328       "General.exnMessage"!  The former relies on a certain hook to be
1329       initialized, and that often does not happen in the stand-alone case.)
1330    
1331    ----------------------------------------------------------------------
1332    Name: Lal George
1333    Date: 2001/11/21  13:56:18 EST
1334    Tag: george-2001121-pseudo-ops
1335    Description:
1336    
1337      Implemented a complete redesign of MLRISC pseudo-ops. Now there
1338      ought to never be any question of incompatabilities with
1339      pseudo-op syntax expected by host assemblers.
1340    
1341      For now, only modules supporting GAS syntax are implemented
1342      but more should follow, such as MASM, and vendor assembler
1343      syntax, e.g. IBM as, Sun as, etc.
1344    
1345    ----------------------------------------------------------------------
1346    Name: Matthias Blume
1347    Date: 2001/11/14 11:52:00 EST
1348    Tag: blume-20011114-srcname
1349    Description:
1350    
1351    1. Routed the name of the current source file to mlriscgen where it
1352       should be directly emitted into the code object.  (This last part
1353       is yet to be done.)
1354    
1355    2. Some cleanup of the pgraph code to make it match the proposal that
1356       I put out the other day.  (The proposal notwithstanding, things are
1357       still in flux here.)
1358    
1359    ----------------------------------------------------------------------
1360    Name: Lal George
1361    Date: 2001/11/14 09:44:04 EST
1362    Tag:
1363    Description:
1364    
1365      Fix for a backpatching bug reported by Allen.
1366    
1367      Because the boundary between short and long span-dependent
1368      instructions is +/- 128, there are an astounding number of
1369      span-dependent instructions whose size is over estimated.
1370    
1371      Allen came up with the idea of letting the size of span
1372      dependent instructions be non-monotonic, for a maxIter
1373      number of times, after which the size must be monotonically
1374      increasing.
1375    
1376      This table shows the number of span-dependent instructions
1377      whose size was over-estimated as a function of maxIter, for the
1378      file Parse/parse/ml.grm.sml:
1379    
1380         maxIter            # of instructions:
1381            10                      687
1382            20                      438
1383            30                      198
1384            40                        0
1385    
1386      In compiling the compiler, there is no significant difference in
1387      compilation speed between maxIter=10 and maxIter=40. Actually,
1388      my measurements showed that maxIter=40 was a tad faster than
1389      maxIter=10! Also 96% of the  files in the compiler reach a fix
1390      point within 13 iterations, so fixing maxIter at 40, while high,
1391      is okay.
1392    
1393    ----------------------------------------------------------------------
1394    Name: Matthias Blume
1395    Date: 2001/10/31 15:25:00 EST
1396    Tag: blume-20011031-pgraph
1397    Description:
1398    
1399    CKIT:
1400    * Changed the "Function" constructor of type Ast.ctype to carry optional
1401      argument identifiers.
1402    * Changed the return type of TypeUtil.getFunction accordingly.
1403    * Type equality ignores the argument names.
1404    * TypeUtil.composite tries to preserve argument names but gives up quickly
1405      if there is a mismatch.
1406    
1407    installation script:
1408    * attempts to use "curl" if available (unless "wget" is available as well)
1409    
1410    CM:
1411    * has an experimental implementation of "portable graphs" which I will
1412      soon propose as an implementation-independent library format
1413    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
1414    
1415    NLFFI-LIB:
1416    * some cleanup (all cosmetic)
1417    
1418    NLFFIGEN:
1419    * temporarily disabled the mechanism that suppresses ML output for
1420      C definitions whose identifiers start with an underscore character
1421    * generate val bindings for enum constants
1422    * user can request that only one style (light or heavy) is being used;
1423      default is to use both (command-line arguments: -heavy and -light)
1424    * fixed bug in handling of function types involving incomplete pointers
1425    * generate ML entry points that take record arguments (i.e., using
1426      named arguments) for C functions that have a prototype with named
1427      arguments
1428      (see changes to CKIT)
1429    
1430    ----------------------------------------------------------------------
1431    Name: Allen Leung
1432    Date: 2001/10/27 20:34:00 EDT
1433    Tag: leunga-20011027-x86-fast-fp-call
1434    Description:
1435    
1436       Fixed the bug described in blume-20010920-slowfp.
1437    
1438       The fix involves
1439          1. generating FCOPYs in FSTP in ia32-svid
1440          2. marking a CALL with the appropriate annotation
1441    
1442    ----------------------------------------------------------------------
1443    Name: Matthias Blume
1444    Date: 2001/10/16 11:32:00 EDT
1445    Tag: blume-20011016-netbsd
1446    Description:
1447    
1448    Underscore patch from Chris Richards (fixing problem with compiling
1449    runtime system under recent NetBSD).
1450    
1451    ----------------------------------------------------------------------
1452    Name: Allen Leung
1453    Date: 2001/10/12 17:18:32 EDT 2001
1454    Tag: leung-20011012-x86-printflowgraph
1455    Description:
1456    
1457    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
1458    
1459    ----------------------------------------------------------------------
1460    Name: Lal George
1461    Date: 2001/10/11 23:51:34 EDT
1462    Tag: george-20011011-too-many-instrs
1463    Description:
1464    
1465    The representation of a program point never expected to see more
1466    than 65536 instructions in a basic block!
1467    
1468    ----------------------------------------------------------------------
1469    Name: Lal George
1470    Date: 2001/10/09 09:41:37 EDT
1471    Tag: george-20011008-mlrisc-labels
1472    Description:
1473    
1474    Changed the machine description files to support printing of
1475    local and global labels in assembly code, based on host assembler
1476    conventions.
1477    
1478    ----------------------------------------------------------------------
1479    Name: Matthias Blume
1480    Date: 2001/09/25 15:25:00 EDT
1481    Tag: blume-20010925-exninfo
1482    Description:
1483    
1484    I provided a non-hook implementation of exnName (at the toplevel) and
1485    made the "dummy" implementation of exnMessage (at the toplevel) more
1486    useful: if nothing gets "hooked in", then at least you are going to
1487    see the exception name and a message indicating why you don't see more.
1488    
1489    [For the time being, programs that need exnMessage and want to use
1490    ml-build should either use General.exnMessage (strongly recommended) or
1491    refer to structure General at some other point so that CM sees a
1492    static dependency.]
1493    
1494    [Similar remarks go for "print" and "use":  If you want to use their
1495    functionality in stand-alone programs generated by ml-build, then use
1496    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
1497    
1498    ----------------------------------------------------------------------
1499    Name: Matthias Blume
1500    Date: 2001/09/20 17:28:00 EDT
1501    Tag: blume-20010920-slowfp
1502    Description:
1503    
1504    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
1505    
1506    ----------------------------------------------------------------------
1507    Name: Matthias Blume
1508    Date: 2001/09/20 17:20:00 EDT
1509    Tag: blume-20010920-canonicalpaths
1510    Description:
1511    
1512    0. Updated the BOOT file (something that I forgot to do earlier).
1513    
1514    1. Small internal change to CM so that it avoids "/../" in filenames
1515       as much as possible (but only where it is safe).
1516    
1517    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
1518       that arguments that contain delimiters are passed through correctly.
1519       This change also means that all "special" arguments of the form
1520       @SMLxxx... must come first.
1521    
1522    3. Changed install script to put relative anchor names for tool commands
1523       into pathconfig.
1524    
1525    ----------------------------------------------------------------------
1526    Name: Matthias Blume
1527    Date: 2001/09/18 15:35:00 EDT
1528    Tag: blume-20010918-readme11036
1529    Description:
1530    
1531    Added README files.
1532    
1533    ----------------------------------------------------------------------
1534    Name: Matthias Blume
1535    Date: 2001/09/18 11:45:00 EDT
1536    Tag: Release_110_36 (retag)
1537    Description:
1538    
1539    Fixed mistake in config/preloads. Retagged as 110.36.
1540    
1541    ----------------------------------------------------------------------
1542    Name: Matthias Blume
1543    Date: 2001/09/18 09:40:00 EDT
1544    Tag: Release_110_36_orig (tag changed)
1545    Description:
1546    
1547    New version (110.36).  New bootfiles.
1548    
1549    ----------------------------------------------------------------------
1550    Name: Matthias Blume
1551    Date: 2001/09/14 16:15:00 EDT
1552    Tag: blume-20010914-x86fastfp
1553    Description:
1554    
1555    John committed some changes that Allen made, in particular a (hopefully)
1556    correctly working version of the x86-fp module.
1557    
1558    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
1559    flag to "true".  Everything seems to compile to a fixpoint ok, and
1560    "mandelbrot" speeds up by about 15%.
1561    
1562    ----------------------------------------------------------------------
1563    Name: Matthias Blume
1564    Date: 2001/09/13 11:20:00 EDT
1565    Tag: blume-20010913-minimal
1566    Description:
1567    
1568    1. Stefan Monnier's patch to fix a miscompilation problem that
1569       was brought to light by John Reppy's work on Moby.
1570    
1571    2. Implemented a minimal "structure Compiler" that contains just
1572       "version" and "architecture".  The minimal version will be
1573       available when the full version is not.  This is for backward-
1574       compatibility with code that wants to test Compiler.version.
1575    
1576    ----------------------------------------------------------------------
1577    Name: Matthias Blume
1578    Date: 2001/08/28 14:03:00 EDT
1579    Tag: blume-20010828-ml-lex
1580    Description:
1581    
1582    Fix for bug 1581, received from Neophytos Michael.
1583    
1584    ----------------------------------------------------------------------
1585    Name: Matthias Blume
1586    Date: 2001/08/27 11:20:00 EDT
1587    Tag: blume-20010827-readme11035
1588    Description:
1589    
1590    Fleshed out the README file for 110.35.
1591    
1592    ----------------------------------------------------------------------
1593    Name: Matthias Blume
1594    Date: 2001/08/24 17:10:00 EDT
1595    Tag: Release_110_35
1596    Description:
1597    
1598    New version number (110.35).  New bootfiles.
1599    
1600    ----------------------------------------------------------------------
1601    Name: Lal George
1602    Date: 2001/08/24 13:47:18 EDT 2001
1603    Tag: george-20010824-MLRISC-graphs
1604    Description:
1605    
1606     removed clusters from MLRISC completely and replaced with graphs.
1607    
1608    ----------------------------------------------------------------------
1609    Name: Matthias Blume
1610    Date: 2001/08/23 17:50:00 EDT
1611    Tag: blume-20010823-toplevel
1612    Description:
1613    
1614    - some reorganization of the code that implements various kinds of
1615      environments in the compiler (static, dynamic, symbolic, combined)
1616    - re-implemented the EnvRef module so that evalStream works properly
1617      (if the stream contains references to "use", "CM.make", etc.)
1618    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
1619    
1620    ----------------------------------------------------------------------
1621    Name: Matthias Blume
1622    Date: 2001/08/20 15:50 EDT
1623    Tag: blume20010820-slipup
1624    Description:
1625    
1626    I forgot to commit a few files.  Here they are...
1627    
1628    ----------------------------------------------------------------------
1629    Name: Matthias Blume
1630    Date: 2001/08/20 15:35:00 EDT
1631    Tag: blume-20010820-debugprof
1632    Description:
1633    
1634    !!!! NEW BOOTFILES !!!!
1635    
1636    This is another round of reorganizing the compiler sources.  This
1637    time the main goal was to factor out all the "instrumentation"
1638    passes (for profiling and backtracing) into their own library.
1639    The difficulty was to do it in such a way that it does not depend
1640    on elaborate.cm but only on elabdata.cm.
