-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathChangeLog
1921 lines (1282 loc) · 75.5 KB
/
ChangeLog
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.6 (v0_9_6) on Linux (lc2322-tx30) on 7/25/2017 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.6 (v0_9_6) on Linux (lc2322-tx30) on 7/25/2017 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.6 (v0_9_6) on Linux (lc2242-tx30) on 7/25/2017 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.6 (v0_9_6) on Linux (lc2243-tx30) on 7/25/2017 ==
2017-07-25 Brian Kahne <[email protected]>
* Miscellaneous automake/autoconf cleanup.
* Moved to automake 1.15 to make working on Ubuntu 16.04 easier.
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.5 (v0_9_5) on Linux (lc2322-tx30) on 7/14/2017 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.5 (v0_9_5) on Linux (lc2322-tx30) on 7/14/2017 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.5 (v0_9_5) on Linux (lc2243-tx30) on 7/14/2017 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.5 (v0_9_5) on Linux (lc2243-tx30) on 7/14/2017 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.5 (v0_9_5) on Linux (lc2289-tx30) on 7/14/2017 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.5 (v0_9_5) on Linux (lc2289-tx30) on 7/14/2017 ==
2017-07-07 Brian Kahne <bkahne@BrianUbuntu>
* Cleanup of warnings and autoconf changes to remove 'ar' issues.
2017-07-06 Brian Kahne <[email protected]>
* Porting Plasma to Ubuntu 16.04. Included gc7.2, as the prior
libgc segfaulted on 16.04. Note that, when porting libgc, rules
for creating a symlink have to be added to the end of
gc/Makefile.am, so that includes have the same location in the
installation directory as the sandbox. I'm also getting weird
behavior on RHEL6 when trying to run autoreconf, so I turned off
recursive behavior for now.
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.4 (v0_9_4) on Linux (lc2631-tx30) on 7/12/2016 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.4 (v0_9_4) on Linux (lc2075-tx30) on 7/12/2016 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.4 (v0_9_4) on Linux (lc2631-tx30) on 7/12/2016 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.4 (v0_9_4) on Linux (lc2075-tx30) on 7/12/2016 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.4 (v0_9_4) on Linux (lc2320-tx30) on 8/27/2015 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.4 (v0_9_4) on Linux (lc2320-tx30) on 8/27/2015 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.4 (v0_9_4) on Linux (lc2541-tx30) on 8/27/2015 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.4 (v0_9_4) on Linux (lc2294-tx30) on 8/27/2015 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.3 (v0_9_3) on Linux (lc2242-tx30) on 4/24/2015 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.3 (v0_9_3) on Linux (lc2243-tx30) on 4/24/2015 ==
2015-04-24 Brian Kahne <bkahne@ubuntu>
* opencxx/parser/Lex.cc: Fixed a stack overrun error caused by an
incorrect cast of an integer to a pointer- this can clobber the
stack on a 64b Linux platform.
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.2 (v0_9_2) on Linux (lc2249-tx30) on 8/20/2014 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.2 (v0_9_2) on Linux (lc2249-tx30) on 8/20/2014 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.2 (v0_9_2) on Linux (lc2249-tx30) on 8/20/2014 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.2 (v0_9_2) on Linux (lc2320-tx30) on 4/28/2014 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.2 (v0_9_2) on Linux (lc2294-tx30) on 4/28/2014 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.2 (v0_9_2) on Linux (lc2320-tx30) on 4/28/2014 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.2 (v0_9_2) on Linux (lc2294-tx30) on 4/28/2014 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.2 (v0_9_2) on Linux (lc2320-tx30) on 4/28/2014 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.2 (v0_9_2) on Linux (lc2294-tx30) on 4/28/2014 ==
2014-04-28 Brian Kahne <bkahne@lc2320-tx30>
* configure.ac: Flex is now optional: If we don't have it, we
just can't rebuild the tokenizer in the tests/cc directory.
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.1 (v0_9_1) on Linux (lc2320-tx30) on 4/28/2014 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.1 (v0_9_1) on Linux (lc2294-tx30) on 4/28/2014 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.1 (v0_9_1) on Linux (lc2320-tx30) on 4/25/2014 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.1 (v0_9_1) on Linux (lc2294-tx30) on 4/25/2014 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.0 (v0_9_0) on Linux (lc2467-tx30) on 2/28/2013 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.0 (v0_9_0) on Linux (lc2496-tx30) on 2/28/2013 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.0 (v0_9_0) on Linux (lc2015-tx30) on 2/5/2013 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.9.0 (v0_9_0) on Linux (lc2015-tx30) on 2/5/2013 ==
2013-02-05 Brian Kahne <[email protected]>
* Various build-system modifications so that we detect if we're
compiling with gcc > 4.4.x and disable compilation of the language
parser executable and runtime if that's the case. The only
regression that's then run is the simple parser test. This avoids
issues with C++11 constructs found in gcc 4.5.x and higher, which
confuse the parser.
