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/NOTES/HISTORY
ViewVC logotype

Diff of /sml/trunk/NOTES/HISTORY

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 2387, Sat Apr 7 13:30:41 2007 UTC revision 2790, Tue Oct 2 23:26:59 2007 UTC
# Line 14  Line 14 
14  Description:  Description:
15    
16  ----------------------------------------------------------------------  ----------------------------------------------------------------------
17    Name: Mike Rainey (http://www.cs.uchicago.edu/~mrainey)
18    Date: 2007/10/2
19    Description:
20    
21    Added demo support for AMD64 for quick testing and fixed support for
22    64-bit label constants.
23    ----------------------------------------------------------------------
24    Name: Mike Rainey (http://www.cs.uchicago.edu/~mrainey)
25    Date: 2007/09/20
26    Description:
27    
28    Added support for the MLRISC COND instruction and the x86-64 CMOVCC
29    instruction.
30    ----------------------------------------------------------------------
31    Name: Mike Rainey (http://www.cs.uchicago.edu/~mrainey)
32    Date: 2007/09/17
33    Description:
34    
35    Fixed an instruction-selection bug when loading 64-bit labels.
36    
37    ----------------------------------------------------------------------
38    Name: George Kuan (http://www.cs.uchicago.edu/~gkuan)
39    Date: 2007/09/14
40    Description:
41    
42    Fixed Elaborator and Translate performance bugs
43    
44    ----------------------------------------------------------------------
45    Name: Mike Rainey (http://www.cs.uchicago.edu/~mrainey)
46    Date: 2007/09/12
47    Description:
48    
49    Fixed xorl memory argument bug.
50    
51    ----------------------------------------------------------------------
52    Name: George Kuan (http://www.cs.uchicago.edu/~gkuan)
53    Date: 2007/07/27
54    Description:
55    
56    Added f64sgn (for Real64.signBit) as a primop defined
57    in MLRiscGen. signBit(~0.0) not handled correctly.
58    
59    ----------------------------------------------------------------------
60    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
61    Date: 2007/06/21
62    Description:
63    
64    Fixed bug in {TextIO,BinIO}.StreamIO.endOfStream that would incorrectly
65    signal end of stream.
66    
67    ----------------------------------------------------------------------
68    Name: Matthias Blume (http://www.tti-c.org/blume/)
69    Date: 2007/06/12 11:22:00 CDT
70    Description:
71    
72    Eliminated config/allsources.  The information is now drawn directly
73    from config/actions.
74    
75    ----------------------------------------------------------------------
76    Name: Matthias Blume (http://www.tti-c.org/blume/)
77    Date: 2007/06/07 23:30:00 CDT
78    Tag: release-110.65
79    Description:
80    
81    Version 110.65. NEW BOOTFILES!
82    
83    ----------------------------------------------------------------------
84    Name: Matthias Blume (http://www.tti-c.org/blume/)
85    Date: 2007/06/06 21:45:00 CDT
86    Description:
87    
88    - Aaron: fixed a number of bugs in ml-ulex.
89    - Matthias:
90       - added CM control "cm.force-tools"; this is false by default;
91         when set to true, then tools like ml-yacc, ml-lex, ml-ulex, etc.
92         will be forced to run regardless of whether or not their targets
93         are up-to-date
94       - changed installer code so that config/install.sh will re-build
95         heap images for all tools even if those heap images already existed
96    
97    ----------------------------------------------------------------------
98    Name: Matthias Blume (http://www.tti-c.org/blume/)
99    Date: 2007/06/04 21:21:00 CDT
100    Description:
101    
102    After Aaron Turon's bug fix for ml-ulex (handling the ^ character
103    in legacy mode), re-ran the lexer generator on all lex input files
104    and committed the results.
105    
106    This should fix the problem with ckit and nlffi that was reported
107    by Vesa A. Norrman.
108    
109    ----------------------------------------------------------------------
110    Name: Matthias Blume (http://www.tti-c.org/blume/)
111    Date: 2007/05/31 11:40:00 CDT
112    Tag: release-110.64
113    Description:
114    
115    Version 110.64. NEW BOOTFILES!
116    
117    ----------------------------------------------------------------------
118    Name: Dave MacQueen (dbm@cs.uchicago.edu)
119    Date: 2007/05/31 14:55:00 CDT
120    Revision: 2650
121    Description:
122    
123    3rd merge of base from primop-branch-3 into the trunk.