1641    
1642    Therefore there have been further changes to both elaborate.cm and
1643    elabdata.cm -- more "generic" things have been moved from the former
1644    to the latter.  As a result, I was forced to split the assignment
1645    of numbers indicating "primtyc"s into two portions: SML-generic and
1646    SML/NJ-specific.  Since it would have been awkward to maintain,
1647    I bit the bullet and actually _changed_ the mapping between these
1648    numbers and primtycs.  The bottom line of this is that you need
1649    a new set of bin- and bootfiles.
1650    
1651    I have built new bootfiles for all architectures, so doing a fresh
1652    checkout and config/install.sh should be all you need.
1653    
1654    The newly created library's name is
1655    
1656        $smlnj/viscomp/debugprof.cm
1657    
1658    and its sources live under
1659    
1660        src/compiler/DebugProf
1661    
1662    ----------------------------------------------------------------------
1663    Name: Matthias Blume
1664    Date: 2001/08/15 17:15:00 EDT
1665    Tag: blume-20010815-compreorg
1666    Description:
1667    
1668    This is a first cut at reorganizing the CM libraries that make up the
1669    core of the compiler.  The idea is to separate out pieces that could
1670    be used independently by tools, e.g., the parser, the typechecker, etc.
1671    
1672    The current status is a step in this direction, but it is not quite
1673    satisfactory yet.  Expect more changes in the future.
1674    
1675    Here is the current (new) organization...
1676    
1677        What used to be $smlnj/viscomp/core.cm is now divided into
1678        six CM libraries:
1679    
1680             $smlnj/viscomp/basics.cm
1681                           /parser.cm
1682                           /elabdata.cm
1683                           /elaborate.cm
1684                           /execute.cm
1685                           /core.cm
1686    
1687        The CM files for these libraries live under src/system/smlnj/viscomp.
1688        All these libraries are proxy libraries that contain precisely
1689        one CM library component.  Here are the locations of the components
1690        (all within the src/compiler tree):
1691    
1692             Basics/basics.cm
1693             Parse/parser.cm
1694             ElabData/elabdata.cm
1695             Elaborator/elaborate.cm
1696             Execution/execute.cm
1697             core.cm
1698    
1699         [This organization is the same that has been used already
1700         for a while for the architecture-specific parts of the visible
1701         compiler and for the old version of core.cm.]
1702    
1703         As you will notice, many source files have been moved from their
1704         respective original locations to a new home in one of the above
1705         subtrees.
1706    
1707         The division of labor between the new libraries is the following:
1708    
1709             basics.cm:
1710                - Simple, basic definitions that pertain to many (or all) of
1711                  the other libraries.
1712             parser.cm:
1713                - The SML parser, producing output of type Ast.dec.
1714                - The type family for Ast is also defined and exported here.
1715             elabdata.cm:
1716                - The datatypes that describe input and output of the elaborator.
1717                  This includes types, absyn, and static environments.
1718             elaborator.cm:
1719                - The SML/NJ type checker and elaborator.
1720                  This maps an Ast.dec (with a given static environment) to
1721                  an Absyn.dec (with a new static environment).
1722                - This libraries implements certain modules that used to be
1723                  structures as functors (to remove dependencies on FLINT).
1724             execute.cm:
1725                - Everything having to do with executing binary code objects.
1726                - Dynamic environments.
1727             core.cm:
1728                - SML/NJ-specific instantiations of the elaborator and MLRISC.
1729                - Top-level modules.
1730                - FLINT (this should eventually become its own library)
1731    
1732    Notes:
1733    
1734    I am not 100% happy with the way I separated the elaborator (and its
1735    data structures) from FLINT.  Two instances of the same problem:
1736    
1737        1. Data structures contain certain fields that carry FLINT-specific
1738           information.  I hacked around this using exn and the property list
1739           module from smlnj-lib.  But the fact that there are middle-end
1740           specific fields around at all is a bit annoying.
1741    
1742        2. The elaborator calculates certain FLINT-related information.  I tried
1743           to make this as abstract as I could using functorization, but, again,
1744           the fact that the elaborator has to perform calculations on behalf
1745           of the middle-end at all is not nice.
1746    
1747        3. Having to used exn and property lists is unfortunate because it
1748           weakens type checking.  The other alternative (parameterizing
1749           nearly *everything*) is not appealing, though.
1750    
1751    I removed the "rebinding =" warning hack because due to the new organization
1752    it was awkward to maintain it.  As a result, the compiler now issues some of
1753    these warnings when compiling init.cmi during bootstrap compilation. On
1754    the plus side, you also get a warning when you do, for example:
1755       val op = = Int32.+
1756    which was not the case up to now.
1757    
1758    I placed "assign" and "deref" into the _Core structure so that the
1759    code that deals with the "lazy" keyword can find them there.  This
1760    removes the need for having access to the primitive environment
1761    during elaboration.
1762    
1763    ----------------------------------------------------------------------
1764    Name: Matthias Blume
1765    Date: 2001/08/13
1766    Tag: blume-20010813-closures
1767    Description:
1768    
1769    This fix was sent to us by Zhong Shao.  It is supposed to improve the
1770    performance of certain loops by avoiding needless closure allocation.
1771    
1772    ----------------------------------------------------------------------
1773    Name: Lal George
1774    Date: 2001/07/31 10:03:23 EDT 2001
1775    Tag: george-20010731-x86-fmalloc
1776    Description: Fixed bug in x86 calls
1777    
1778        There was a bug where call instructions would mysteriously
1779        vanish. The call instruction had to be one that returned
1780        a floating point value.
1781    
1782    ----------------------------------------------------------------------
1783    Name: Lal George
1784    Date: 2001/07/19 16:36:29 EDT 2001
1785    Tag: george-20010719-simple-cells
1786    Description:
1787    
1788    I have dramatically simplified the interface for CELLS in MLRISC.
1789    
1790    In summary, the cells interface is broken up into three parts:
1791    
1792      1. CellsBasis : CELLS_BASIS
1793    
1794            CellsBasis is a top level structure and common for all
1795            architectures.  it contains the definitions of basic datatypes
1796            and utility  functions over these types.
1797    
1798      2. functor Cells() : CELLS
1799    
1800            Cells generates an interface for CELLS that incorporates the
1801            specific resources on the target architecture, such as the
1802            presence of special register classes, their number and size,
1803            and various useful substructures.
1804    
1805      3. <ARCH>CELLS
1806    
1807            e.g. SparcCells: SPARCCELLS
1808    
1809            <ARCH>CELLS usually contains additional bindings for special
1810            registers  on the architecture, such as:
1811    
1812                    val r0 : cell           (* register zero *)
1813                    val y : cell            (* Y register *)
1814                    val psr : cell          (* processor status register *)
1815                    ...
1816    
1817            The structure returned by applying the Cells functor is opened
1818            in this interface.
1819    
1820    The main implication of all this is that the datatypes for cells is
1821    split between CellsBasis and CELLS -- a fairly simple change for user
1822    code.
1823    
1824    In the old scheme the CELLS interface had a definitional binding of
1825    the form:
1826    
1827            signature CELLS = sig
1828    
1829               structure CellsBasis = CellsBasis
1830    
1831               ...
1832    
1833            end
1834    
1835    With all the sharing constraints that goes on in MLRISC, this old
1836    design  quickly leads to errors such as:
1837    
1838            "structure definition spec inside of sharing ... "
1839    
1840    
1841    and appears to require an unacceptable amount of sharing and where
1842    constraint hackery.
1843    
1844    I think this error message (the interaction of definitional specs and
1845    sharing) requires more explanation on our web page.
1846    
1847    ----------------------------------------------------------------------
1848    Name: Matthias Blume
1849    Date: 2001/07/19 15:00:00 EDT
1850    Tag: blume-20010719-libreorg
1851    Description:
1852    
1853    This update puts together a fairly extensive but straightforward change
1854    to the way the libraries that implement the interactive system are
1855    organized:
1856    
1857       The biggest change is the elimination of structure Compiler.  As a
1858       replacement for this structure, there is now a CM library
1859       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
1860       that exports all the substructures of the original structure Compiler
1861       directly.  So instead of saying Compiler.Foo.bar one now simply
1862       says Foo.bar.  (The CM libraries actually export a collection of
1863       structures that is richer than the collection of substructures of
1864       structure Compiler.)
1865    
1866       To make the transition smooth, there is a separate library called
1867       $smlnj/compiler/compiler.cm which puts together and exports the
1868       original structure Compiler (or at least something very close to it).
1869    
1870       There are five members of the original structure Compiler
1871       that are not exported directly but which instead became members
1872       of a new structure Backend (described by signature BACKEND).  These are:
1873       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
1874       Interact (: INTERACT), structure Machine (: MACHINE), and val
1875       architecture (: string).
1876    
1877       Structure Compiler.Version has become structure CompilerVersion.
1878    
1879       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
1880       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
1881       or x86, respectively.
1882       Each of these exports the same frontend structures that
1883       $smlnj/compiler.cm exports.  But they do not have a structure Backend
1884       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
1885       Hppa, PPC, Sparc, or X86, respectively.
1886    
1887       Library $smlnj/compiler/all.cm exports the union of the exports of
1888       $smlnj/compiler/<arch>.cm
1889    
1890       There are no structures <Arch>Compiler anymore, use
1891       $smlnj/compiler/<arch>.cm instead.
1892    
1893       Library host-compiler-0.cm is gone.  Instead, the internal library
1894       that instantiates CM is now called cm0.cm.  Selection of the host
1895       compiler (backend) is no longer done here but. (Responsibility for it
1896       now lies with $smlnj/compiler/current.cm.  This seems to be more
1897       logical.)
1898    
1899       Many individual files have been moved or renamed.  Some files have
1900       been split into multiple files, and some "dead" files have been deleted.
1901    
1902    Aside from these changes to library organization, there are also changes
1903    to the way the code itself is organized:
1904    
1905       Structure Binfile has been re-implemented in such a way that it no
1906       longer needs any knowledge of the compiler.  It exclusively deals
1907       with the details of binfile layout.  It no longer invokes the
1908       compiler (for the purpose of creating new prospective binfile
1909       content), and it no longer has any knowledge of how to interpret
1910       pickles.
1911    
1912       Structure Compile (: COMPILE) has been stripped down to the bare
1913       essentials of compilation.  It no longer deals with linking/execution.
1914       The interface has been cleaned up considerably.
1915    
1916       Utility routines for dealing with linking and execution have been
1917       moved into their own substructures.
1918    
1919       (The ultimate goal of these changes is to provide a light-weight
1920       binfile loader/linker (at least for, e.g., stable libraries) that
1921       does not require CM or the compiler to be present.)
1922    
1923    CM documentation has been updated to reflect the changes to library
1924    organization.
1925    
1926    ----------------------------------------------------------------------
1927    Name: Matthias Blume
1928    Date: 2001/07/10 17:30:00 EDT
1929    Tag: Release_110_34
1930    Description:
1931    
1932    Minor tweak to 110.34 (re-tagged):
1933    
1934      - README.html file added to CVS repository
1935      - runtime compiles properly under FreeBSD 3.X and 4.X
1936    
1937    ----------------------------------------------------------------------
1938    Name: Matthias Blume
1939    Date: 2001/07/10 17:30:00 EDT
1940    Tag: Release_110_34
1941    Description:
1942    
1943    New version number (110.34). New bootfiles.
1944    
1945    ----------------------------------------------------------------------
1946    Name: Matthias Blume
1947    Date: 2001/07/09 16:00:00 EDT
1948    Tag: blume-20010709-more-varargs
1949    Description:
1950    
1951    I changed the handling of varargs in ml-nlffigen again:
1952    The ellipsis ... will now simply be ignored (with an accompanying warning).
1953    
1954    The immediate effect is that you can actually call a varargs function
1955    from ML -- but you can't actually supply any arguments beyond the ones
1956    specified explicitly.  (For example, you can call printf with its format
1957    string, but you cannot pass additional arguments.)