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.8.50 (v0_8_50) on Linux (lc2031-tx30) on 12/7/2012 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.8.50 (v0_8_50) on Linux (lc2015-tx30) on 12/7/2012 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.8.50 (v0_8_50) on Linux (lc2031-tx30) on 12/6/2012 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.8.50 (v0_8_50) on Linux (lc2031-tx30) on 12/6/2012 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.8.50 (v0_8_50) on Linux (lc2015-tx30) on 12/6/2012 ==
== file:///proj/.ppc_43/vc_noec/git/plasma release 0.8.50 (v0_8_50) on Linux (lc2031-tx30) on 12/6/2012 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.50 (v0_8_50) on Linux (lc0205-tx30) on 10/28/2011 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.50 (v0_8_50) on Linux (lc0093) on 10/28/2011 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.50 (v0_8_50) on Linux (lc0205-tx30) on 10/28/2011 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.50 (v0_8_50) on Linux (lc0093) on 10/28/2011 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.50 (v0_8_50) on Linux (lc0205-tx30) on 10/28/2011 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.50 (v0_8_50) on Linux (lc0093) on 10/28/2011 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.50 (v0_8_50) on Linux (lc0205-tx30) on 10/28/2011 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.50 (v0_8_50) on Linux (lc0093) on 10/28/2011 ==
2011-10-28 Brian Kahne <[email protected]>
* Minor improvements to get this to work on gcc 4.4.x and autoconf
2.68.
2011-01-25 Brian Kahne <[email protected]>
* Modified plasma.in so that the autoconf substituted variables
are in single-quotes so that weird pathnames won't be interpolated
by Perl.
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.49 (v0_8_49) on Linux (lc0100) on 7/29/2010 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.49 (v0_8_49) on Linux (lc0101) on 7/29/2010 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.49 (v0_8_49) on Linux (lc0089) on 7/29/2010 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.48 (v0_8_48) on Linux (bkahne-ec1) on 7/13/2010 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.48 (v0_8_48) on Linux (bkahne-ec1) on 7/13/2010 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.48 (v0_8_48) on Linux (lc0089) on 7/13/2010 ==
2010-07-13 Brian Kahne <bkahne@lc0089>
* Updated plasma for gcc 4.4 compatibility.
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.47 (v0_8_47) on Linux (bkahne-ec1) on 11/9/2009 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.47 (v0_8_47) on Linux (lc0100) on 11/9/2009 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.47 (v0_8_47) on Linux (lc0101) on 11/9/2009 ==
2009-11-09 Brian Kahne <[email protected]>
* Modified try/catch block parsing so that the catch declaration
is treated as an actual declaration and translated so that the
declaration is added to the environment.
2009-08-14 Brian Kahne <[email protected]>
* Modified how the temporary file-name routine is supplied. This
used to be done via a weakly linked function, but that doesn't
work on Cygwin, so I've moded to using a functor with a function,
SetMakeTempFilename, that sets the local static variable.
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.46 (v0_8_46) on Linux (lc0100) on 8/12/2009 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.46 (v0_8_46) on Linux (lc0089) on 8/12/2009 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.46 (v0_8_46) on Linux (bkahne-ec1) on 8/12/2009 ==
2009-08-11 Brian Kahne <[email protected]>
* Upgraded to Boehm GC 6.8.
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.45 (v0_8_45) on Linux (bkahne-ec1) on 7/20/2009 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.45 (v0_8_45) on Linux (lc0100) on 7/20/2009 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.45 (v0_8_45) on Linux (lc0089) on 7/20/2009 ==
2009-07-20 Brian Kahne <[email protected]>
* Added --enable-frontend and --enable-runtime configure options
to allow the user to turn off various aspects of compiling Plasma.
2009-07-14 Brian Kahne <[email protected]>
* Upgraded to libtool 2.2.6a and autoconf 2.63.
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.43 (v0_8_43) on Linux (ld2059) on 1/30/2009 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.43 (v0_8_43) on Linux (ld2058) on 1/30/2009 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.43 (v0_8_43) on Linux (ld2058) on 1/30/2009 ==
2009-01-30 Brian Kahne <[email protected]>
* Got make distcheck to work.
* Added checking code to configure.ac to modify tar arguments to
avoid tarfile corruption due to large UIDs.
