aboutsummaryrefslogtreecommitdiff
path: root/contrib/bc/NEWS.md
blob: c66e7b164c8a95e54e915c9f1a280cf6e4053e0c (plain) (blame)
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
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
# News

## 4.0.1

This is a production release that only adds one thing: flushing output when it
is printed with a print statement.

## 4.0.0

This is a production release with many fixes, a new command-line option, and a
big surprise:

* A bug was fixed in `dc`'s `P` command where the item on the stack was *not*
  popped.
* Various bugs in the manuals have been fixed.
* A known bug was fixed where history did not interact well with prompts printed
  by user code without newlines.
* A new command-line option, `-R` and `--no-read-prompt` was added to disable
  just the prompt when using `read()` (`bc`) or `?` (`dc`).
* And finally, **official support for Windows was added**.

The last item is why this is a major version bump.

Currently, only one set of build options (extra math and prompt enabled, history
and NLS/locale support disabled, both calculators enabled) is supported on
Windows. However, both debug and release builds are supported.

In addition, Windows builds are supported for the the library (`bcl`).

For more details about how to build on Windows, see the [README][5] or the
[build manual][13].

## 3.3.4

This is a production release that fixes a small bug.

The bug was that output was not flushed before a `read()` call, so prompts
without a newline on the end were not flushed before the `read()` call.

This is such a tiny bug that users only need to upgrade if they are affected.

## 3.3.3

This is a production release with one tweak and fixes for manuals.

The tweak is that `length(0)` returns `1` instead of `0`. In `3.3.1`, I changed
it so `length(0.x)`, where `x` could be any number of digits, returned the
`scale`, but `length(0)` still returned `0` because I believe that `0` has `0`
significant digits.

After request of FreeBSD and considering the arguments of a mathematician,
compatibility with other `bc`'s, and the expectations of users, I decided to
make the change.

The fixes for manuals fixed a bug where `--` was rendered as `-`.

## 3.3.2

This is a production release that fixes a divide-by-zero bug in `root()` in the
[extended math library][16]. All previous versions with `root()` have the bug.

## 3.3.1

This is a production release that fixes a bug.

The bug was in the reporting of number length when the value was 0.

## 3.3.0

This is a production release that changes one behavior and fixes documentation
bugs.

The changed behavior is the treatment of `-e` and `-f` when given through
`BC_ENV_ARGS` or `DC_ENV_ARGS`. Now `bc` and `dc` do not exit when those options
(or their equivalents) are given through those environment variables. However,
`bc` and `dc` still exit when they or their equivalents are given on the
command-line.

## 3.2.7

This is a production release that removes a small non-portable shell operation
in `configure.sh`. This problem was only noticed on OpenBSD, not FreeBSD or
Linux.

Non-OpenBSD users do ***NOT*** need to upgrade, although NetBSD users may also
need to upgrade.

## 3.2.6

This is a production release that fixes the build on FreeBSD.

There was a syntax error in `configure.sh` that the Linux shell did not catch,
and FreeBSD depends on the existence of `tests/all.sh`.

All users that already upgraded to `3.2.5` should update to this release, with
my apologies for the poor release of `3.2.5`. Other users should skip `3.2.5` in
favor of this version.

## 3.2.5

This is a production release that fixes several bugs and adds a couple small
things.

The two most important bugs were bugs that causes `dc` to access memory
out-of-bounds (crash in debug builds). This was found by upgrading to `afl++`
from `afl`. Both were caused by a failure to distinguish between the same two
cases.

Another bug was the failure to put all of the licenses in the `LICENSE.md` file.

Third, some warnings by `scan-build` were found and eliminated. This needed one
big change: `bc` and `dc` now bail out as fast as possible on fatal errors
instead of unwinding the stack.

Fourth, the pseudo-random number now attempts to seed itself with `/dev/random`
if `/dev/urandom` fails.

Finally, this release has a few quality-of-life changes to the build system. The
usage should not change at all; the only thing that changed was making sure the
`Makefile.in` was written to rebuild properly when headers changed and to not
rebuild when not necessary.

## 3.2.4

This is a production release that fixes a warning on `gcc` 6 or older, which
does not have an attribute that is used.

Users do ***NOT*** need to upgrade if they don't use `gcc` 6 or older.

## 3.2.3

This is a production release that fixes a bug in `gen/strgen.sh`. I recently
changed `gen/strgen.c`, but I did not change `gen/strgen.sh`.

Users that do not use `gen/strgen.sh` do not need to upgrade.

## 3.2.2

This is a production release that fixes a portability bug in `configure.sh`. The
bug was using the GNU `find` extension `-wholename`.

## 3.2.1

This is a production release that has one fix for `bcl(3)`. It is technically
not a bug fix since the behavior is undefined, but the `BclNumber`s that
`bcl_divmod()` returns will be set to `BCL_ERROR_INVALID_NUM` if there is an
error. Previously, they were not set.

## 3.2.0

This is a production release that has one bug fix and a major addition.

The bug fix was a missing `auto` variable in the bessel `j()` function in the
math library.

The major addition is a way to build a version of `bc`'s math code as a library.
This is done with the `-a` option to `configure.sh`. The API for the library can
be read in `./manuals/bcl.3.md` or `man bcl` once the library is installed with
`make install`.

This library was requested by developers before I even finished version 1.0, but
I could not figure out how to do it until now.

If the library has API breaking changes, the major version of `bc` will be
incremented.