1958    
1959    This behavior is only marginally more useful than the one before, but
1960    it has the advantage that a function or, more importantly, a function
1961    type never gets dropped on the floor, thus avoiding follow-up problems with
1962    other types that refer to the offending one.
1963    
1964    ----------------------------------------------------------------------
1965    Name: Matthias Blume
1966    Date: 2001/07/09 11:25:00 EDT
1967    Tag: blume-20010709-varargs
1968    Description:
1969    
1970    1. ckit-lib.cm now exports structure Error
1971    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1972       with a warning accompanied by a source location.  Moreover, it
1973       merely skips the offending function or type and proceeds with the
1974       rest of its work.u  As a result, one can safely feed C code containing
1975       "..." to ml-nlffigen.
1976    3. There are some internal improvements to CM, providing slightly
1977       more general string substitutions in the tools subsystem.
1978    
1979    ----------------------------------------------------------------------
1980    Name: Matthias Blume
1981    Date: 2001/06/27 15:10:00 EDT
1982    Tag: blume-20010627-concur
1983    Description:
1984    
1985    Fixed a small bug in CM's handling of parallel compilation.
1986    (You could observe the bug by Control-C-interrupting an ordinary
1987    CMB.make or CM.stabilize and then attaching some compile servers.
1988    The result was that all of a sudden the previously interrupted
1989    compilation would continue on its own.  This was because of
1990    an over-optimization: CM did not bother to clean out certain queues
1991    when no servers were attached "anyway", resulting in the contents
1992    of these queues to grab control when new servers did get attached.)
1993    
1994    There is also another minor update to the CM manual.
1995    
1996    ----------------------------------------------------------------------
1997    Name: Matthias Blume
1998    Date: 2001/06/26 16:15:00 EDT
1999    Tag: blume-20010626-cmdoc
2000    Description:
2001    
2002    Minor typo fixed in CM manual (syntax diagram for libraries).
2003    
2004    ----------------------------------------------------------------------
2005    Name: Matthias Blume
2006    Date: 2001/06/25 22:55:00 EDT
2007    Tag: blume-20010625-x86pc
2008    Description:
2009    
2010    Fixed a nasty bug in the X86 assembly code that caused signal
2011    handlers to fail (crash) randomly.
2012    
2013    ----------------------------------------------------------------------
2014    Name: Matthias Blume
2015    Date: 2001/06/25 12:05:00 EDT
2016    Tag: blume-20010625-nlffigen
2017    Description:
2018    
2019    This update fixes a number of minor bugs in ml-nlffigen as reported by
2020    Nick Carter <nbc@andrew.cmu.edu>.
2021    
2022      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
2023      2. Default names for generated files are now derived from the name of
2024         the C file *without its directory*.  In particular, this causes generated
2025         files to be placed locally even if the C file is in some system directory.
2026      3. Default names for generated signatures and structures are also derived
2027         from the C file name without its directory.  This avoids silly things
2028         like "structure GL/GL".
2029         (Other silly names are still possible because ml-nlffigen does not do
2030          a thorough check of whether generated names are legal ML identifiers.
2031          When in doubt, use command line arguments to force particular names.)
2032    
2033    ----------------------------------------------------------------------
2034    Name: Matthias Blume
2035    Date: 2001/06/21 12:25:00 EDT
2036    Tag: blume-20010621-eXene
2037    Description:
2038    
2039    eXene now compiles and (sort of) works again.
2040    
2041    The library name (for version > 110.33) is $/eXene.cm.
2042    
2043    I also added an new example in src/eXene/examples/nbody.  See the
2044    README file there for details.
2045    
2046    ----------------------------------------------------------------------
2047    Name: Matthias Blume
2048    Date: 2001/06/20 16:40:00 EDT
2049    Tag: blume-20010620-cml
2050    Description:
2051    
2052    CML now compiles and works again.
2053    
2054    Libraries (for version > 110.33):
2055    
2056      $cml/cml.cm            Main CML library.
2057      $cml/basis.cm          CML's version of $/basis.cm.
2058      $cml/cml-internal.cm   Internal helper library.
2059      $cml/core-cml.cm       Internal helper library.
2060      $cml-lib/trace-cml.cm  Tracing facility.
2061      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
2062    
2063    The installer (config/install.sh) has been taught how to properly
2064    install this stuff.
2065    
2066    ----------------------------------------------------------------------
2067    Name: Matthias Blume
2068    Date: 2001/06/19 17:55:00 EDT
2069    Tag: blume-20010619-instantiate
2070    Description:
2071    
2072    This un-breaks the fix for bug 1432.
2073    (The bug was originally fixed in 110.9 but I broke it again some
2074    time after that.)
2075    
2076    ----------------------------------------------------------------------
2077    Name: Matthias Blume
2078    Date: 2001/06/19 17:25:00 EDT
2079    Tag: blume-20010619-signals
2080    Description:
2081    
2082    This should (hopefully) fix the long-standing signal handling bug.
2083    (The runtime system was constructing a continuation record with an
2084    incorrect descriptor which would cause the GC to drop data on the floor...)
2085    
2086    ----------------------------------------------------------------------
2087    Name: Matthias Blume
2088    Date: 2001/06/15 15:05:00 EDT
2089    Tag: blume-20010615-moresparc
2090    Description:
2091    
2092    Here is a short late-hour update related to Sparc c-calls:
2093    
2094     -- made handling of double-word arguments a bit smarter
2095    
2096     -- instruction selection phase tries to collapse certain clumsily
2097        constructed ML-Trees; typical example:
2098    
2099            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
2100    
2101        This currently has no further impact on SML/NJ since mlriscGen does
2102        not seem to generate such patterns in the first place, and c-calls
2103        (which did generate them in the beginning) has meanwhile been fixed
2104        so as to avoid them as well.
2105    
2106    ----------------------------------------------------------------------
2107    Name: Matthias Blume
2108    Date: 2001/06/15 15:05:00 EDT
2109    Tag: blume-20010615-sparc
2110    Description:
2111    
2112    The purpose of this update is to provide an implementation of NLFFI
2113    on Sparc machines.
2114    
2115    Here are the changes in detail:
2116    
2117       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
2118       the Sparc implementation of the c-calls API.
2119       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
2120       for accessing the ML frame.  Thus, we have a real frame pointer and
2121       can freely modify %sp without need for an omit-frame-ptr phase.
2122       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
2123       this case.
2124       * ml-nlffigen has been taught to produce code for different architectures
2125       and calling conventions.
2126       * In a way similar to what was done in the x86 case, the Sparc
2127       backend uses its own specific extension to mltree.  (For example,
2128       it needs to be able to generate UNIMP instructions which are part
2129       of the calling convention.)
2130       * ml-nlffi-lib was reorganized to make it more modular (in particular,
2131       to make it easier to plug in new machine- and os-dependent parts).
2132    
2133    There are some other fairly unrelated bug fixes and cleanups as well:
2134    
2135       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
2136       that they properly share their libraries with existing SML/NJ libraries.
2137       * I fixed a minor cosmetic bug in CM, supressing certain spurious
2138       follow-up error messages.
2139       * Updates to CM/CMB documentation.
2140    
2141    TODO items:
2142    
2143       * MLRISC should use a different register as its asmTemp on the Sparc.
2144         (The current %o2 is a really bad choice because it is part of the
2145         calling conventions, so things might interfere in unexpected ways.)
2146    
2147    ----------------------------------------------------------------------
2148    Name: Matthias Blume
2149    Date: 2001/06/07
2150    Tag: blume-20010607-calls
2151    Description:
2152    
2153    A number of internal changes related to C calls and calling conventions:
2154    
2155    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
2156       number of bytes popped implicitly (by the callee).  In most cases
2157       this field is 0 but on x86/win32 it is some non-zero value.  This
2158       is information provided for the benefit of the "omit-frameptr" pass.
2159    2. The CALL instruction on the x86 carries a similar "pops" field.
2160       The instruction selection phase copies its value from the ML-Tree
2161       CALL statement.
2162    3. On all other architectures, the instruction selection phase checks
2163       whether "pops=0" and complains if not.
2164    4. The c-calls implementation for x86 now accepts two calling conventions:
2165       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
2166       up after the call and pops is set to 0.  For "stdcall", the caller
2167       does nothing, leaving the cleanup to the callee; pops is set to
2168       the number of bytes that were pushed onto the stack.
2169    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
2170       distinguish between "ccall" and "stdcall".
2171    6. The UNIMP instruction has been added to the supported Sparc instruction
2172       set. (This is needed for implementing the official C calling convention
2173       on this architecture.)
2174    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
2175       work with the latest CM.
2176    
2177    ----------------------------------------------------------------------
2178    Name: Matthias Blume
2179    Date: 2001/06/05 15:10:00 EDT
2180    Tag: blume-20010605-cm-index
2181    Description:
2182    
2183    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
2184    
2185    1. CM can now generate "index files".  These are human-readable files
2186       that list on a per-.cm-file basis each toplevel symbol defined or
2187       imported.  The location of the index file for
2188       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
2189       To enable index-file generation, set CM.Control.generate_index to true
2190       or export an environment-symbol: export CM_GENERATE_INDEX=true.
2191    
2192       The CM manual has been updated accordingly.
2193    
2194    2. I made some slight modifications to the c-calls API in MLRISC.
2195    
2196         a) There is now a callback to support saving/restoring of
2197            dedicated but caller-save registers around the actual call
2198            instruction.
2199         b) One can optionally specify a comment-annotation for the
2200            call instruction.
2201    
2202    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
2203       (For example, the comment annotation shows the C prototype of
2204        the function being called.)
2205    
2206    ----------------------------------------------------------------------
2207    Name: Matthias Blume
2208    Date: 2001/06/01 13:30:00 EDT
2209    Tag: blume-20010601-nlffi-cleanup
2210    Description:
2211    
2212    This is mostly a cleanup of MLFFI stuff:
2213    
2214       - some signature files have been put into a more exposed place
2215       - the ugly 'f type parameter is gone (simplifies types tremendously!)
2216       - ml-nlffigen changed accordingly
2217       - tutorial updated
2218    
2219    Other changes:
2220    
2221       - author's affiliation in CM manual(s) updated
2222       - some more recognized keywords added to Allen's sml.sty
2223    
2224    ----------------------------------------------------------------------
2225    Name: Matthias Blume
2226    Date: 2001/05/25 15:30:00 EDT
2227    Tag: blume-20010525-iptr
2228    Description:
2229    
2230      - put the official 110.33-README (as it appears on the ftp server) under
2231        CVS
2232      - fixed a small bug related to incomplete pointer types in
2233        ml-nlffigen
2234      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
2235        (it does not need the 'f type parameter)
2236    
2237    ----------------------------------------------------------------------
2238    Name: Matthias Blume
2239    Date: 2001/05/23 14:30:00 EDT
2240    Tag: Release_110_33
2241    Description:
2242    
2243    New version number (110.33).  New bootfiles.
2244    
2245    ----------------------------------------------------------------------
2246    Name: Matthias Blume
2247    Date: 2001/05/22 18:06:00 EDT
2248    Tag: blume-20010522-targets
2249    Description:
2250    
2251    Made install.sh use file config/targets.customized if it exists, falling
2252    back to config/targets if it doesn't.  This way one can have a customized
2253    version of the targets file without touching the "real thing", thus
2254    eliminating the constant fear of accidentally checking something bogus
2255    back into the CVS repository...  (File config/targets.customized must
2256    not be added to the repository!)
2257    
2258    ----------------------------------------------------------------------
2259    Name: Matthias Blume
2260    Date: 2001/05/22 16:30:00 EDT
2261    Tag: blume-20010522-minitut
2262    Description:
2263    
2264    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
2265       struct returns.
