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 876, Wed Jul 11 15:42:20 2001 UTC revision 976, Wed Nov 14 14:54:03 2001 UTC
# Line 13  Line 13 
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Lal George
17    Date: 2001/11/14 09:44:04 EST
18    Tag:
19    Description:
20    
21      Fix for a backpatching bug reported by Allen.
22    
23      Because the boundary between short and long span-dependent
24      instructions is +/- 128, there are an astounding number of
25      span-dependent instructions whose size is over estimated.
26    
27      Allen came up with the idea of letting the size of span
28      dependent instructions be non-monotonic, for a maxIter
29      number of times, after which the size must be monotonically
30      increasing.
31    
32      This table shows the number of span-dependent instructions
33      whose size was over-estimated as a function of maxIter, for the
34      file Parse/parse/ml.grm.sml:
35    
36         maxIter            # of instructions:
37            10                      687
38            20                      438
39            30                      198
40            40                        0
41    
42      In compiling the compiler, there is no significant difference in
43      compilation speed between maxIter=10 and maxIter=40. Actually,
44      my measurements showed that maxIter=40 was a tad faster than
45      maxIter=10! Also 96% of the  files in the compiler reach a fix
46      point within 13 iterations, so fixing maxIter at 40, while high,
47      is okay.
48    
49    ----------------------------------------------------------------------
50    Name: Matthias Blume
51    Date: 2001/10/31 15:25:00 EST
52    Tag: blume-20011031-pgraph
53    Description:
54    
55    CKIT:
56    * Changed the "Function" constructor of type Ast.ctype to carry optional
57      argument identifiers.
58    * Changed the return type of TypeUtil.getFunction accordingly.
59    * Type equality ignores the argument names.
60    * TypeUtil.composite tries to preserve argument names but gives up quickly
61      if there is a mismatch.
62    
63    installation script:
64    * attempts to use "curl" if available (unless "wget" is available as well)
65    
66    CM:
67    * has an experimental implementation of "portable graphs" which I will
68      soon propose as an implementation-independent library format
69    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
70    
71    NLFFI-LIB:
72    * some cleanup (all cosmetic)
73    
74    NLFFIGEN:
75    * temporarily disabled the mechanism that suppresses ML output for
76      C definitions whose identifiers start with an underscore character
77    * generate val bindings for enum constants
78    * user can request that only one style (light or heavy) is being used;
79      default is to use both (command-line arguments: -heavy and -light)
80    * fixed bug in handling of function types involving incomplete pointers
81    * generate ML entry points that take record arguments (i.e., using
82      named arguments) for C functions that have a prototype with named
83      arguments
84      (see changes to CKIT)
85    
86    ----------------------------------------------------------------------
87    Name: Allen Leung
88    Date: 2001/10/27 20:34:00 EDT
89    Tag: leunga-20011027-x86-fast-fp-call
90    Description:
91    
92       Fixed the bug described in blume-20010920-slowfp.
93    
94       The fix involves
95          1. generating FCOPYs in FSTP in ia32-svid
96          2. marking a CALL with the appropriate annotation
97    
98    ----------------------------------------------------------------------
99    Name: Matthias Blume
100    Date: 2001/10/16 11:32:00 EDT
101    Tag: blume-20011016-netbsd
102    Description:
103    
104    Underscore patch from Chris Richards (fixing problem with compiling
105    runtime system under recent NetBSD).
106    
107    ----------------------------------------------------------------------
108    Name: Allen Leung
109    Date: 2001/10/12 17:18:32 EDT 2001
110    Tag: leung-20011012-x86-printflowgraph
111    Description:
112    
113    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
114    
115    ----------------------------------------------------------------------
116    Name: Lal George
117    Date: 2001/10/11 23:51:34 EDT
118    Tag: george-20011011-too-many-instrs
119    Description:
120    
121    The representation of a program point never expected to see more
122    than 65536 instructions in a basic block!
123    
124    ----------------------------------------------------------------------
125    Name: Lal George
126    Date: 2001/10/09 09:41:37 EDT
127    Tag: george-20011008-mlrisc-labels
128    Description:
129    
130    Changed the machine description files to support printing of
131    local and global labels in assembly code, based on host assembler
132    conventions.