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.42 (v0_8_42) on Linux (ld2058) on 1/15/2009 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.42 (v0_8_42) on Linux (ld2058) on 1/15/2009 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.42 (v0_8_42) on Linux (ld2058) on 1/15/2009 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.42 (v0_8_42) on Linux (ld2061) on 12/4/2008 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.42 (v0_8_42) on Linux (ld2059) on 12/4/2008 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.42 (v0_8_42) on Linux (ld2059) on 12/4/2008 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.42 (v0_8_42) on Linux (ld2061) on 12/4/2008 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.42 (v0_8_42) on Linux (ld2059) on 12/4/2008 ==
2008-12-04 Brian Kahne <[email protected]>
* Added patches from Todd Dukes for fixing const* and regression
problems, found when using gcc 4.2.x.
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.41 (v0_8_41) on Linux (lb32001) on 9/4/2008 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.41 (v0_8_41) on Linux (ld2058) on 9/4/2008 ==
2008-07-03 Brian Kahne <[email protected]>
* Added the ability to shut-off the GC. This is done by defining
GC_DISABLED in gc/include/{gc_cpp.h,gc_allocator.h}. Right now,
you have to hand-edit these files; we could modify this so that it
can be done by a configure variable, but I haven't implemented
that yet.
2008-06-05 Brian Kahne <[email protected]>
* Updated the libtool macro so that it works properly with the gcc
4.1.1 found in /pkg.
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.40 (v0_8_40) on Linux (lb32001) on 5/12/2008 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.40 (v0_8_40) on Linux (ld2058) on 2/26/2008 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.40 (v0_8_40) on Linux (lb32001) on 2/26/2008 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.39 (v0_8_39) on Linux (ld2058) on 2/26/2008 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.39 (v0_8_39) on Linux (lb32001) on 2/26/2008 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.39 (v0_8_39) on Linux (lb32001) on 2/26/2008 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.39 (v0_8_39) on Linux (ld2058) on 2/26/2008 ==
2008-02-05 Brian Kahne <[email protected]>
* Added an '@' operator at the same precedence as '->' in order to
support new features in uADL. Added parsing/in11.C to test this.
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.38 (v0_8_38) on Linux (lb32001) on 12/4/2007 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.38 (v0_8_38) on Linux (lb0172) on 12/4/2007 ==
2007-12-04 Brian Kahne <[email protected]>
* Fixed a bug with the boolean version of Reify. It was
incorrectly matching as a boolean anything that was just a portion
of a boolean constant, e.g. 't' would match as 'true'. Modified
parsing/in5.C to test this.
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.37 (v0_8_37) on Linux (lb32001) on 7/9/2007 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.37 (v0_8_37) on Linux (lb0170) on 6/28/2007 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.37 (v0_8_37) on Linux (lb32001) on 6/28/2007 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.37 (v0_8_37) on Linux (lb0170) on 6/28/2007 ==
2007-06-28 Brian Kahne <[email protected]>
* Fixed some distribution problems: We weren't distributing
QuickThread files needed for building on x86_64.
* Fatal encoding errors now throw a runtime_error with a message,
rather than using the error log, so that the entire error can be
squashed if not relevant to the user.
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.36 (v0_8_36) on Linux (lb32001) on 5/22/2007 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.36 (v0_8_36) on Linux (lb0169) on 5/22/2007 ==
2007-05-22 Brian Kahne <[email protected]>
* Various tweaks to get this to compile with --disable-shared. If
shared objects are disabled, then we add pthread as a system
library and we also define --allow-multiple-definition so that
some of our symbols can replace symbols in other libraries.
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.35 (v0_8_35) on Linux (lb32001) on 5/18/2007 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.35 (v0_8_35) on Linux (lb0169) on 5/18/2007 ==
2007-05-18 Brian Kahne <[email protected]>
* Removed the building of libgccpp.la because it caused link
errors when linked statically against libstdc++ (it overrides new
and delete, which causes multiply-defined-symbol errors if done
statically). We don't use this library, so that shouldn't be a
problem.
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.34 (v0_8_34) on Linux (lb0169) on 4/4/2007 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.34 (v0_8_34) on Linux (lb32001) on 4/4/2007 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.34 (v0_8_34) on Linux (lb0169) on 4/4/2007 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.34 (v0_8_34) on Linux (lb32001) on 4/4/2007 ==
2007-04-04 Brian Kahne <[email protected]>
* Fixed a bug in Program::LineNumber. It was incorrectly looking
at string contents, so a string that contained something that
looked like a line number would be counted as a line number.
Added parsing/in10.C to test this.
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.33 (v0_8_33) on Linux (lb32001) on 4/3/2007 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.33 (v0_8_33) on Linux (lb0169) on 4/3/2007 ==
2007-04-03 Brian Kahne <[email protected]>
* Make MakeTempName non-static so that a user of the library can
override it to allow for alternate intermediate file names.
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.32 (v0_8_32) on Linux (lb32001) on 3/23/2007 ==
2007-03-23 Brian Kahne <[email protected]>
* I no longer build meas in the qt directory b/c for some reason,
it doesn't build on i686 platforms. This is just a test program,
so it's not required by anything else.