2266    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
2267       incomplete, preliminary documentation for NLFFI.
2268    
2269    ----------------------------------------------------------------------
2270    Name: Matthias Blume
2271    Date: 2001/05/14 11:30:00 EDT
2272    Tag: blume-20010514-script
2273    Description:
2274    
2275    Some bugs in install script fixed.
2276    
2277    In addition to that I also made a slight change to the NLFFI API:
2278    Functors generated by ml-nlffigen now take the dynamic library as a
2279    straight functor argument, not as a suspended one.  (The original
2280    functor code used to force the suspension right away anyway, so there
2281    was nothing gained by this complication of the interface.)
2282    
2283    ----------------------------------------------------------------------
2284    Name: Matthias Blume
2285    Date: 2001/05/11 14:35:00 EDT
2286    Tag: blume-20010511-ml-nlffi
2287    Description:
2288    
2289    I finally took the plunge and added my new FFI code to the main
2290    repository.  For x86-linux it is now ready for prime-time.
2291    
2292    There are two new subdirectories of "src":
2293    
2294      - ml-nlffi-lib:
2295           The utility library for programs using the FFI interface.
2296           Here is the implementation of $/c.cm and its associated low-level
2297           partners $/c-int.cm and $/memory.cm.
2298      - ml-nlffigen:
2299           A stand-alone program for generating ML glue code from C source
2300           code.
2301    
2302    Building ml-nlffigen requires $/ckit-lib.cm.
2303    
2304    The config/install.sh script has been updates to do the Right Thing
2305    (hopefully).
2306    
2307    Notice that the source tree for the C-Kit will not be put under "src"
2308    but directly under the installation root directory.  (This is the
2309    structure that currently exists on the CVS server when you check out
2310    module "sml".)  Fortunately, config/install.sh knows about this oddity.
2311    
2312    Bugs: No documentation yet.
2313    
2314    ----------------------------------------------------------------------
2315    Name: Matthias Blume
2316    Date: 2001/05/09 16:35:00 EDT
2317    Tag: blume-20010509-cpscontract
2318    Description:
2319    
2320    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
2321    wrapper/unwrapper elimination did not decrement usage counts and some
2322    dead variables got overlooked by the dead-up logic.)
2323    
2324    ----------------------------------------------------------------------
2325    Name: Lal George
2326    Date: 2001/05/08  17:26:09 EDT
2327    Tag: george-20010508-omit-frameptr
2328    Description:
2329    
2330    Changes to implement the omit-frame-pointer optimization to support
2331    raw C calls. For now, there is only support on the Intel x86, but
2332    other architectures will follow as more experience is gained with this.
2333    
2334    
2335    ----------------------------------------------------------------------
2336    Name: Matthias Blume
2337    Date: 2001/05/07 14:40:00 EDT
2338    Tag: blume-20010507-proxies
2339    Description:
2340    
2341    I made into "proxy libraries" all libraries that qualify for such a
2342    change.  (A qualifying library is a library that has another library or
2343    groups as its sole member and repeats that member's export list
2344    verbatim.  A proxy library avoids this repetition by omitting its export
2345    list, effectively inheriting the list that its (only) member exports.
2346    See the CM manual for more explanation.)
2347    The main effect is that explicit export lists for these libraries
2348    do not have to be kepts in sync, making maintenance a bit easier.
2349    
2350    I also added copyright notices to many .cm-files.
2351    
2352    Last but not least, I made a new set of bootfiles.
2353    
2354    ----------------------------------------------------------------------
2355    Name: Matthias Blume
2356    Date: 2001/05/04 17:00:00 EDT
2357    Tag: blume-20010504-cm-lsplit
2358    Description:
2359    
2360    0. John merged pending changes to $/smlnj-lib.cm
2361    
2362    1. Allen's previous change accidentally backed out of one of Lal's
2363       earlier changes.  I undid this mistake (re-introducing Lal's change).
2364    
2365    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
2366       within the compiler where applicable.  There is some code simplification
2367       because of that.
2368    
2369    3. The "split" phase (in FLINT) is now part of the default list of phases.
2370       Compiler.Control.LambdaSplitting.* can be used to globally control the
2371       lambda-splitting (cross-module-inlining) engine.  In addition to that,
2372       it can now also be controlled on a per-source basis: CM has been taught
2373       a new tool parameter applicable to ML source files.
2374    
2375       - To turn lambda-splitting off completely:
2376            local open Compiler.Control.LambdaSplitting in
2377                val _ = set Off
2378            end
2379       - To make "no lambda-splitting" the global default (but allow per-source
2380         overriding); this is the initial setting:
2381            local open Compiler.Control.LambdaSplitting in
2382                val _ = set (Default NONE)
2383            end
2384       - To make "lambda-splitting with aggressiveness a" the global default
2385         (and allow per-source overriding):
2386            local open Compiler.Control.LambdaSplitting in
2387                val _ = set (Default (SOME a))
2388            end
2389    
2390       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
2391         write (in the respective .cm-file):
2392            a.sml (lambdasplitting:off)
2393       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
2394            a.sml (lambdasplitting:on)
2395       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
2396         <a> is a decimal non-negative integer):
2397            a.sml (lambdasplitting:<a>)
2398       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
2399            a.sml (lambdasplitting:infinity)
2400       - To use the global default for a.sml:
2401            a.sml (lambdasplitting:default)
2402         or simply
2403            a.sml
2404    
2405    ----------------------------------------------------------------------
2406    Name: Allen Leung
2407    Date: 2001/05/04 01:57:00 EDT
2408    Tag: leunga-20010504-sync
2409    Description:
2410    
2411      MLRISC features.
2412    
2413      1. Fix to CMPXCHG instructions.
2414      2. Changed RA interface to allow annotations in callbacks.
2415      3. Added a new method to the stream interface to allow annotations updates.
2416    
2417    ----------------------------------------------------------------------
2418    Name: Matthias Blume
2419    Date: 2001/05/01 11:45:00 EDT
2420    Tag: blume-20010501-pcedittmp
2421    Description:
2422    
2423    Changed install.sh to use the current working directory instead of
2424    /usr/tmp for a temporary file (pcedittmp).  The previous choice
2425    of /usr/tmp caused trouble with MacOS X because of file premission
2426    problems.
2427    
2428    ----------------------------------------------------------------------
2429    Name: Matthias Blume
2430    Date: 2001/04/20 11:10:00 EDT
2431    Tag: blume-20010420-inMLflag
2432    Description:
2433    
2434     - added vp_limitPtrMask to vproc-state.h
2435       (for use by the raw-C-calls mechanism to implement proper interrupt
2436        handling)
2437     - made the ML compiler aware of various data-structure offsets so it
2438       can generate code for accessing the vp_inML flag and vp_limitPtrMask
2439     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
2440       raw C-calls
2441    
2442    ----------------------------------------------------------------------
2443    Name: Lal George
2444    Date: 2001/04/20 09:15:28 EDT
2445    Tag: george-20010420-macosX
2446    Description:
2447    
2448     - Changes to port to Mac OS X; Darwin.
2449    
2450     - In the process I found that sqrt was broken on the PPC, because the
2451       fsqrt instruction is not implemented.
2452    
2453    ----------------------------------------------------------------------
2454    Name: Matthias Blume
2455    Date: 2001/04/18 12:45:00 EDT
2456    Tag: blume-20010418-ccalls
2457    Description:
2458    
2459     - fixed two off-by-4 errors in the x86-specific c-calls implementation
2460       (this bug prevented structure arguments containing pointers from being
2461        passed correctly)
2462     - changed the raw-C-call code in mlriscGen.sml in such a way that
2463       structure arguments are represented as a pointer to the beginning
2464       of the structure (instead of having a series of synthesized arguments,
2465       one for each structure member)
2466    
2467     - made makeml script's verbosity level configurable via environment
2468       variable (MAKEML_VERBOSITY)
2469    
2470     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
2471       in rawmem-x86.sml; we are now using the real thing
2472    
2473    ----------------------------------------------------------------------
2474    Name: Matthias Blume
2475    Date: 2001/03/22 16:25:00 EST
2476    Tag: blume-20010322-bootfiles
2477    Description:
2478    
2479    Created a new set of bootfiles (for your automatic installation convenience).
2480    
2481    ----------------------------------------------------------------------
2482    Name: Matthias Blume
2483    Date: 2001/03/22 15:10:00 EST
2484    Tag: blume-20010322-rawmem-parcm
2485    Description:
2486    
2487    1. All "raw memory access" primitives for the new FFI are implemented now
2488       (at least on the x86).
2489    2. Some further cleanup of CM's parallel make mechanism.
2490    
2491    ----------------------------------------------------------------------
2492    Name: Matthias Blume
2493    Date: 2001/03/19 17:53:00 EST
2494    Tag: blume-20010319-parallel
2495    Description:
2496    
2497    Parallel make (using compile servers) now works again.
2498    
2499    To this end, CM.stabilize and CMB.make have been modified to work in
2500    two passes when compile servers are attached:
2501       1. Compile everything, do not perform stabilization; this pass
2502          uses compile servers
2503       2. Stabilize everything; this pass does not use compile servers
2504    If there are no compile servers, the two passes are combined into one
2505    (as before).  Splitting the passes increases the inherent parallelism
2506    in the dependency graph because the entire graph including all
2507    libraries is available at the same time.  This, in turn, improves
2508    server utilization.  The downside is that the master process will
2509    have to do some extra work after compilation is done (because for
2510    technical reasons it must re-read all the binfiles during stabilization).
2511    
2512    ----------------------------------------------------------------------
2513    Name: Matthias Blume
2514    Date: 2001/03/16 12:22:00 EST
2515    Tag: blume-20010316-bootfiles
2516    Description:
2517    
2518    Created a new set of bootfiles (for your automatic installation convenience).
2519    
2520    ----------------------------------------------------------------------
2521    Name: Matthias Blume
2522    Date: 2001/03/16 11:00:00 EST
2523    Tag: blume-20010316-MLTREE-fixup
2524    Description:
2525    
2526    This is a minor fixup for an (untagged) earlier commit by Allen.
2527    (A file was missing).
2528    
2529    ----------------------------------------------------------------------
2530    Name: Allen Leung
2531    Date: Mon Mar  5 18:54:57 EST 2001
2532    Tag: leunga-20010305-cut-support
2533    
2534    1. New support for alternative control-flow in MLTREE.
2535       Currently we support
2536    
2537          FLOW_TO(CALL ...., [k1,...,kn])
2538    
2539       This is needed for 'cuts to' in C-- and try/handle-like constructs
2540       in Moby
2541    
2542       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
2543    
2544    2. Register Allocator
2545    
2546       Changes in interface [from Fermin, John]
2547    
2548    3. Alpha 8-bit SLL support [Fermin]
2549    
2550    4. All architectures
2551    
2552       A new module (ClusterExpandCopies) for expanding parallel copies.
2553    
2554    ----------------------------------------------------------------------
2555    Name: Allen Leung
2556    Date: 2001/02/27 23:07:00 EST
2557    Tag: leunga-20010227-minor-stuff
2558    
2559    1. Alpha bug fix for CMOVNE
2560    2. Handle mltree COND(..,FCMP ...,...)
2561    3. Bug fix in simplifier
2562    
2563    ----------------------------------------------------------------------
2564    Name: Matthias Blume
2565    Date: 2001/01/30 17:50:00 EST
2566    Tag: blume-20010130-sync
2567    Description:
2568    
2569    This is just a minor update to sync my devel branch with the main brach.
2570    The only visible change is the addition of some README files.
2571    
2572    ----------------------------------------------------------------------
2573    Name: Matthias Blume
2574    Date: 2001/01/12 23:30:00 JST
2575    Tag: blume-20010112-bootfiles
2576    Description:
2577    
2578    Made a new set of bootfiles that goes with the current state of the
2579    repository.