133    
134    ----------------------------------------------------------------------
135    Name: Matthias Blume
136    Date: 2001/09/25 15:25:00 EDT
137    Tag: blume-20010925-exninfo
138    Description:
139    
140    I provided a non-hook implementation of exnName (at the toplevel) and
141    made the "dummy" implementation of exnMessage (at the toplevel) more
142    useful: if nothing gets "hooked in", then at least you are going to
143    see the exception name and a message indicating why you don't see more.
144    
145    [For the time being, programs that need exnMessage and want to use
146    ml-build should either use General.exnMessage (strongly recommended) or
147    refer to structure General at some other point so that CM sees a
148    static dependency.]
149    
150    [Similar remarks go for "print" and "use":  If you want to use their
151    functionality in stand-alone programs generated by ml-build, then use
152    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
153    
154    ----------------------------------------------------------------------
155    Name: Matthias Blume
156    Date: 2001/09/20 17:28:00 EDT
157    Tag: blume-20010920-slowfp
158    Description:
159    
160    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
161    
162    ----------------------------------------------------------------------
163    Name: Matthias Blume
164    Date: 2001/09/20 17:20:00 EDT
165    Tag: blume-20010920-canonicalpaths
166    Description:
167    
168    0. Updated the BOOT file (something that I forgot to do earlier).
169    
170    1. Small internal change to CM so that it avoids "/../" in filenames
171       as much as possible (but only where it is safe).
172    
173    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
174       that arguments that contain delimiters are passed through correctly.
175       This change also means that all "special" arguments of the form
176       @SMLxxx... must come first.
177    
178    3. Changed install script to put relative anchor names for tool commands
179       into pathconfig.
180    
181    ----------------------------------------------------------------------
182    Name: Matthias Blume
183    >>>>>>> 1.169
184    Date: 2001/09/18 15:35:00 EDT
185    Tag: blume-20010918-readme11036
186    Description:
187    
188    Added README files.
189    
190    ----------------------------------------------------------------------
191    Name: Matthias Blume
192    Date: 2001/09/18 11:45:00 EDT
193    Tag: Release_110_36 (retag)
194    Description:
195    
196    Fixed mistake in config/preloads. Retagged as 110.36.
197    
198    ----------------------------------------------------------------------
199    Name: Matthias Blume
200    Date: 2001/09/18 09:40:00 EDT
201    Tag: Release_110_36_orig (tag changed)
202    Description:
203    
204    New version (110.36).  New bootfiles.
205    
206    ----------------------------------------------------------------------
207    Name: Matthias Blume
208    Date: 2001/09/14 16:15:00 EDT
209    Tag: blume-20010914-x86fastfp
210    Description:
211    
212    John committed some changes that Allen made, in particular a (hopefully)
213    correctly working version of the x86-fp module.
214    
215    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
216    flag to "true".  Everything seems to compile to a fixpoint ok, and
217    "mandelbrot" speeds up by about 15%.
218    
219    ----------------------------------------------------------------------
220    Name: Matthias Blume
221    Date: 2001/09/13 11:20:00 EDT
222    Tag: blume-20010913-minimal
223    Description:
224    
225    1. Stefan Monnier's patch to fix a miscompilation problem that
226       was brought to light by John Reppy's work on Moby.
227    
228    2. Implemented a minimal "structure Compiler" that contains just
229       "version" and "architecture".  The minimal version will be
230       available when the full version is not.  This is for backward-
231       compatibility with code that wants to test Compiler.version.
232    
233    ----------------------------------------------------------------------
234    Name: Matthias Blume
235    Date: 2001/08/28 14:03:00 EDT
236    Tag: blume-20010828-ml-lex
237    Description:
238    
239    Fix for bug 1581, received from Neophytos Michael.
240    
241    ----------------------------------------------------------------------
242    Name: Matthias Blume
243    Date: 2001/08/27 11:20:00 EDT
244    Tag: blume-20010827-readme11035
245    Description:
246    
247    Fleshed out the README file for 110.35.
248    
249    ----------------------------------------------------------------------
250    Name: Matthias Blume
251    Date: 2001/08/24 17:10:00 EDT
252    Tag: Release_110_35
253    Description:
254    
255    New version number (110.35).  New bootfiles.
256    
257    ----------------------------------------------------------------------
258    Name: Lal George
259    Date: 2001/08/24 13:47:18 EDT 2001
260    Tag: george-20010824-MLRISC-graphs
261    Description:
262    
263     removed clusters from MLRISC completely and replaced with graphs.