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.31 (v0_8_31) on Linux (lb32001) on 3/23/2007 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.31 (v0_8_31) on Linux (lb0169) on 3/23/2007 ==
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.31 (v0_8_31) on Linux (lb32001) on 3/23/2007 ==
2007-03-23 Brian Kahne <[email protected]>
* Added support for nested template functions. As with nested
functions, this can be disabled by undefining NESTED_FUNCS_OKAY.
Added parsing/in9.C to test this.
== file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.30 (v0_8_30) on Linux (lb32001) on 3/7/2007 ==
file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.30 (v0_8_30) on SunOS (sb0183) on 1/22/2007
file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.30 (v0_8_30) on Linux (lb32001) on 1/19/2007
file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.30 (v0_8_30) on Linux (lb32001) on 1/19/2007
2007-01-19 Brian Kahne <[email protected]>
* For Solaris, we'll compile, but we only test the parser.
file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.29 (v0_8_29) on Linux (lb0169) on 12/8/2006
file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.29 (v0_8_29) on Linux (lb32001) on 12/8/2006
file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.29 (v0_8_29) on Linux (lb0169) on 12/8/2006
file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.29 (v0_8_29) on Linux (lb32001) on 12/8/2006
file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.29 (v0_8_29) on Linux (lb32001) on 12/8/2006
file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.29 (v0_8_29) on Linux (lb32001) on 11/30/2006
file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.29 (v0_8_29) on Linux (lb32001) on 11/30/2006
file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.29 (v0_8_29) on Linux (lb32001) on 11/30/2006
file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.29 (v0_8_29) on Linux (lb32001) on 11/30/2006
file:///proj/.ppc_43/vc_noec/svn/plasma release 0.8.29 (v0_8_29) on Linux (lb0169) on 11/30/2006
2006-11-30 Brian Kahne <[email protected]>
* Tested against gcc 4.1.1.
* We now try to discover FlexLexer.h's location using several
tests- we use "which" first, then "w.which".
file:///proj/ppctools/svn/plasma release 0.8.28 (v0_8_28) on Linux (lb0169) on 11/9/2006
file:///proj/ppctools/svn/plasma/ release 0.8.28 (v0_8_28) on Linux (lb0169) on 11/9/2006
file:///proj/ppctools/svn/plasma release 0.8.28 (v0_8_28) on Linux (lb32001) on 11/9/2006
file:///proj/ppctools/svn/plasma release 0.8.28 (v0_8_28) on Linux (lb32001) on 11/9/2006
file:///proj/ppctools/svn/plasma release 0.8.28 (v0_8_28) on Linux (lb32001) on 11/9/2006
file:///proj/ppctools/svn/plasma release 0.8.28 (v0_8_28) on Linux (lb32001) on 11/9/2006
2006-11-08 Brian Kahne <[email protected]>
* Plasma has been ported to the AMD64 architecture. For
QuickThreads support, I used a setjmp/longjmp piece of code from
another project which extended QuickThreads. No other major
porting problems existed, except for a few minor little issues,
such as needing to modify some casts in the ecu tests directory
and to ensure that the assembler used -m32 when assembling the
output of minicc in tests/cc.
2006-10-30 Brian Kahne <[email protected]>
* Plasma should now build correctly if configured and built from a
different directory than where the source is located.
======== plasma release 0.8.27 (v0_8_27) on Linux (lb32001) on 10/30/2006 ========
======== plasma release 0.8.27 (v0_8_27) on Linux (lb32001) on 10/30/2006 ========
======== plasma release 0.8.27 (v0_8_27) on Linux (lb32001) on 10/30/2006 ========
======== plasma release 0.8.27 (v0_8_27) on Linux (lb32001) on 10/30/2006 ========
======== plasma release 0.8.27 (v0_8_27) on Linux (lb32001) on 10/30/2006 ========
======== plasma release 0.8.27 (v0_8_27) on Linux (lb32001) on 10/30/2006 ========
======== plasma release 0.8.27 (v0_8_27) on Linux (lb32001) on 10/27/2006 ========
2006-10-27 Brian Kahne <[email protected]>
* opencxx/Environment.cc (Append): Removed an explicit delete
which was causing a free() error b/c the memory was not allocated
by the system but rather by GC.
2006-08-14 Brian Kahne <[email protected]>
* opencxx/Environment.cc (DumpAll): Added a dump-everything
function for debugging purposes.