124    Additional bug fixes included:
125    
126     * Timer.cpu_timer, etc. type printing corrected (by making
127         Timer have opaque sig constraint in basis/Implementation/timer.sml)
128     * Infinite loop in FLINT (tests/typing/tests/25.sml) (fix by
129         Stefan Monnier)
130    
131    ----------------------------------------------------------------------
132    Name: Matthias Blume (http://www.tti-c.org/blume/)
133    Date: 2007/05/31 11:40:00 CDT
134    Description:
135    
136    Fixed some bugs in new Div code in FLINT/trans/translate.sml.
137    
138    Also, changed the handling of "no core access":  When translate.sml
139    needs access to a core exception at a time when the core has not
140    been set up yet (this only happens when compiling system/smlnj/init/*),
141    then don't bother generating the corresponding tests.
142    
143    The old scheme was to generate a bogus value to be used in place of
144    the exception.  Unfortunately, that confuses the plambda type checker.
145    Moreover, it does not do any good, because at runtime we don't expect
146    such an exception to be ever raised.  (The code in system/smlnj/init/*
147    has to be written very carefully with this in mind!)
148    
149    ----------------------------------------------------------------------
150    Name: Mike Rainey (http://www.cs.uchicago.edu/~mrainey)
151    Date: 2007/05/29
152    Description:
153    
154    Added FSQRT instructions for the AMD64 code generator.
155    
156    ----------------------------------------------------------------------
157    Name: Matthias Blume (http://www.tti-c.org/blume/)
158    Date: 2007/05/29 19:00:00 CDT
159    Description:
160    
161    FLINT/trans/translate.sml now wraps all DIV/MOD/QUOT/REM operations
162    with an explicit test for zero division.  This should fixes several
163    regressions and makes it possible for downstream optimization phases
164    to treat these operations as "pure" when they are applied to unsigned
165    operands.
166    
167    ----------------------------------------------------------------------
168    Name: Mike Rainey (http://www.cs.uchicago.edu/~mrainey)
169    Date: 2007/05/29
170    Description:
171    
172    Added the new MLRISC code generator for the AMD64.  This version,
173    in contrast to the previous one, uses SSE registers and instructions
174    for all floating-point computations.
175    
176    ----------------------------------------------------------------------
177    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
178    Date: 2007/05/29
179    Description:
180    
181    A number of fixes related to the formatting of dates.  These include fixes
182    for bugs #1415 and #1416.  We also now correctly handle format characters that
183    lie outside the specified set.
184    
185    ----------------------------------------------------------------------
186    Name: Matthias Blume (http://www.tti-c.org/blume/)
187    Date: 2007/05/23 22:30:00 CDT
188    Description:
189    
190    CMB (and CM) now automatically defines the CM "preprocessor" symbol
191    NO_PLUGINS during "makeml -rebuild" or when CM operates in "slave" mode.
192    
193      (In addition, CMB_REBUILD_MODE is defined for makeml -rebuild,
194       CM_SLAVE_MODE is defined in attached slaves while running CM.make
195       or CM.rebuild, and CMB_SLAVE_MODE is defined in attached slaves
196       while running CMB.make.  The point is that the single symbol
197       NO_PLUGINS is definde in all three cases.)
198    
199    I changed the three locations within the sources that get compiled
200    during CMB.make where ml-yacc or ml-lex input is processed: When
201    NO_PLUGINS defined, then the use of the mlyacc and mllex tools
202    is bypassed.
203    
204    When bootstrapping new versions of the compiler, there can be
205    situations where the plugin tools for ml-yacc and ml-lex (or ml-ulex)
206    are not available or otherwise not operational.  In this case one can
207    manually define the NO_PLUGINS symbol prior to running CMB.make().  To
208    do so, the following command should be issued at the interactive
209    prompt:
210    
211       #set (CMB.symval "NO_PLUGINS") (SOME 1);
212    
213    A CMB.make with NO_PLUGINS defined relies on the existence of the
214    files that normally would be generated by ml-yacc and ml-lex.