264    
265    ----------------------------------------------------------------------
266    Name: Matthias Blume
267    Date: 2001/08/23 17:50:00 EDT
268    Tag: blume-20010823-toplevel
269    Description:
270    
271    - some reorganization of the code that implements various kinds of
272      environments in the compiler (static, dynamic, symbolic, combined)
273    - re-implemented the EnvRef module so that evalStream works properly
274      (if the stream contains references to "use", "CM.make", etc.)
275    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
276    
277    ----------------------------------------------------------------------
278    Name: Matthias Blume
279    Date: 2001/08/20 15:50 EDT
280    Tag: blume20010820-slipup
281    Description:
282    
283    I forgot to commit a few files.  Here they are...
284    
285    ----------------------------------------------------------------------
286    Name: Matthias Blume
287    Date: 2001/08/20 15:35:00 EDT
288    Tag: blume-20010820-debugprof
289    Description:
290    
291    !!!! NEW BOOTFILES !!!!
292    
293    This is another round of reorganizing the compiler sources.  This
294    time the main goal was to factor out all the "instrumentation"
295    passes (for profiling and backtracing) into their own library.
296    The difficulty was to do it in such a way that it does not depend
297    on elaborate.cm but only on elabdata.cm.
298    
299    Therefore there have been further changes to both elaborate.cm and
300    elabdata.cm -- more "generic" things have been moved from the former
301    to the latter.  As a result, I was forced to split the assignment
302    of numbers indicating "primtyc"s into two portions: SML-generic and
303    SML/NJ-specific.  Since it would have been awkward to maintain,
304    I bit the bullet and actually _changed_ the mapping between these
305    numbers and primtycs.  The bottom line of this is that you need
306    a new set of bin- and bootfiles.
307    
308    I have built new bootfiles for all architectures, so doing a fresh
309    checkout and config/install.sh should be all you need.
310    
311    The newly created library's name is
312    
313        $smlnj/viscomp/debugprof.cm
314    
315    and its sources live under
316    
317        src/compiler/DebugProf
318    
319    ----------------------------------------------------------------------
320    Name: Matthias Blume
321    Date: 2001/08/15 17:15:00 EDT
322    Tag: blume-20010815-compreorg
323    Description:
324    
325    This is a first cut at reorganizing the CM libraries that make up the
326    core of the compiler.  The idea is to separate out pieces that could
327    be used independently by tools, e.g., the parser, the typechecker, etc.
328    
329    The current status is a step in this direction, but it is not quite
330    satisfactory yet.  Expect more changes in the future.
331    
332    Here is the current (new) organization...
333    
334        What used to be $smlnj/viscomp/core.cm is now divided into
335        six CM libraries:
336    
337             $smlnj/viscomp/basics.cm
338                           /parser.cm
339                           /elabdata.cm
340                           /elaborate.cm
341                           /execute.cm
342                           /core.cm
343    
344        The CM files for these libraries live under src/system/smlnj/viscomp.
345        All these libraries are proxy libraries that contain precisely
346        one CM library component.  Here are the locations of the components
347        (all within the src/compiler tree):
348    
349             Basics/basics.cm
350             Parse/parser.cm
351             ElabData/elabdata.cm
352             Elaborator/elaborate.cm
353             Execution/execute.cm
354             core.cm
355    
356         [This organization is the same that has been used already
357         for a while for the architecture-specific parts of the visible
358         compiler and for the old version of core.cm.]
359    
360         As you will notice, many source files have been moved from their
361         respective original locations to a new home in one of the above
362         subtrees.
363    
364         The division of labor between the new libraries is the following:
365    
366             basics.cm:
367                - Simple, basic definitions that pertain to many (or all) of
368                  the other libraries.
369             parser.cm:
370                - The SML parser, producing output of type Ast.dec.