2580    
2581    ----------------------------------------------------------------------
2582    Name: Matthias Blume
2583    Date: 2001/01/12 21:20:00 JST
2584    Tag: blume-20010112-sync
2585    Description:
2586    
2587    I am just flushing out some minor changes that had accumulated in
2588    my private branch in order to sync with the main tree.  (This is
2589    mainly because I had CVS trouble when trying to merge _into_ my
2590    private branch.)
2591    
2592    Most people should be completely unaffected by this.
2593    
2594    ----------------------------------------------------------------------
2595    Name: Allen Leung
2596    Date: Thu Jan 11 21:03:00 EST 2001
2597    Tag: leunga-20010111-labexp=mltree
2598    Description:
2599    
2600    1.  Removed the type LabelExp and replace it by MLTree.
2601    2.  Rewritten mltree-simplify with the pattern matcher tool.
2602    3.  There were some bugs in alpha code generator which would break
2603        64-bit code generation.
2604    4.  Redo the tools to generate code with the
2605    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
2606        are now generated by perl scripts.
2607    
2608    ----------------------------------------------------------------------
2609    Name: Matthias Blume
2610    Date: 2001/01/10 21:55:00 JST
2611    Tag: blume-20010110-rcc
2612    Description:
2613    
2614    The RCC stuff now seems to work (but only on the x86).
2615    This required hacking of the c-calls interface (and -implementation) in
2616    MLRISC.
2617    
2618    Normal compiler users should be unaffected.
2619    
2620    ----------------------------------------------------------------------
2621    Name: Matthias Blume
2622    Date: 2001/01/09 01:20:00 JST
2623    Tag: blume-20010109-rcc
2624    Description:
2625    
2626    This is a fairly big patch, flushing out a large number of pending
2627    changes that I made to my development copy over the last couple of days.
2628    
2629    Of practical relevance at this moment is a workaround for a pickling
2630    bug that Allen ran into the other day.  The cause of the bug itself is
2631    still unknown and it might be hard to fix it properly, but the
2632    workaround has some merits of its own (namely somewhat reducing pickling
2633    overhead for certain libraries).  Therefore, I think this solution should
2634    be satisfactory at this time.
2635    
2636    The rest of the changes (i.e., the vast majority) has to do with my
2637    ongoing efforts of providing direct support for C function calls from
2638    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
2639    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
2640    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
2641    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
2642    handle RCC.
2643    
2644    The last part (the changes to mlriscGen.sml) are still known to be
2645    wrong on the x86 and not implemented on all other architectures.  But
2646    the infrastructure is in place. I had to change a few functor
2647    signatures in the backend to be able to route the CCalls interface
2648    from MLRISC there, and I had to specialize the mltree type (on the
2649    x86) to include the necessary extensions. (The extensions themselves
2650    were already there and redy to go in MLRISC/x86).
2651    
2652    Everything should be very happy as soon as someone helps me with
2653    mlriscGen.sml...
2654    
2655    In any case, nothing of this should matter to anyone as long as the
2656    new primop is not being used (which is going to be the case unless you
2657    find it where I hid it :). The rest of the compiler is completely
2658    unaffected.
2659    
2660    ----------------------------------------------------------------------
2661    Name: Matthias Blume
2662    Date: 2001/01/05 00:30:00 JST
2663    Tag: blume-20010105-primops
2664    Description:
2665    
2666    Added some experimental support for work that I am doing right now.
2667    These changes mostly concern added primops, but there is also a new
2668    experimental C library in the runtime system (but currently not enabled
2669    anywhere except on Linux/X86).
2670    
2671    In the course of adding primops (and playing with them), I discovered that
2672    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
2673    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
2674    To recover, I made type information in INL_PRIM mandatory and changed
2675    prim.sml as well as built-in.sml accordingly.  The InLine structure now
2676    has complete, correct type information (i.e., no bottom types).
2677    
2678    Since all these changes mean that we need new binfiles, I also bumped the
2679    version number to 110.32.1.
2680    
2681    ----------------------------------------------------------------------
2682    Name: Matthias Blume
2683    Date: 2000/12/30 22:10:00 JST
2684    Tag: blume-20001230-various
2685    Description:
2686    
2687    Added proxy libraries for MLRISC and let MLRISC libraries refer
2688    to each other using path anchors.  (See CM manual for explanation.)
2689    
2690    Updated CM documentation.
2691    
2692    Fixed some bugs in CM.
2693    
2694    Implemented "proxy" libraries (= syntactic sugar for CM).
2695    
2696    Added "-quiet" option to makeml and changed runtime system accordingly.
2697    
2698    Added cleanup handler for exportML to reset timers and compiler stats.
2699    
2700    ----------------------------------------------------------------------
2701    Name: Lal George
2702    Date: 2000/12/22 22:22:58 EST 2000
2703    Tag: Release_110_32
2704    Description:
2705    
2706            Infinite precision used throughout MLRISC.
2707            see MLRISC/mltree/machine-int.sig
2708    
2709    ----------------------------------------------------------------------
2710    Name: Matthias Blume
2711    Date: 2000/12/22 23:16:00 JST
2712    Tag: blume-20001222-warn
2713    Description:
2714    
2715    Corrected wording and formatting of some CM warning message which I
2716    broke in my previous patch.
2717    
2718    ----------------------------------------------------------------------
2719    Name: Matthias Blume
2720    Date: 2000/12/22 21:20:00 JST
2721    Tag: blume-20001222-anchorenv
2722    Description:
2723    
2724    Fixed CM's handling of anchor environments in connection with CMB.make.
2725    
2726    ----------------------------------------------------------------------
2727    Name: Matthias Blume
2728    Date: 2000/12/22 13:15:00 JST
2729    Tag: blume-20001222-cleanup
2730    Description:
2731    
2732    Removed src/cm/ffi which does not (and did not) belong here.
2733    
2734    ----------------------------------------------------------------------
2735    Name: Matthias Blume
2736    Date: 2000/12/21 23:55:00 JST
2737    Tag: blume-20001221-exn
2738    Description:
2739    
2740    Probably most important: CM no longer silently swallows all exceptions
2741    in the compiler.
2742    Plus: some other minor CM changes.  For example, CM now reports some
2743    sizes for generated binfiles (code, data, envpickle, lambdapickle).
2744    
2745    ----------------------------------------------------------------------
2746    Name: Matthias Blume
2747    Date: 2000/12/15 00:01:05 JST
2748    Tag: blume-20001215-dirtool
2749    Description:
2750    
2751    - "dir" tool added.
2752    - improvements and cleanup to Tools structure
2753    - documentation updates
2754    
2755    ----------------------------------------------------------------------
2756    Name: Allen Leung
2757    Date: Thu Dec 14 03:45:24 EST 2000
2758    Description:
2759    Tag:  leunga-20001214-int-inf
2760    Description:
2761    
2762       In IntInf, added these standard functions, which are missing from our
2763    implementation:
2764    
2765        andb : int * int -> int
2766        xorb : int * int -> int
2767        orb  : int * int -> int
2768        notb : int -> int
2769         <<   : int * word -> int
2770        ~>>  : int * word -> int
2771    
2772       Not tested, I hope they are correct.
2773    
2774    ----------------------------------------------------------------------
2775    Name: Allen Leung
2776    Date: Fri Dec  8 19:23:26 EST 2000
2777    Description:
2778    Tag:  leunga-20001208-nowhere
2779    Description:
2780    
2781      Slight improvements to the 'nowhere' tool to handle OR-patterns,
2782    to generate better error messages etc.  Plus a brief manual.
2783    
2784    ----------------------------------------------------------------------
2785    Name: Lal George
2786    Date: 2000/12/08 09:54:02 EST 2000
2787    Tag: Release_110_31
2788    Description:
2789    
2790    - Version 110.31
2791    ----------------------------------------------------------------------
2792    Name: Allen Leung
2793    Date: Thu Dec  7 22:01:04 EST 2000
2794    Tag:  leunga-20001207-cell-monster-hack
2795    Description:
2796    
2797    Major MLRISC internal changes.  Affect all clients.
2798    Summary:
2799    
2800    1.  Type CELLS.cell = int is now replaced by a datatype.
2801        As a result, the old regmap is now gone.  Almost all interfaces
2802        in MLRISC change as a consequence.
2803    
2804    2.  A new brand version of machine description tool (v3.0) that generates
2805        modules expecting the new interface.  The old version is removed.
2806    
2807    3.  The RA interface has been further abstracted into two new functors.
2808        RISC_RA and X86RA.  These functors have much simpler interfaces.
2809        [See also directory MLRISC/demo.]
2810    
2811    4.  Some other new source->source code generation tools are available:
2812    
2813        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
2814        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
2815           I use this tool to generate the peephole optimizers---with the new
2816           cell type changes, peephole rules are becoming difficult to write
2817           without conditional pattern matching.
2818    
2819    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
2820        cover the entire MLRISC source tree so many things broke.
2821    
2822    6.  CM files have been moved to the subdirectory MLRISC/cm.
2823        They are moved because there are a lot of them and they clutter up the
2824        root dir.
2825    
2826    7.  More detailed documentation to come...
2827    
2828        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
2829        a makeml -rebuild first.  This is because of other other
2830        changes that Matthias has made (see below).
2831    
2832    
2833    ----------------------------------------------------------------------
2834    Name: Matthias Blume
2835    Date: 2000/11/30 23:12:00 JST
2836    Tag: blume-20001130-filereorg
2837    Description:
2838    
2839    Some manual updates and some file reorganizations in CM.
2840    
2841    ----------------------------------------------------------------------
2842    Name: Matthias Blume
2843    Date: 2000/11/24 17:45:00 JST
2844    Tag: blume-20001124-link
2845    Description:
2846    
2847    Drastically improved link traversal code for the case that the dynamic
2848    value was already loaded at bootstrap time.  As a result, CM and CMB
2849    now both load blazingly fast -- even on a very slow machine.  Also,
2850    memory consumption has been further reduced by this.
2851    
2852    Warning: The format of the PIDMAP file has changed.  THerefore, to
2853    bootstrap you have to do this:
2854    
2855    1. Run CMB.make
2856    2. Make a symbolic link for the boot directory:
2857         ln -s sml.boot.ARCH-OS xxx
2858    3. "Rebuild" the boot directory:
2859         ./makeml -boot xxx -rebuild sml ; rm xxx
2860    4. Boot normally:
2861          ./makeml
2862    
2863    ----------------------------------------------------------------------
2864    Name: Matthias Blume
2865    Date: 2000/11/21 21:20:00 JST
2866    Tag: blume-20001121-tools
2867    Description:
2868    
2869    Continued hacking on autoloading problem -- with success this time.
2870    Also changed tool-plugin mechanism.  See new CM manual.
2871    
2872    ----------------------------------------------------------------------
2873    Name: Matthias Blume
2874    Date: 2000/11/19 14:30:00 JST
2875    Tag:  blume-20001119-autoload
2876    Description:
2877    
2878    Some hacking to make autoloading faster.  Success for CMB, no success
2879    so far for CM.  There is a reduced structure CM' that autoloads faster.
2880    (This is a temporary, non-documented hack to be eliminated again when
2881    the general problem is solved.)
2882    
2883    ----------------------------------------------------------------------
2884    Name: Matthias Blume
2885    Date: 2000/11/17 14:10:00 JST
2886    Tag: blume-20001117-pickle-lib
2887    Description:
2888    
2889    1. Eliminated comp-lib.cm
2890    2. Made pickle-lib.cm
2891    3. Eliminated all uses of intset.sml (from comp-lib.cm)
2892    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
2893       equivalent constructs from smlnj-lib.cm (INtHashTable).
2894    5. Point 4. also goes for those uses of intmap.* in MLRISC.