===== plasma release 0.8.26 (v0_8_26) on Linux (cde-tx32-ldt189) on 8/4/2006 =====
===== plasma release 0.8.26 (v0_8_26) on Linux (cde-tx32-ldt189) on 8/4/2006 =====
===== plasma release 0.8.26 (v0_8_26) on Linux (cde-tx32-ldt189) on 8/4/2006 =====
===== plasma release 0.8.26 (v0_8_26) on Linux (cde-tx32-ldt189) on 8/4/2006 =====
====== plasma release 0.8.26 (v0_8_26) on Linux (cde-tx32-ldt189) ======
====== plasma release 0.8.26 (v0_8_26) on Linux (cde-tx32-ldt189) ======
====== plasma release 0.8.26 (v0_8_26) on Linux (cde-tx32-ldt189) ======
====== plasma release 0.8.26 (v0_8_26) on Linux (cde-tx32-ldt189) ======
2006-03-21 Brian Kahne <[email protected]>
* Moved the website to Compass so that everyone can access it.
====== plasma release 0.8.26 (v0_8_26) on Linux (cde-tx32-ldt189) ======
====== plasma release 0.8.26 (v0_8_26) on Linux (cde-tx32-ldt189) ======
2006-03-17 Brian Kahne <[email protected]>
* Fixed minor issues in order to port to gcc 4.1. The problems
were that the new compiler eliminated a lot of timing loops,
making some regressions break when compiled with optimization.
====== plasma release 0.8.25 (v0_8_25) on Linux (cde-tx32-ldt189) ======
2006-03-17 Brian Kahne <[email protected]>
* Various minor changes to port Plasma to gcc 4.1.
* The garbage collector was upgraded to version 6.7.
====== plasma release 0.8.24 (v0_8_24) on Linux (cde-tx32-ldt189) ======
====== plasma release 0.8.24 (v0_8_24) on Linux (cde-tx32-ldt189) ======
====== plasma release 0.8.24 (v0_8_24) on Linux (cde-tx32-ldt189) ======
2006-03-07 Brian Kahne <[email protected]>
* Fixed bugs in TypeInfo:
* Added ability to get encoding string.
* Fixed the nth-template-argument accessor function.
* Fixed nested functions so that they work (I think). This
feature is now turned on by default.
2006-02-08 Brian Kahne <[email protected]>
* Removed trailing whitespace from shbang lines to avoid an
issue with running them under Cygwin.
====== plasma release 0.8.23 (v0_8_23) on Linux (cde-tx32-ldt189) ======
2006-01-05 Brian Kahne <[email protected]>
* Fixed a bug with parsing long strings: We didn't handle the case
where we received a null character, indicating that we had a badly
formed string (had run to the end of the buffer w/o finding the
terminator). This meant that we ended up in an infinite loop.
Fixed this and added tests/parsing/in7.C to verify this.
====== plasma release 0.8.22 (v0_8_22) on Linux (cde-tx32-ldt189) ======
====== plasma release 0.8.22 (v0_8_22) on Linux (cde-tx32-ldt189) ======
2005-11-09 Brian Kahne <[email protected]>
* Removed references to Somerset and FIL from documentation.
* Modified the documentation and configure/Makefile stuff so that
we print an URL location that's specified using a command-line
option (--web-url).
====== plasma release 0.8.21 (v0_8_21) on Linux (cde-tx32-ldt189) ======
2005-11-02 Brian Kahne <[email protected]>
* Added copy-right notices to all C++/Plasma files. Created a
script called "add-copyright" that tries to do this in an
intelligent fashion.
====== plasma release 0.8.20 (v0_8_20) on Linux (cde-tx32-ldt189) ======
2005-05-26 Brian Kahne <[email protected]>
* If configure finds that the path to the compiler is a link, it
tries to follow it.
* Modified plasma.in so that the contents of an autoconf variable
CONFOPTS will be added to cpp and gcc options when plasma is run.
* Modified configure.ac so that if you configure on an IA32
machine, -m32 is added to CONFOPTS. The idea is that we want
32-bit code generated even on an Opteron if Plasma is built on an
IA32 machine.
====== plasma release 0.8.19 (v0_8_19) on Linux (cde-tx32-ldt189) ======
2005-05-18 Brian Kahne <[email protected]>
* Modified Encoding so that integer template parameters are
encoded using an 'N' prefix followed by the number as a string.
2005-04-27 Brian Kahne <[email protected]>
* configure and Makefile.in files are now checked in so that the
user does not have to run automake/autoconf to build from CVS.
====== plasma release 0.8.18 (v0_8_18) on Linux (cde-tx32-ldt189) ======
====== plasma release 0.8.18 (v0_8_18) on Linux (cde-tx32-ldt189) ======
====== plasma release 0.8.18 (v0_8_18) on Linux (cde-tx32-ldt189) ======
====== plasma release 0.8.18 (v0_8_18) on Linux (cde-tx32-ldt189) ======
====== plasma release 0.8.18 (v0_8_18) on Linux (cde-tx32-ldt189) ======
2005-04-20 Brian Kahne <[email protected]>
* Added support for a "let" statement. It has the form of:
let (decl [, decl ...]) { body }
where decl is a variable declaration with or without a type. If
no type is present, its type is inferenced using gcc's typeof operator.