371                - The type family for Ast is also defined and exported here.
372             elabdata.cm:
373                - The datatypes that describe input and output of the elaborator.
374                  This includes types, absyn, and static environments.
375             elaborator.cm:
376                - The SML/NJ type checker and elaborator.
377                  This maps an Ast.dec (with a given static environment) to
378                  an Absyn.dec (with a new static environment).
379                - This libraries implements certain modules that used to be
380                  structures as functors (to remove dependencies on FLINT).
381             execute.cm:
382                - Everything having to do with executing binary code objects.
383                - Dynamic environments.
384             core.cm:
385                - SML/NJ-specific instantiations of the elaborator and MLRISC.
386                - Top-level modules.
387                - FLINT (this should eventually become its own library)
388    
389    Notes:
390    
391    I am not 100% happy with the way I separated the elaborator (and its
392    data structures) from FLINT.  Two instances of the same problem:
393    
394        1. Data structures contain certain fields that carry FLINT-specific
395           information.  I hacked around this using exn and the property list
396           module from smlnj-lib.  But the fact that there are middle-end
397           specific fields around at all is a bit annoying.
398    
399        2. The elaborator calculates certain FLINT-related information.  I tried
400           to make this as abstract as I could using functorization, but, again,
401           the fact that the elaborator has to perform calculations on behalf
402           of the middle-end at all is not nice.
403    
404        3. Having to used exn and property lists is unfortunate because it
405           weakens type checking.  The other alternative (parameterizing
406           nearly *everything*) is not appealing, though.
407    
408    I removed the "rebinding =" warning hack because due to the new organization
409    it was awkward to maintain it.  As a result, the compiler now issues some of
410    these warnings when compiling init.cmi during bootstrap compilation. On
411    the plus side, you also get a warning when you do, for example:
412       val op = = Int32.+
413    which was not the case up to now.
414    
415    I placed "assign" and "deref" into the _Core structure so that the
416    code that deals with the "lazy" keyword can find them there.  This
417    removes the need for having access to the primitive environment
418    during elaboration.
419    
420    ----------------------------------------------------------------------
421    Name: Matthias Blume
422    Date: 2001/08/13
423    Tag: blume-20010813-closures
424    Description:
425    
426    This fix was sent to us by Zhong Shao.  It is supposed to improve the
427    performance of certain loops by avoiding needless closure allocation.
428    
429    ----------------------------------------------------------------------
430    Name: Lal George
431    Date: 2001/07/31 10:03:23 EDT 2001
432    Tag: george-20010731-x86-fmalloc
433    Description: Fixed bug in x86 calls
434    
435        There was a bug where call instructions would mysteriously
436        vanish. The call instruction had to be one that returned
437        a floating point value.
438    
439    ----------------------------------------------------------------------
440    Name: Lal George
441    Date: 2001/07/19 16:36:29 EDT 2001
442    Tag: george-20010719-simple-cells
443    Description:
444    
445    I have dramatically simplified the interface for CELLS in MLRISC.
446    
447    In summary, the cells interface is broken up into three parts:
448    
449      1. CellsBasis : CELLS_BASIS
450    
451            CellsBasis is a top level structure and common for all
452            architectures.  it contains the definitions of basic datatypes
453            and utility  functions over these types.
454    
455      2. functor Cells() : CELLS
456    
457            Cells generates an interface for CELLS that incorporates the
458            specific resources on the target architecture, such as the
459            presence of special register classes, their number and size,
460            and various useful substructures.
461    
462      3. <ARCH>CELLS
463    
464            e.g. SparcCells: SPARCCELLS
465    
466            <ARCH>CELLS usually contains additional bindings for special
467            registers  on the architecture, such as:
468    
469                    val r0 : cell           (* register zero *)
470                    val y : cell            (* Y register *)
471                    val psr : cell          (* processor status register *)
472                    ...
473    
474            The structure returned by applying the Cells functor is opened
475            in this interface.
476    
477    The main implication of all this is that the datatypes for cells is
478    split between CellsBasis and CELLS -- a fairly simple change for user
479    code.