2895       Duplicated intmap modules thrown out.
2896    6. Hunted down all duplicated SCC code and replaced it with
2897       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
2898    7. Rewrote Feedback module.
2899    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
2900       should be thrown out and equivalent modules from smlnj-lib.cm
2901       should be used (IntRedBlackSet, IntListSet, ...).
2902    
2903    Confirmed that compiler compiles to fixpoint.
2904    
2905    ----------------------------------------------------------------------
2906    Name: Allen Leung
2907    Date: 2000/11/10 18:00:00
2908    Tag: leunga-20001110-new-x86-fp
2909    
2910    A new x86 floating point code generator has been added.
2911    By default this is turned off.  To turn this on, do:
2912    
2913        CM.autoload "$smlnj/compiler.cm";
2914        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
2915    
2916    Changes:
2917    
2918    1.  Changed FTAN to FPTAN so that the assembly output is correct.
2919    2.  Changed the extension callback for FTANGENT to generate:
2920    
2921              fptan
2922              fstp  %st(0)
2923        instead of
2924              fptan
2925              fstpl ftempmem
2926    
2927    3.  Numerous assembly fixes for x86.
2928    
2929    5.  Cleaned up the machine code output module x86/x86MC.sml and added
2930        support for a whole bunch of instructions and addressing modes:
2931    
2932          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
2933          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
2934          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
2935          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
2936          fxch %st(n)
2937          fld %st(n)
2938          fst %st(n)
2939          fst mem
2940          fstp %st(n)
2941          fucom %st(n)
2942          fucomp %st(n)
2943    
2944        All these are now generated when the fast fp mode is turned on.
2945    
2946    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
2947    
2948    ----------------------------------------------------------------------
2949    Name: Matthias Blume
2950    Date: 2000/11/09 11:20:00 JST
2951    Tag: blume-20001109-scc
2952    Description:
2953    
2954    Eliminated some code duplication:
2955    
2956    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
2957       (Otherwise the functor is useless.)
2958    2. Used GraphSCCFn where SCCUtilFun was used previously.
2959    3. Got rid of SCCUtilFun (in comp-lib.cm).
2960    
2961    ----------------------------------------------------------------------
2962    Name: Lal George
2963    Date: 2000/11/06 09:02:21 EST 2000
2964    Tag: Release_110_30
2965    Description:
2966    
2967    - Version 110.30
2968    ----------------------------------------------------------------------
2969    Name: Matthias Blume
2970    Date: 2000/11/04 14:45:00
2971    Tag: blume-20001104-mlbuild
2972    Description:
2973    
2974    - Made ml-build faster on startup.
2975    - Documentation fixes.
2976    
2977    ----------------------------------------------------------------------
2978    Name: Matthias Blume
2979    Date: 2000/11/02 17:00:00 JST
2980    Tag: blume-20001102-condcomp
2981    Description:
2982    
2983    - Small tweaks to pickler -- new BOOTFILES!
2984    - Version bumped to 110.29.2.
2985    - Added conditional compilation facility to init.cmi (see comment there).
2986    ----------------------------------------------------------------------
2987    Name: Allen Leung
2988    Date: 2000/10/23 19:31:00
2989    Tag: leunga-20001023-demo-ra
2990    
2991    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
2992    2. Test programs for the graph library updated
2993    3. Some new MLRISC demo programs added
2994    
2995    ----------------------------------------------------------------------
2996    Name: Matthias Blume
2997    Date: 2000/08/31 22:15:00 JST
2998    Tag: blume-20001017-errmsg
2999    Description:
3000    
3001    More error message grief: Where there used to be no messages, there
3002    now were some that had bogus error regions.  Fixed.
3003    
3004    ----------------------------------------------------------------------
3005    Name: Matthias Blume
3006    Date: 2000/08/31 17:30:00 JST
3007    Tag: blume-20001017-v110p29p1
3008    Description:
3009    
3010    I made a version 110.29.1 with new bootfiles.
3011    
3012    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
3013              CM documentation changes and a small bugfix in CM's error reporting.
3014    
3015    ----------------------------------------------------------------------
3016    Name: Lal George
3017    Date: 2000/09/27 14:42:35 EDT
3018    Tag: george-20000927-nodestatus
3019    Description:
3020    
3021    Changed the type of the nodestatus, so that:
3022    
3023            SPILLED(~1)             is now SPILLED
3024            SPILLED(m) where m>=0   is now MEMREG(m)
3025            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
3026    
3027    ----------------------------------------------------------------------
3028    Name: Matthias Blume
3029    Date: 2000/09/07 14:45:00 JST
3030    Tag: blume-20000907-cmerrmsg
3031    Description:
3032    
3033    Small tweak to CM to avoid getting ML syntax error messages twice.
3034    
3035    ----------------------------------------------------------------------
3036    Name: Matthias Blume
3037    Date: 2000/08/31 18:00:00 JST
3038    Tag: blume-20000831-cvsbootfiles
3039    Description:
3040    
3041    New URL for boot files (because the 110.29 files on the BL server do
3042    now work correctly with my updated install scripts for yacc and lex).
3043    
3044    ----------------------------------------------------------------------
3045    Name: Matthias Blume
3046    Date: 2000/08/08 12:33:00 JST
3047    Tag: blume-20000808-manual
3048    Description:
3049    
3050    Tiny update to CM manual.
3051    
3052    ----------------------------------------------------------------------
3053    Name: Allen Leung
3054    Date: 2000/08/7 19:31:00
3055    Tag: leunga-20000807-a-whole-bunch-of-stuff
3056    
3057      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
3058    and Mobdy.
3059    
3060    1.  x86
3061    
3062       a.  Fixes to peephole module by John and Dan.
3063       b.  Assembly fix to SETcc by Allen.
3064       c.  Fix to c-call by John.
3065       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
3066       e.  Instruction selection optimization to SETcc as suggested by John.
3067    
3068           For example,
3069    
3070            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
3071    
3072           should generate:
3073    
3074            MOVL a, x
3075            SUBL b, x
3076            SHRL 31, x
3077    
3078    2.  IR stuff
3079    
3080         A bunch of new DJ-graph related algorithms added.  These
3081         speed up SSA construction.
3082    
3083    3.  SSA + Scheduling
3084    
3085         Added code for SSA and scheduling to the repository
3086    
3087    ----------------------------------------------------------------------
3088    Name: Lal George
3089    Date: 2000/07/27 11:53:14 EDT
3090    
3091    Tag: lal-20000727-linux-ppc
3092    Description:
3093    
3094     Made changes to support Linux PPC.
3095     p.s. I have confirmation that the 110.29 boot files work fine.
3096    
3097    ----------------------------------------------------------------------
3098    Name: Matthias Blume
3099    Date: 2000/07/27 17:40:00 JST
3100    Tag: blume-20000727-scripts
3101    Description:
3102    
3103    !!!! WARNING !!!!
3104    You must recompile the runtime system!
3105    !!!! WARNING !!!!
3106    
3107    This is basically another round of script-enhancements:
3108    
3109    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
3110       and undefine CM preprocessor symbols.
3111    
3112    2. ml-build avoids generating a new heap image if it finds that the
3113       existing one is still ok.  (The condition is that no ML file had to
3114       be recompiled and all ML files are found to be older that the heap
3115       file.)
3116    
3117       To make this work smoothly, I also hacked the runtime system as
3118       well as SMLofNJ.SysInfo to get access to the heap image suffix
3119       (.sparc-solaris, ...) that is currently being used.
3120    
3121       Moreover, the signature of CM.mk_standalone has changed.  See the
3122       CM manual.
3123    
3124    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
3125       CM manual for details.)
3126    
3127    4. More CM manual updates:
3128        - all of the above has been documented.
3129        - there is now a section describing the (CM-related) command line
3130          arguments that are accepted by the "sml" command
3131    
3132    ----------------------------------------------------------------------
3133    Name: Matthias Blume
3134    Date: 2000/07/25 16:20:00 JST
3135    Tag: blume-20000725-makedepend
3136    Description:
3137    
3138    Added a script called ml-makedepend.  This can be used in makefiles
3139    for Unix' make in a way very similar to the "makedepend" command for
3140    C.
3141    
3142    The script internally uses function CM.sources.
3143    
3144    Synopsis:
3145    
3146        ml-makedepend [-f makefile] cmfile targetname
3147    
3148    The default for the makefile is "makefile" (or "Makefile" should
3149    "makefile" not exist).
3150    
3151    ml-makedepend adds a cmfile/targetname-specific section to this
3152    makefile (after removing the previous version of this section).  The
3153    section contains a single dependency specification with targetname on
3154    the LHS (targetname is an arbitrary name), and a list of files derived
3155    from the cmfile on the RHS.  Some of the files on the RHS are
3156    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
3157    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
3158    path names.  The makefile writer is responsible for making sure that
3159    these variables have correct at the time "make" is invoked.
3160    
3161    ----------------------------------------------------------------------
3162    Name: Matthias Blume
3163    Date: 2000/07/22 23:30:00 JST
3164    Tag: blume-20000722-urlupdate
3165    Description:
3166    
3167    Changed BOOT and config/srcarchiveurl to point to BL server:
3168    
3169        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
3170    
3171    ----------------------------------------------------------------------
3172    Name: Matthias Blume
3173    Date: 2000/07/18 18:00:00 JST
3174    Tag: blume-20000718-Version_110_29
3175    Description:
3176    
3177    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
3178    
3179    2. Updated config/version to 110.29
3180    
3181    3. Updated config/srcarchiveurl
3182    
3183    3. New boot files!
3184       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
3185    
3186    ----------------------------------------------------------------------
3187    Name: Matthias Blume
3188    Date: 2000/07/11 13:58:00 JST
3189    Tag: blume-20000711-doctypo
3190    Description:
3191    
3192    Fixed a few typos in CM manual.
3193    
3194    ----------------------------------------------------------------------
3195    Name: Allen Leung
3196    Date: 2000/06/15 00:38:00
3197    Tag: leunga-20000704-sparc-x86
3198    
3199    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
3200    2. fix to x86 RET bug [found by Dan Grossman]
3201    3. sparc assembly bug fix for ticc instructions [found by Fermin]
3202    
3203       Affects c-- and moby only
3204    
3205    ----------------------------------------------------------------------
3206    Name: Matthias Blume
3207    Date: 2000/07/04 15:26:00
3208    Tag: blume-20000704-trigger
3209    Description:
3210    
3211    1. Improvements to CM manual.
3212    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
3213       of getting a back-trace.  The function, when called, raises an
3214       internal exception which explicitly carries the full back-trace history,
3215       so it is unaffected by any intervening handle-raise pairs ("trivial"
3216       or not).  The interactive loop will print that history once it arrives
3217       at top level.
3218       Short of having all exceptions implicitly carry the full history, the
3219       recommended way of using this facility is:
3220         - compile your program with instrumentation "on"
3221         - run it, when it raises an exception, look at the history
3222         - if the history is "cut off" because of some handler, go and modify
3223           your program so that it explicitly calls BTrace.trigger
3224         - recompile (still instrumented), and rerun; look at the full history
3225    
3226    ----------------------------------------------------------------------
3227    Name: Matthias Blume
3228    Date: 2000/07/03 15:36:00 JST
3229    Tag: blume-20000702-manual
3230    Description:
3231    
3232    Small corrections and updates to CM manual.
3233    
3234    ----------------------------------------------------------------------
3235    Name: Matthias Blume
3236    Date: 2000/06/29 16:04:00 JST
3237    Tag: blume-20000629-yacctool
3238    Description:
3239    
3240    Changes:
3241    
3242    1. Class "mlyacc" now takes separate arguments to pass options to
3243       generated .sml- and .sig-files independently.