====== plasma release 0.8.17 (v0_8_17) on Linux (cde-tx32-ldt189) ======
2005-03-16 Brian Kahne <[email protected]>
* Added a Ptree::Reify function to handle booleans and added to
tests/parsing to test this.
* Modified the SourceLocation function to recursively find the
first leaf- this is a better implementation than what I had
before.
====== plasma release 0.8.16 (v0_8_16) on Linux (cde-tx32-ldt189) ======
2005-03-10 Brian Kahne <[email protected]>
* Additional minor changes, such as better source location support
for error messages: Given a ptree, if it's not a leaf, we attempt
to find a leaf item.
2005-03-09 Brian Kahne <[email protected]>
* Enhanced the parser to accept several new things:
Multi-line quotes are now supported using Python's doc-string
format: """ string """.
Nested code blocks are allowed:
foo = { ... };
Bare user-statements are allowed in expressions:
foo = func(int i) { ... };
Support for nested functions was added, but is currently commented
out in Parser.cc.
====== plasma release 0.8.15 (v0_8_15) on Linux (cde-tx32-ldt189) ======
2005-02-22 Brian Kahne <[email protected]>
* Upgraded to OpenC++ 2.8 (current CVS snapshot). I merged in my
changes and got everything working. The interface is pretty much
the same, with a few minor tweeks. In the future, I think that
we'll want to move towards using OpenC++ simply as a filter and
have plasma (the perl program) call the preprocessor and
compiler. However, that can wait for now.
====== plasma release 0.8.14 (v0_8_14) on Linux (cde-tx32-ldt189) ======
2005-02-15 Brian Kahne <[email protected]>
* Included OpenC++ within the Plasma distribution to make it
easier to distribute the code.
====== plasma release 0.8.13 (v0_8_13) on Linux (cde-tx32-ldt189) ======
2005-01-28 Brian Kahne <[email protected]>
* Added an API to the channels in order to allow for some
reflection/connectivity deduction. It works like this: Each
channel needs to define the following:
bool multiple_producers_allowed() const: Returns true if multiple
producers are allowed for this channel.
bool multiple_consumers_allowed() const: Returns true if multiple
producers are allowed for this channel.
void *get_source_channel() const: Return a pointer to the channel
that feeds this one. This pointer should be initialized
to the 'this' pointer of the channel at first. Make sure
that if the void pointer is stored in a base class, that
the base class occurs first in the list of parent classes.
void *set_source_channel() const: Set the source channel.
The idea with the source channel is that broadcaster objects set
the source to the source channel- this allows a program to deduce
connectivity information. Channels which don't have a source,
such as timeout and result, don't need a source channel pointer.
====== plasma release 0.8.12 (v0_8_12) on Linux (cde-tx32-ldt189) ======
2005-01-27 Brian Kahne <[email protected]>
* Modified the broadcaster classes so that you don't specify how
many you need- you simply call get_sink() and that appends to the
list. I changed the data structure from a vector to a list so
that you don't have to worry about relocations.
* Modified the clocked channels so that you can have one with a
zero clock period. This becomes equivalent to a non-clocked
channel. The reasoning behind this is that you can write a design
where everything is a clocked channel, then modify its behavior if
you want to, rather than having to change types.
====== plasma release 0.8.11 (v0_8_11) on Linux (cde-tx32-ldt189) ======
2005-01-25 Brian Kahne <[email protected]>
* Fixed a minor issue with alt code generation- we don't create a
top loop label if we have a default block because we don't loop:
If no channel is ready, we execute the default block and continue.
2005-01-20 Brian Kahne <[email protected]>
* Fixed bug w/clearing 0-sized clocked channels: We used to not
re-queue a waiting writer, so it would stay blocked. Fixed this
with a tail clause in clear().
* Removed the handle from threads. I still have a valid flag for
use by busysleep/busywake, but that's it. Removed HandleType from
all channels.
* Re-wrote how alt works: It no longer uses a handle. Instead, a
thread goes asleep and when it wakes, it jumps back to the
checking logic. This was done to fix a subtle bug with prialt:
If two channels become ready at the same time, the priorities
should be obeyed even if the lower priority thread is what wakes
the channel. Using the handle, it wasn't possible to ensure
this. The handle is still there, but will be removed next.
* Added chan25 to test the prialt issue mentioned above.
======= plasma release 0.8.9 (v0_8_9) on Linux (cde-tx32-ldt189) =======
2005-01-17 Brian Kahne <[email protected]>
* Changed the clear() functions- they simply do a get() if ready()
is true. This was because the other behavior of clear() could
lead to non-deterministic behavior and runaway conditions.