215    (Copies of these files are in the repository.)
216    
217    ----------------------------------------------------------------------
218    Name: David MacQueen (dbm@cs.uchicago.edu)
219    Date: 2007/5/23 17:23:00 CDT
220    Revision: 2603
221    Description:
222    
223    2nd merge of base from primop-branch-3 into the trunk.
224    Additional bug fixes included:
225     Date.scan and Date.fromString fixed;
226     Overloading resolution fixed and some type printing problems corrected.
227    
228    ----------------------------------------------------------------------
229    Name: Matthias Blume (http://www.tti-c.org/blume/)
230    Date: 2007/05/23 09:50:00 CDT
231    Description:
232    
233    Changed the installation mechanism for CM tool plugins.  These are
234    just libaries and now get installed like ordinary libaries.
235    
236    There are now a number of new installation targets that give some
237    fine-grain control over what classes and suffixes are known, and what
238    they will map to.
239    See config/targets for details.
240    
241    The code that caused plugin installation as part of running a
242    tool's "build" script has been removed.  (The build script is for
243    building, not for installing.)
244    
245    ----------------------------------------------------------------------
246    Name: Matthias Blume (http://www.tti-c.org/blume/)
247    Date: 2007/05/22 11:50:00 CDT
248    Description:
249    
250    Added a boolean control named cm.tolerate-tool-failures (env. variable
251    name: CM_TOLERATE_TOOL_FAILURES).  The default is false and makes
252    CM fail if a shell tool reports a non-success exit status.  If
253    the control is set to true, then CM will press on after tool failures
254    in the event that the target files exist (even though they are
255    considered outdated).  Turning the control to true can be useful for
256    bootstrapping.
257    
258    ----------------------------------------------------------------------
259    Name: Aaron Turon (http://www.cs.uchicago.edu/~adrassi)
260    Date: 2007/05/19
261    Description:
262    
263    Merge of ml-lpt revisions for 110.64.
264    
265    The name of several ml-lpt-lib modules has changed:
266      Repair        => AntlrRepair
267      StreamPos     => AntlrStreamPos
268      ErrHandlerFn  => AntlrErrHandler
269      EBNF          => AntlrEBNF
270    
271    The ml-antlr specification format has changed: declarations such as
272    %tokens and nonterminal definitions can occur *multiple* times in the
273    same specification.  The semantics are such that each new declaration
274    extends the previous ones.  This does not apply to %start or %name,
275    of course.
276    
277    Importing a grammar via %import now includes *all* declarations in from
278    the imported grammar, except for %name, %entry, and %start.  Tokens
279    and nonterminals can be dropped using the new %dropping clause of the
280    %import directive; the separate %drop and %extend have been removed.
281    
282    We now allow optional type annotations on nonterminals, using the %nonterms
283    directive as in ml-yacc.
284    
285    The refcell construct is now implemented using SML's regular reference
286    cells, so the :== and !! notation has been deprecated.
287    
288    The ml-antlr tool now does much more checking of specifications, and
289    its error messages have been greatly improved.  Error repair for
290    generated parsers has been completely rewritten, and is now both
291    much faster and more accurate.
292    
293    ml-ulex is now more lenient with escape codes (non-SML-standard
294    escape codes are now interpreted literally, so e.g. \| denotes "|").
295    Also, character classes may now include a "-" character at the beginning
296    as is standard in most other regexp tools.
297    
298    All of these changes are documented in the user guide, which has
299    been updated and improved with this merge.
300    
301    ----------------------------------------------------------------------
302    Name: David MacQueen (http://www.cs.uchicago.edu/~jhr)
303    Date: 2007/05/03
304    Revision: r2492
305    Description:
306    
307    Merge of the primop3 branch (base) into the trunk to create 110.63.1.
308    Significant changes in FLINT and the front end, mostly having to do
309    with a reorganized system for handling primops.  Various bug fixes
310    and improvements in printing signatures.