480    
481    In the old scheme the CELLS interface had a definitional binding of
482    the form:
483    
484            signature CELLS = sig
485    
486               structure CellsBasis = CellsBasis
487    
488               ...
489    
490            end
491    
492    With all the sharing constraints that goes on in MLRISC, this old
493    design  quickly leads to errors such as:
494    
495            "structure definition spec inside of sharing ... "
496    
497    
498    and appears to require an unacceptable amount of sharing and where
499    constraint hackery.
500    
501    I think this error message (the interaction of definitional specs and
502    sharing) requires more explanation on our web page.
503    
504    ----------------------------------------------------------------------
505    Name: Matthias Blume
506    Date: 2001/07/19 15:00:00 EDT
507    Tag: blume-20010719-libreorg
508    Description:
509    
510    This update puts together a fairly extensive but straightforward change
511    to the way the libraries that implement the interactive system are
512    organized:
513    
514       The biggest change is the elimination of structure Compiler.  As a
515       replacement for this structure, there is now a CM library
516       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
517       that exports all the substructures of the original structure Compiler
518       directly.  So instead of saying Compiler.Foo.bar one now simply
519       says Foo.bar.  (The CM libraries actually export a collection of
520       structures that is richer than the collection of substructures of
521       structure Compiler.)
522    
523       To make the transition smooth, there is a separate library called
524       $smlnj/compiler/compiler.cm which puts together and exports the
525       original structure Compiler (or at least something very close to it).
526    
527       There are five members of the original structure Compiler
528       that are not exported directly but which instead became members
529       of a new structure Backend (described by signature BACKEND).  These are:
530       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
531       Interact (: INTERACT), structure Machine (: MACHINE), and val
532       architecture (: string).
533    
534       Structure Compiler.Version has become structure CompilerVersion.
535    
536       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
537       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
538       or x86, respectively.
539       Each of these exports the same frontend structures that
540       $smlnj/compiler.cm exports.  But they do not have a structure Backend
541       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
542       Hppa, PPC, Sparc, or X86, respectively.
543    
544       Library $smlnj/compiler/all.cm exports the union of the exports of
545       $smlnj/compiler/<arch>.cm
546    
547       There are no structures <Arch>Compiler anymore, use
548       $smlnj/compiler/<arch>.cm instead.
549    
550       Library host-compiler-0.cm is gone.  Instead, the internal library
551       that instantiates CM is now called cm0.cm.  Selection of the host
552       compiler (backend) is no longer done here but. (Responsibility for it
553       now lies with $smlnj/compiler/current.cm.  This seems to be more
554       logical.)
555    
556       Many individual files have been moved or renamed.  Some files have
557       been split into multiple files, and some "dead" files have been deleted.
558    
559    Aside from these changes to library organization, there are also changes
560    to the way the code itself is organized:
561    
562       Structure Binfile has been re-implemented in such a way that it no
563       longer needs any knowledge of the compiler.  It exclusively deals
564       with the details of binfile layout.  It no longer invokes the
565       compiler (for the purpose of creating new prospective binfile
566       content), and it no longer has any knowledge of how to interpret
567       pickles.
568    
569       Structure Compile (: COMPILE) has been stripped down to the bare
570       essentials of compilation.  It no longer deals with linking/execution.
571       The interface has been cleaned up considerably.
572    
573       Utility routines for dealing with linking and execution have been
574       moved into their own substructures.
575    
576       (The ultimate goal of these changes is to provide a light-weight
577       binfile loader/linker (at least for, e.g., stable libraries) that
578       does not require CM or the compiler to be present.)
579    
580    CM documentation has been updated to reflect the changes to library
581    organization.
582    
583    ----------------------------------------------------------------------
584  Name: Matthias Blume  Name: Matthias Blume
585  Date: 2001/07/10 17:30:00 EDT  Date: 2001/07/10 17:30:00 EDT
586  Tag: Release_110_34  Tag: Release_110_34
# Line 2589  Line 3157 
3157       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3158       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3159       Please, don't be too harsh on me for messing with this code a bit more       Please, don't be too harsh on me for messing with this code a bit more
3160       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3161       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3162       Dave's hands. :)       Dave's hands. :)
3163    

Legend:
Removed from v.876  
changed lines
  Added in v.976

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