* Added a zero-size capability to clocked channels and made this
the default size. If the size is 0, a write will block until a
read removes the data. This lets you do a fully interlocked
pipeline.
* Fixed a bug with priorities- terminate now adds all waiters,
then queries for the next ready thread. This fixes a problem
whereby we used to get the next ready thread before adding the
waiters which could result in us executing a lower-priority thread
before others of higher priority.
======= plasma release 0.8.8 (v0_8_8) on Linux (cde-tx32-ldt189) =======
2005-01-04 Brian Kahne <[email protected]>
* Fixed bug with pLowestPriority(): It was returning 0, rather
than the 1 less than the total number of priorities (which is the
lowest priority from the user's point of view). Modified pri2 to
test this.
======= plasma release 0.8.7 (v0_8_7) on Linux (cde-tx32-ldt189) =======
2004-12-22 Brian Kahne <[email protected]>
* Fixed bug: We could get a crash if we tried accessing the
current priority from a global constructor because the current
processor wasn't set up yet. I now return the lowest priority in
such a situation.
2004-12-22 Brian Kahne <[email protected]>
* Created a simple broadcaster template that takes input from a
channel and writes it to its output channels, clearing them
first. Added a convenience class for clocked channels that
creates a broadcaster with an input clocked channel and an output
that's non-clocked. The channels are cleared so that the
broadcast can be ignored without backing anything up, e.g. if one
consumer doesn't care, the data is wiped out on the next read.
======= plasma release 0.8.6 (v0_8_6) on Linux (cde-tx32-ldt189) =======
2004-12-21 Brian Kahne <[email protected]>
* Added clear methods to the channels. This will clear out the
data and any pending writers.
2004-12-14 Brian Kahne <[email protected]>
* Fixed bug with single-consumer clocked channels: We weren't
clearing _readt so on a second pass to an alt where we had to
sleep, we'd assert.
* Added clock9 to test clocked channels interacting with
non-clocked channels in an alt.
======= plasma release 0.8.5 (v0_8_5) on Linux (cde-tx32-ldt189) =======
======= plasma release 0.8.5 (v0_8_5) on Linux (cde-tx32-ldt189) =======
======= plasma release 0.8.5 (v0_8_5) on Linux (cde-tx32-ldt189) =======
2004-12-03 Brian Kahne <[email protected]>
* Added support for multiple consumers to clocked channels. You
can create a multi-consumer clocked channel by specifying
MultiConsumerClockChannel as the second template argument to a
ClockChan type. Added clock7 and clock8 to test this.
* Timeout::ready() now returns true after a timeout as issued.
Get "consumes" this readiness, setting the ready state to false.
You can now call get() and it will sleep until the timeout issues,
so you can use a timeout outside of an alt, though it's pretty
much equivalent to just calling pDelay.
* Modified the code generated for an alt to handle multi-consumers
properly. Once a thread has been awakened, each switch element,
before the get, calls ready on the channel. If it's not ready
(somebody else consumed the data), then we jump back to the
beginning and start over.
2004-11-30 Brian Kahne <[email protected]>
* spawn now supports void functions: They're treated as returning
an int that's always 0. So, the result object should be of type
int. Since that's now the default type for Result, you can just
omit it and have an empty template declaration:
Result<> r = spawn(foo());
You can then wait or kill r if necessary.
* Refactored the channels to take advantage of common code. Thus,
channels which allow for multiple producers inherit from
MultiProducerChannel. Channels which have a single piece of data
inherit from SingleDataChannel. There is also
SingleConsumerChannel and MultiConsumerChannel base classes, which
are now template arguments for Channel, QueueChan, and BusyChan.
This allows the channel to support multiple consumers, if desired.
======= plasma release 0.8.4 (v0_8_4) on Linux (cde-tx32-ldt189) =======
======= plasma release 0.8.4 (v0_8_4) on Linux (cde-tx32-ldt189) =======
2004-11-16 Brian Kahne <[email protected]>
======= plasma release 0.8.4 (v0_8_4) on Linux (cde-tx32-ldt189) =======
* Fixed bug: Empty statements, i.e. just a semicolon, in a par
block would cause the parser to crash because it would try to
extract stuff from an empty object.
======= plasma release 0.8.4 (v0_8_4) on Linux (cde-tx32-ldt189) =======
2004-11-12 Brian Kahne <[email protected]>
======= plasma release 0.8.3 (v0_8_3) on Linux (cde-tx32-ldt189) =======
* Fixed next_phi() equation b/c clocked channels with a skew could
have multiple reads at the same time if the second read occurred
on a cycle boundary.
2004-11-11 Brian Kahne <[email protected]>
* Clocked channels now take three parameters: Period, skew, and
size. Skew defaults to 0. This lets a user have a channel that's
delayed with respect to other clocked channels, but with the same
period, e.g. a half-cycle delay. Regression "clock5" tests this.