311    
312    ----------------------------------------------------------------------
313    Name: Aaron Turon (http://www.cs.uchicago.edu/~adrassi)
314    Date: 2007/05/02
315    Description:
316    
317    Preliminary commit of large ml-lpt revisions (more to come for 110.64).
318    The ml-antlr specification format has changed: declarations such as
319    %tokens and nonterminal definitions can occur *multiple* times in the
320    same specification.  The semantics are such that each new declaration
321    extends the previous ones.  Grammar extension constructs have also
322    changed.  We now allow type annotations on nonterminals.  Finally,
323    the refcell construct is now implemented using SML's regular reference
324    cells, so the :== and !! notation has been deprecated.  All of these
325    changes will appear in the 110.64 user guide for ml-lpt.
326    
327    The ml-antlr tool now does much more checking of specifications, and
328    its error messages have been greatly improved.  There has also been
329    some work on the error repair process for generated parsers, but this
330    will be further improved in 110.64.
331    
332    ----------------------------------------------------------------------
333    Name: David MacQueen (http://www.cs.uchicago.edu/~jhr)
334    Date: 2007/05/03
335    Revision: r2492
336    Description:
337    
338    Merge of the primop3 branch (base) into the trunk to create 110.63.1.
339    Significant changes in FLINT and the front end, mostly having to do
340    with a reorganized system for handling primops.  Various bug fixes
341    and improvements in printing signatures.
342    
343    ----------------------------------------------------------------------
344    Name: Aaron Turon (http://www.cs.uchicago.edu/~adrassi)
345    Date: 2007/05/02
346    Description:
347    
348    Preliminary commit of large ml-lpt revisions (more to come for 110.64).
349    The ml-antlr specification format has changed: declarations such as
350    %tokens and nonterminal definitions can occur *multiple* times in the
351    same specification.  The semantics are such that each new declaration
352    extends the previous ones.  Grammar extension constructs have also
353    changed.  We now allow type annotations on nonterminals.  Finally,
354    the refcell construct is now implemented using SML's regular reference
355    cells, so the :== and !! notation has been deprecated.  All of these
356    changes will appear in the 110.64 user guide for ml-lpt.
357    
358    The ml-antlr tool now does much more checking of specifications, and
359    its error messages have been greatly improved.  There has also been
360    some work on the error repair process for generated parsers, but this
361    will be further improved in 110.64.
362    
363    ----------------------------------------------------------------------
364    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
365    Date: 2007/04/24
366    Description:
367    
368    More Basis fixes: The Char.fromString (etc.) functions did not handle
369    the "\uxxxx" escape sequence.  There is still an outstanding bug with
370    String.fromString the tail is a format escape.  I added a comment to
371    this effect in Basis/Implementation/string.sml.
372    Thanks to Andreas Rossberg.
373    
374    ----------------------------------------------------------------------
375    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
376    Date: 2007/04/24
377    Description:
378    
379    Added next function to Fifo and Queue modules in the SML/NJ Library.
380    
381    ----------------------------------------------------------------------
382    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
383    Date: 2007/04/23
384    Description:
385    
386    More Basis fixes: Time.fmt dropped the leading "~" for negative time
387    values that had no whole part.
388    
389    ----------------------------------------------------------------------
390    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
391    Date: 2007/04/23
392    Description:
393    
394    More Basis fixes: the OS.Path module did not include the InvalidArc
395    exception and did not do sufficient argument checking.
396    Thanks to Stephen Weeks and Adam Chilpala.
397    
398    ----------------------------------------------------------------------
399    Name: Mike Rainey (http://www.cs.uchicago.edu/~mrainey)
400    Date: 2007/04/16
401    Description:
402    
403    Fixed an unbound functor bug in the AMD64 CM file.
404    
405    ----------------------------------------------------------------------
406    Name: Mike Rainey (http://www.cs.uchicago.edu/~mrainey)
407    Date: 2007/04/12
408    Description:
409    
410    Substantially changed the signature and implementation of AMD64
411    SVID.  It now looks similar to the ia32 SVID, yet uses staged
412    allocation.
413    
414    ----------------------------------------------------------------------
415  Name: John Reppy (http://www.cs.uchicago.edu/~jhr)  Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
416  Date: 2007/04/07  Date: 2007/04/07
417  Description:  Description:

Legend:
Removed from v.2387  
changed lines
  Added in v.2790

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