3244    2. Corresponding CM manual updates.
3245    3. BTrace module now also reports call sites.  (However, for loop clusters
3246       it only shows from where the cluster was entered.)  There are associated
3247       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
3248    
3249    ----------------------------------------------------------------------
3250    Name: Matthias Blume
3251    Date: 2000/06/27 16:51:00 JST
3252    Tag: blume-20000627-noweb
3253    Description:
3254    
3255    Changes:
3256    
3257     1. Implemented "subdir" and "witness" options for noweb tool.
3258        This caused some slight internal changes in CM's tool implementation.
3259     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
3260        some remaining issues from earlier path anchor changes.
3261     3. Updated CM manual accordingly.
3262    
3263     4. Changed implementation of back-tracing so that I now consider it
3264        ready for prime-time.
3265    
3266        In particular, you don't have to explicitly trigger the back-trace
3267        anymore.  Instead, if you are running BTrace-instrumented code and
3268        there is an uncaught exception (regardless of whether or not it was
3269        raised in instrumented code), the top-level evalloop will print
3270        the back-trace.
3271    
3272        Features:
3273    
3274          - Instrumented and uninstrumented code work together seemlessly.
3275            (Of course, uninstrumented code is never mentioned in actual
3276             back-traces.)
3277    
3278          - Asymptotic time- and space-complexity of instrumented code is
3279            equal to that of uninstrumented code.  (This means that
3280            tail-recursion is preserved by the instrumentation phase.)
3281    
3282          - Modules whose code has been instrumented in different sessions
3283            work together without problem.
3284    
3285          - There is no penalty whatsoever on uninstrumented code.
3286    
3287          - There is no penalty on "raise" expressions, even in
3288            instrumented code.
3289    
3290        A potential bug (or perhaps it is a feature, too):
3291    
3292          A back-trace reaches no further than the outermost instrumented
3293          non-trivial "raise".  Here, a "trivial" raise is one that is the
3294          sole RHS of a "handle" rule.  Thus, back-traces reach trough
3295    
3296               <exp> handle e => raise e
3297    
3298          and even
3299    
3300               <exp> handle Foo => raise Bar
3301    
3302          and, of course, through
3303    
3304               <exp> handle Foo => ...
3305    
3306         if the exception was not Foo.
3307    
3308         Back-traces always reach right through any un-instrumented code
3309         including any of its "handle" expressions, trivial or not.
3310    
3311       To try this out, do the following:
3312    
3313         - Erase all existing binfiles for your program.
3314           (You may keep binfiles for those modules where you think you
3315            definitely don't need back-tracing.)
3316         - Turn on back-trace instrumentation:
3317              SMLofNJ.Internals.BTrace.mode (SOME true);
3318         - Recompile your program.  (I.e., run "CM.make" or "use".)
3319         - You may now turn instrumentation off again (if you want):
3320              SMLofNJ.Internals.BTrace.mode (SOME false);
3321         - Run your program as usual.  If it raises an exception that
3322           reaches the interactive toplevel, then a back-trace will
3323           automatically be printed.  After that, the toplevel loop
3324           will print the exception history as usual.
3325    
3326    ----------------------------------------------------------------------
3327    Name: Matthias Blume
3328    Date: 2000/06/26 09:56:46 JST
3329    Tag: blume-20000626-setup
3330    Description:
3331    
3332    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
3333          ML code before and after compiling a file (e.g., to set compiler
3334          flags)
3335    
3336    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
3337              - associated changes to btrace.sml (BTrace instrumentation pass)
3338              - cleaner implementation of btimp.sml (BTrace tracing and report
3339                module)
3340    
3341    CM manual: * new path encoding documented
3342               * description of setup-parameter to "sml" added
3343    
3344    The biggest user-visible change to back-tracing is that it is no
3345    longer necessary to compile all traced modules within the same
3346    session.  (This was a real limitation.)
3347    
3348    ----------------------------------------------------------------------
3349    Name: Matthias Blume
3350    Date: 2000/06/24 12:40:00 JST
3351    Tag: blume-20000624-startup
3352    Description:
3353    
3354    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
3355    bit too often -- to put it mildly. :)
3356    
3357    ----------------------------------------------------------------------
3358    Name: Matthias Blume
3359    Date: 2000/06/23 18:20:00 JST
3360    Tag: blume-20000623-btrace
3361    Description:
3362    
3363    This updates adds a backtrace facility to aid programmers in debugging
3364    their programs.  This involves the following changes:
3365    
3366    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
3367       keeping track of the current call stack.  When programs are compiled
3368       in a special mode, the compiler will insert calls to these hooks
3369       into the user program.
3370       "Hook" means that it is possible for different implementations of
3371       back-tracing to register themselves (at different times).
3372    
3373    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
3374       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
3375       It can be turned on using this call:
3376         SMLofNJ.Internals.BTrace.mode (SOME true);
3377       Turning it off again:
3378         SMLofNJ.Internals.BTrace.mode (SOME false);
3379       Querying the current status:
3380         SMLofNJ.Internals.BTrace.mode NONE;
3381       Annotated programs are about twice as big as normal ones, and they
3382       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
3383       where all hooks do nothing).  The slowdown with a plugin that is
3384       actually useful (such as the one supplied by default) is even greater,
3385       but in the case of the default plugin it is still only an constant
3386       factor (amortized).
3387    
3388    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
3389       with a sub-structure BTrace for controlling back-tracing.  In particular,
3390       the above-mentioned function "mode" controls whether the annotation
3391       phase is invoked by the compiler.  Another important function is
3392       "trigger": when called it aborts the current execution and causes
3393       the top-level loop to print a full back-trace.
3394    
3395    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
3396       for back-tracing.  It keeps track of the dynamic call stack and in
3397       addition to that it keeps a partial history at each "level" of that
3398       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
3399       d and b (at separate times, dynamically), then the report will show:
3400    
3401       GOTO   d
3402             /c
3403       GOTO  \b
3404       CALL   a
3405    
3406       This shows that there was an initial non-tail call of a, then a
3407       tail-call to b or c, looping behavior in a cluster of functions that
3408       consist of b and c, and then a goto from that cluster (i.e., either from
3409       b or from c) to d.
3410    
3411       Note that (depending on the user program) the amount of information
3412       that the back-trace module has to keep track of at each level is bounded
3413       by a constant.  Thus, the whole implementation has the same asymptotical
3414       complexity as the original program (both in space and in time).
3415    
3416    5. compiler/TopLevel/interact/evalloop.sml has been modified to
3417       handle the special exception SMLofNJ.Internals.BTrace.BTrace
3418       which is raised by the "trigger" function mentioned above.
3419    
3420    Notes on usage:
3421    
3422    - Annotated code works well together with unannotated code:
3423    Unannotated calls simply do not show up at all in the backtrace.
3424    
3425    - It is not a good idea to let modules that were annotated during
3426    different sessions run at the same time.  This is because the compiler
3427    chooses small integers to identify individual functions, and there
3428    will be clashes if different modules were compiled in separate sessions.
3429    (Nothing will crash, and you will even be told about the clashes, but
3430    back-trace information will in general not be useful.)
3431    
3432    - Back-tracing can be confused by callcc and capture.
3433    
3434    - The only way of getting a back-trace right now is to explicitly
3435    invoke the "trigger" function from your user program.  Eventually, we
3436    should make every exception carry back-trace information (if
3437    available).  But since this creates more overhead at "raise"-time
3438    (similar to the current exnHistory overhead), I have not yet
3439    implemented this.  (The implementation will be rather easy.)  With
3440    exceptions carrying back-trace information, this facility will be even
3441    more useful because users don't need to modify their programs...
3442    
3443    - While it is possible to compile the compiler with back-trace
3444    annotations turned on (I did it to get some confidence in
3445    correctness), you must make absolutely sure that core.sml and
3446    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
3447    be compiled with annotation because there is no core access yet, but
3448    if you compile btimp.sml with annotation, then the system will go into
3449    an infinite recursion and crash.)
3450    Since CM currently does not know about BTrace, the only way to turn
3451    annotations on and off for different modules of the compiler is to
3452    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
3453    this is awkward and clumsy.
3454    
3455    Sample sessions:
3456    
3457    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
3458    - SMLofNJ.Internals.BTrace.mode (SOME true);
3459    [autoloading]
3460    [autoloading done]
3461    val it = false : bool
3462    - structure X = struct
3463    -     fun main n = let
3464    -         fun a (x, 0) = d x
3465    -           | a (x, n) = b (x, n - 1)
3466    -         and b (x, n) = c (x, n)
3467    -         and c (x, n) = a (x, n)
3468    -         and d x = e (x, 3)
3469    -         and e (x, 0) = f x
3470    -           | e (x, n) = e (x, n - 1)
3471    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
3472    -           | f n = n * g (n - 1)
3473    -         and g n = a (n, 3)
3474    -     in
3475    -         f n
3476    -     end
3477    - end;
3478    structure X : sig val main : int -> int end
3479    - X.main 3;
3480    *** BACK-TRACE ***
3481    GOTO   stdIn:4.2-13.20: X.main[2].f
3482    GOTO-( stdIn:4.2-13.20: X.main[2].e
3483    GOTO   stdIn:4.2-13.20: X.main[2].d
3484         / stdIn:4.2-13.20: X.main[2].a
3485         | stdIn:4.2-13.20: X.main[2].b
3486    GOTO-\ stdIn:4.2-13.20: X.main[2].c
3487    CALL   stdIn:4.2-13.20: X.main[2].g
3488    GOTO   stdIn:4.2-13.20: X.main[2].f
3489    GOTO-( stdIn:4.2-13.20: X.main[2].e
3490    GOTO   stdIn:4.2-13.20: X.main[2].d
3491         / stdIn:4.2-13.20: X.main[2].a
3492         | stdIn:4.2-13.20: X.main[2].b
3493    GOTO-\ stdIn:4.2-13.20: X.main[2].c
3494    CALL   stdIn:4.2-13.20: X.main[2].g
3495    GOTO   stdIn:4.2-13.20: X.main[2].f
3496    GOTO-( stdIn:4.2-13.20: X.main[2].e
3497    GOTO   stdIn:4.2-13.20: X.main[2].d
3498         / stdIn:4.2-13.20: X.main[2].a
3499         | stdIn:4.2-13.20: X.main[2].b
3500    GOTO-\ stdIn:4.2-13.20: X.main[2].c
3501    CALL   stdIn:4.2-13.20: X.main[2].g
3502    GOTO   stdIn:4.2-13.20: X.main[2].f
3503    CALL   stdIn:2.15-17.4: X.main[2]
3504    -
3505    
3506    (Note that because of a FLINt bug the above code currently does not
3507    compile without BTrace turned on.)
3508    
3509    Here is another example, using my modified Tiger compiler:
3510    
3511    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
3512    - SMLofNJ.Internals.BTrace.mode (SOME true);
3513    [autoloading]
3514    [autoloading done]
3515    val it = false : bool
3516    - CM.make "sources.cm";
3517    [autoloading]
3518    ...
3519    [autoloading done]
3520    [scanning sources.cm]
3521    [parsing (sources.cm):parse.sml]
3522    [creating directory CM/SKEL ...]
3523    [parsing (sources.cm):tiger.lex.sml]
3524    ...
3525    [wrote CM/sparc-unix/semant.sml]
3526    [compiling (sources.cm):main.sml]
3527    [wrote CM/sparc-unix/main.sml]
3528    [New bindings added.]