======= plasma release 0.8.2 (v0_8_2) on Linux (cde-tx32-ldt189) =======
2004-11-04 Brian Kahne <[email protected]>
* All channels now define value_type as the type of the channel.
======= plasma release 0.8.1 (v0_8_1) on Linux (cde-tx32-ldt189) =======
2004-11-02 Brian Kahne <[email protected]>
* Fixed a bug with clocked channels: If you wrote to a clocked
channel, then delayed by the clock period, you'd get a segfault
because we would have started a delayed waker, then done our own
reset (because we're now on a clock boundary) without canceling
that waker thread. Fixed this by canceling the thread if it
exists. Added "clock4" to test this.
2004-10-13 Brian Kahne <[email protected]>
* Added plasma.m4, a macro that can be used by plasma users to
detect plasma.
======= plasma release 0.8.0 (v0_8_0) on Linux (cde-tx32-ldt189) =======
2004-10-08 Brian Kahne <[email protected]>
* Port to PowePC: Minor changes to fix regressions, plus I fixed
a bug with the parser- I had some STL data structures that didn't
use a traceable allocator.
2004-10-07 Brian Kahne <[email protected]>
* Added QuickThreads PowerPC port to the qt directory. This
doesn't support varargs, but that should be okay, since we never
use it. Note: This has only been tested a little bit, and not
with these files here- that's the next step!
2004-10-05 Brian Kahne <[email protected]>
* Moved the energy API out of Interface and Proc and into a
separate class. This is a thin wrapper around the ProcValue
class, which is a hash keyed by processors. The idea is that the
user can now record any value- energy just being a common
specialization.
* Created the Quantity class to handle the case where the user
wants to model a limited resource that multiple threads/processors
are trying to get. This uses a new channel, BusyChan, which is
like a simple channel, except that it causes the process to go
busy if a read is attempted and no data is available.
To implement BusyChan, a new form of busy was added (pBusySleep)
which keeps a processor busy until it is awoken (pBusyWake).
The Quantity class works like this:
1. The user declares the class and specifies a timeslice. This
corresponds to the amount of time it takes to do 1 of that
quantity, e.g. a bus object might take 10 time units to do a read.
2. A thread calls Quantity::request(int amount), where amount
specifies the number of items it needs. The class then does a
round-robin on all requesting threads, taking up a timeslice and
decrementing the count of a thread. When a thread's request has
decremented to zero, the requesting thread's processor is un-busied.
======= plasma release 0.7.0 (v0_7_0) on Linux (cde-tx32-ldt189) =======
2004-09-23 Brian Kahne <[email protected]>
* Added some support for power modelling. The basic idea is that
the user can annotate a model with calls to
void pEnergy(energy_t);
This adds to the current processor's value representing consumed
energy. You can then call either:
energy_t pGetEnergy(Processor);
energy_t pReadEnergy(Processor);
The first clears the energy value while the second does not. To
obtain power, just create a thread which regularly samples the
energy of the relevant processors and divides by the sample
period.
2004-09-21 Brian Kahne <[email protected]>
* Got skribe to produce a pdf of the main document (not the source
code listings). Modified configure so that it checks to see if
pdflatex is available and good (a very recent version of the
hyperref package is required).
======= plasma release 0.6.0 (v0_6_0) on Linux (cde-tx32-ldt189) =======
2004-09-20 Brian Kahne <[email protected]>
* Modified configure.ac to specify the version using AC_INIT,
which is the standard method.
2004-09-20 Brian Kahne <[email protected]>
* Added prialt and priafor constructs. These are guaranteed to
read the ports in the order specified by the user, whereas the
ordering is indeterminate for alt/afor.
* Changed the handling of alt blocks so that the channels are read
in reverse order, just to distinguish between alt and prialt. The
afor is handled the same as priafor, but the idea is that a more
sophisticated compiler, which recognized that an afor is a
constant loop, might unroll the loop and do some reordering.
2004-09-17 Brian Kahne <[email protected]>
* Fixed some scheduling bugs related to pooled processors:
1. We used to throw out empty processors in
Cluster::update_proc(). However, this had the effect of getting
rid of pooled processors- they would never then be used. Changed
this so that we no longer have a loop in update_proc. Instead,
get_new_proc() calls ProcQ::get_non_empty() which scans the queue
for a proc w/threads. If it doesn't find any then it advances
time. I like to avoid loops, but this should almost never advance
beyond the first iteration- it'll only occur w/pooled processors
and only when few threads are present.
2. Modified Cluster::busy() so that it removes the scheduling
thread, then finds the next processor (or updates time), then adds
in the scheduler. This fixed the problem of switching to a