3529    val it = true : bool
3530    - Main.compile ("../testcases/merge.tig", "foo.out");
3531    *** BACK-TRACE ***
3532    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
3533    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3534    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
3535    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
3536    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3537    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3538    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
3539         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
3540    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
3541    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3542    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
3543    CALL   main.sml:1.18-118.4: Main.compile[2]
3544    -
3545    
3546    ----------------------------------------------------------------------
3547    Name: Matthias Blumen
3548    Date: 2000/06/21 18:00:00 JST
3549    Tag: blume-20000621-manual
3550    Description:
3551    
3552    CM manual update: Path environments documented.
3553    
3554    ----------------------------------------------------------------------
3555    Name: Matthias Blume
3556    Date: 2000/06/19 13:40:00
3557    Tag: blume-20000619-manual
3558    Description:
3559    
3560    CM manual and system/README update.  This only covers the fact that
3561    there are no more implicit anchors.  (Path environments and the "bind"
3562    option to "cm" have yet to be documented.)
3563    
3564    ----------------------------------------------------------------------
3565    Name: Matthias Blume
3566    Date: 2000/06/19 11:05:00 JST
3567    Tag: blume-20000619-chdir-bugfix
3568    Description:
3569    
3570    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
3571    
3572    ----------------------------------------------------------------------
3573    Name: Matthias Blume
3574    Date: 2000/06/18 22:00:10 JST
3575    Tag: blume-20000618-implicit-anchors-really-gone
3576    Description:
3577    
3578    I updates the previous HISTORY entry where I forgot to mention that
3579    implicit anchors are no longer with us.
3580    
3581    The current update also gets rid of the (now useless) controller
3582    CM.Control.implicit_anchors.
3583    
3584    ----------------------------------------------------------------------
3585    Name: Matthias Blume
3586    Date: 2000/06/16 17:30:00 JST
3587    Tag: blume-20000616-anchorenv
3588    Description:
3589    
3590    This patch implements the long anticipated (just kidding :) "anchor
3591    environment" mechanism.  In the course of doing this, I also
3592    re-implemented CM's internal "SrcPath" module from scratch.  The new
3593    one should be more robust in certain boundary cases.  In any case, it
3594    is a lot cleaner than its predecessor (IMHO).
3595    
3596    This time, although there is yet another boot file format change, I
3597    kept the unpickler backward-compatible.  As a result, no new bootfiles
3598    are necessary and bootstrapping is straightforward.  (You cannot read
3599    new bootfiles into an old system, but the other way around is no
3600    problem.)
3601    
3602    Visible changes:
3603    
3604    ** 0. Implicit path anchors (without the leading $-symbol) are no
3605    longer recognized at all. This means that such path names are not
3606    illegal either.  For example, the name basis.cm simply refers to a
3607    local file called "basis.cm" (i.e, the name is an ordinary path
3608    relative to .cm-files directory).  Or, to put it differently, only
3609    names that start with $ are anchored paths.
3610    
3611    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
3612    vanished.
3613    
3614    John (Reppy) had critizised this as soon as I originally proposed and
3615    implemented it, but at that time I did not really deeply believe
3616    him. :) Now I came full-circle because I need the $<singlearc> syntax
3617    in another place where it cannot be seen as an abbreviation for
3618    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
3619    seems to mean (i.e., it "expands" into the corresponding anchor
3620    value).
3621    
3622    However, when paths are used as members in CM description files, it
3623    continues to be true that there must be at least another arc after the
3624    anchor.  This is now enforced separately during semantic analysis
3625    (i.e., from a lexical/syntactical point of view, the notation is ok.)
3626    
3627    ** 2. The "cm" class now accepts an option "bind".  The option's value
3628    is a sub-option list of precisely two items -- one labeled "anchor"
3629    and the other one labeled "value".  As you might expect, "anchor" is
3630    used to specify an anchor name to be bound, and "value" specifies what
3631    the anchor is being bound to.
3632    
3633    The value must be a directory name and can be given in either standard
3634    syntax (including the possibility that it is itself an anchored path)
3635    or native syntax.
3636    
3637    Examples:
3638    
3639       foo.cm (bind:(anchor:bar value:$mystuff/bar))
3640       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
3641    
3642    and so on.
3643    
3644    The meaning of this is that the .cm-file will be processed with an
3645    augmented anchor environment where the given anchor(s) is/are bound to
3646    the given values(s).
3647    
3648    The rationale for having this feature is this: Suppose you are trying
3649    to use two different (already stable) libraries a.cm and b.cm (that
3650    you perhaps didn't write yourself).  Further, suppose each of these
3651    two libraries internally uses its own auxiliary library $aux/lib.cm.
3652    Normally you would now have a problem because the anchor "lib" can not
3653    be bound to more than one value globally.  Therefore, the project that
3654    uses both a.cm and b.cm must locally redirect the anchor to some other
3655    place:
3656    
3657       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
3658       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
3659    
3660    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
3661    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
3662    
3663    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
3664    you when you do so at the time of CM.stabilize).  Therefore, you can
3665    also use an anchored path as the value:
3666    
3667      a.cm (bind:(anchor:lib value:$a-lib))
3668      b.cm (bind:(anchor:lib value:$b-lib))
3669    
3670    Now you can globally configure (using the usual CM.Anchor.anchor or
3671    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
3672    itself is always locally bound, setting it globally is no longer
3673    meaningful or necessary (but it does not hurt either).  In fact, "lib"
3674    can still be used as a global anchor for separate purposes.  As a
3675    matter of fact, one can locally define "lib" in terms of a global
3676    "lib":
3677    
3678      a.cm (bind:(anchor:lib value:$lib/a))
3679      b.cm (bind:(anchor:lib value:$lib/b))
3680    
3681    ** 3: The encoding of path names has changed.  This affects the way
3682    path names are shown in CM's progress report and also the internal
3683    protocol encoding used for parallel make.
3684    
3685    The encoding now uses one or more ':'-separated segments.  Each
3686    segments corresponds to a file that has been specified relative to the
3687    file given by its preceding segment.  The first segment is either
3688    relative to the CWD, absolute, or anchored.  Each segment itself is
3689    basically a Unix pathname; all segments but the first are relative.
3690    
3691    Example:
3692    
3693       $foo/bar/baz.cm:a/b/c.sml
3694    
3695    This path denotes the file bar/a/b/c.sml relative to the directory
3696    denoted by anchor "foo".  Notice that the encoding also includes
3697    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
3698    paths are resolved relative to the .cm-files directory, so baz.cm must
3699    be ignored to get the "real" pathname.
3700    
3701    To make this fact more obvious, CM puts the names of such "virtual
3702    arcs" into parentheses when they appear in progress reports. (No
3703    parentheses will appear in the internal protocol encoding.)  Thus,
3704    what you really see is:
3705    
3706      $foo/bar/(baz.cm):a/b/c.sml
3707    
3708    I find this notation to be much more informative than before.
3709    
3710    Another new feature of the encoding is that special characters
3711    including parentheses, colons, (back)slashes, and white space are
3712    written as \ddd (where ddd is the decimal encoding of the character).
3713    
3714    *** The CM manual still needs to be updated.
3715    
3716    ----------------------------------------------------------------------
3717    Name: Allen Leung
3718    Date: 2000/06/15 00:38:00
3719    Tag: leunga-20000615-x86-peephole
3720    
3721    x86 Peephole fix by Fermin.  Affects c-- and moby only.
3722    
3723    ----------------------------------------------------------------------
3724    Name: Matthias Blume
3725    Date: 2000/06/12 11:40:00
3726    Tag: blume-20000612-parmakefix
3727    Description:
3728    
3729    More cleanup after changing the file naming scheme: This time I
3730    repaired the parallel make mechanism for CMB.make which I broke earlier.
3731    
3732    ----------------------------------------------------------------------
3733    Name: Allen Leung
3734    Date: 2000/06/09 01:25:00
3735    Tag: leunga-20000609-various
3736    
3737    None of these things should affect normal SML/NJ operations
3738    
3739    1. Peephole improvements provided by Fermin (c--)
3740    2. New annotation DEFUSE for adding extra dependence (moby)
3741    3. New X86 LOCK instructions (moby)
3742    4. New machine description language for reservation tables (scheduling)
3743    5. Fixes to various optimization/analysis modules (branch chaining, dominator
3744       trees etc.)
3745    6. I've changed the CM files so that they can work with versions
3746       110.0.6, 110.25 and 110.28
3747    
3748    ----------------------------------------------------------------------
3749    Name: Matthias Blume
3750    Date: 2000/06/09 12:40:00
3751    Tag: blume-20000609-log
3752    Description:
3753    
3754    - Removed all(?) remaining RCS Log entries from sources.
3755    
3756    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
3757      anchored paths).
3758    
3759    ----------------------------------------------------------------------
3760    Name: Matthias Blume
3761    Date: 2000/06/07 17:00:00 JST
3762    Tag: blume-20000607-no-implicit-anchors
3763    Description:
3764    
3765    1. This update changes the default setting for
3766    CM.Control.implicit_anchors from true to false.  This means that
3767    implicit anchors are no longer permitted by default.  I also tried to
3768    make sure that nothing else still relies on implicit anchors.
3769    (This is the next step on the schedule towards a CM that does not even
3770    have the notion of implicit anchors anymore.)
3771    
3772    2. More CM manual updates.
3773    
3774    3. I managed to track down and fix the pickling bug I mentioned last
3775    time.  Because of the previously existing workaround, this entails no
3776    immediate practical changes.
3777    
3778    ----------------------------------------------------------------------
3779    Name: Matthias Blume
3780    Date: 2000/06/06 11:15:00 JST
3781    Tag: blume-20000606-lazierpickle
3782    Description:
3783    
3784    !!!! NEW BOOT FILES !!!!
3785    
3786    * The main purpose of this update is to make library pickles lazier in
3787    order to reduce the initial space penalty for autoloading a library.
3788    As a result, it is now possible to have $smlnj/compiler.cm
3789    pre-registered.  This should take care of the many complaints or
3790    inquiries about missing structure Compiler.  This required changes to
3791    CM's internal data structures and small tweaks to some algorithms.
3792    
3793    As a neat additional effect, it is no longer necessary (for the sake
3794    of lean heap image files) to distinguish between a "minimal" CM and a
3795    "full" CM.  Now, there is only one CM (i.e., the "full" version:
3796    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
3797    interactive top level. ($smlnj/cm/minimal.cm is gone.)
3798    
3799    To make the life of compiler-hackers easier, "makeml" now also
3800    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
3801    words, after you bootstrap a new sml for the first time, you will not
3802    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
3803    around you will still have to do it, though.)
3804    
3805    * A second change consists of major updates to the CM manual.  There
3806    are now several appendices with summary information and also a full
3807    specification of the CM description file syntax.
3808    
3809    * In directory src/system I added the script "allcross".  This script
3810    invokes sml and cross-compiles the compiler for all supported
3811    architectures.  (Useful when providing a new set of boot files.)
3812    
3813    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
3814    added a small tweak to pickle-util.sml to work around this problem,
3815    but it is not a proper fix yet.  I will investigate further.  (The
3816    effect of the bug was an inflation of library pickle size.)
3817    
3818    * Version number increased to 110.28.1 (to avoid compatibility problems).
3819    
3820    ----------------------------------------------------------------------
3821    Name: Allen Leung
3822    Date: 2000/05/25 17:28 EDT
3823    Tag: leunga-20000525-ra
3824    Description:
3825    
3826      Fixed a bug in freezing phase of the register allocator.
3827    
3828    ----------------------------------------------------------------------
3829    Name: Allen Leung
3830    Date: 2000/05/15 22:53 EDT
3831    Tag: leunga-20000515-alpha-x86-ra
3832    Description:
3833    
3834      1. Alpha
3835    
3836          Slight cleanup.  Removed the instruction SGNXL
3837    
3838      2. X86
3839    
3840          Added the following instructions to the instruction set:
3841    
3842            ROLx, RORx,
3843            BTx, BTSx, BTLx, BTRx,
3844            XCHGx, and variants with the LOCK prefix
3845