## 3.1.6

This is a production release that fixes a new warning from Clang 12 for FreeBSD
and also removes some possible undefined behavior found by UBSan that compilers
did not seem to take advantage of.

Users do ***NOT*** need to upgrade, if they do not want to.

## 3.1.5

This is a production release that fixes the Chinese locales (which caused `bc`
to crash) and a crash caused by `bc` executing code when it should not have been
able to.

***ALL USERS SHOULD UPGRADE.***

## 3.1.4

This is a production release that fixes one bug, changes two behaviors, and
removes one environment variable.

The bug is like the one in the last release except it applies if files are being
executed. I also made the fix more general.

The behavior that was changed is that `bc` now exits when given `-e`, `-f`,
`--expression` or `--file`. However, if the last one of those is `-f-` (using
`stdin` as the file), `bc` does not exit. If `-f-` exists and is not the last of
the `-e` and `-f` options (and equivalents), `bc` gives a fatal error and exits.

Next, I removed the `BC_EXPR_EXIT` and `DC_EXPR_EXIT` environment variables
since their use is not needed with the behavior change.

Finally, I made it so `bc` does not print the header, though the `-q` and
`--quiet` options were kept for compatibility with GNU `bc`.

## 3.1.3

This is a production release that fixes one minor bug: if `bc` was invoked like
the following, it would error:

```
echo "if (1 < 3) 1" | bc
```

Unless users run into this bug, they do not need to upgrade, but it is suggested
that they do.

## 3.1.2

This is a production release that adds a way to install *all* locales. Users do
***NOT*** need to upgrade.

For package maintainers wishing to make use of the change, just pass `-l` to
`configure.sh`.

## 3.1.1

This is a production release that adds two Spanish locales. Users do ***NOT***
need to upgrade, unless they want those locales.

## 3.1.0

This is a production release that adjusts one behavior, fixes eight bugs, and
improves manpages for FreeBSD. Because this release fixes bugs, **users and
package maintainers should update to this version as soon as possible**.

The behavior that was adjusted was how code from the `-e` and `-f` arguments
(and equivalents) were executed. They used to be executed as one big chunk, but
in this release, they are now executed line-by-line.

The first bug fix in how output to `stdout` was handled in `SIGINT`. If a
`SIGINT` came in, the `stdout` buffer was not correctly flushed. In fact, a
clean-up function was not getting called. This release fixes that bug.

The second bug is in how `dc` handled input from `stdin`. This affected `bc` as
well since it was a mishandling of the `stdin` buffer.

The third fixed bug was that `bc` and `dc` could `abort()` (in debug mode) when
receiving a `SIGTERM`. This one was a race condition with pushing and popping
items onto and out of vectors.

The fourth bug fixed was that `bc` could leave extra items on the stack and
thus, not properly clean up some memory. (The memory would still get
`free()`'ed, but it would not be `free()`'ed when it could have been.)

The next two bugs were bugs in `bc`'s parser that caused crashes when executing
the resulting code.

The last two bugs were crashes in `dc` that resulted from mishandling of
strings.

The manpage improvement was done by switching from [ronn][20] to [Pandoc][21] to
generate manpages. Pandoc generates much cleaner manpages and doesn't leave
blank lines where they shouldn't be.

## 3.0.3

This is a production release that adds one new feature: specific manpages.

Before this release, `bc` and `dc` only used one manpage each that referred to
various build options. This release changes it so there is one manpage set per
relevant build type. Each manual only has information about its particular
build, and `configure.sh` selects the correct set for install.

## 3.0.2

This is a production release that adds `utf8` locale symlinks and removes an
unused `auto` variable from the `ceil()` function in the [extended math
library][16].

Users do ***NOT*** need to update unless they want the locales.

## 3.0.1

This is a production release with two small changes. Users do ***NOT*** need to
upgrade to this release; however, if they haven't upgraded to `3.0.0` yet, it
may be worthwhile to upgrade to this release.

The first change is fixing a compiler warning on FreeBSD with strict warnings
on.

The second change is to make the new implementation of `ceil()` in `lib2.bc`
much more efficient.

## 3.0.0

*Notes for package maintainers:*

*First, the `2.7.0` release series saw a change in the option parsing. This made
me change one error message and add a few others. The error message that was
changed removed one format specifier. This means that `printf()` will seqfault
on old locale files. Unfortunately, `bc` cannot use any locale files except the
global ones that are already installed, so it will use the previous ones while
running tests during install. **If `bc` segfaults while running arg tests when
updating, it is because the global locale files have not been replaced. Make
sure to either prevent the test suite from running on update or remove the old
locale files before updating.** (Removing the locale files can be done with
`make uninstall` or by running the `locale_uninstall.sh` script.) Once this is
done, `bc` should install without problems.*

*Second, **the option to build without signal support has been removed**. See
below for the reasons why.*

This is a production release with some small bug fixes, a few improvements,
three major bug fixes, and a complete redesign of `bc`'s error and signal
handling. **Users and package maintainers should update to this version as soon
as possible.**

The first major bug fix was in how `bc` executed files. Previously, a whole file
was parsed before it was executed, but if a function is defined *after* code,
especially if the function definition was actually a redefinition, and the code
before the definition referred to the previous function, this `bc` would replace
the function before executing any code. The fix was to make sure that all code
that existed before a function definition was executed.

The second major bug fix was in `bc`'s `lib2.bc`. The `ceil()` function had a
bug where a `0` in the decimal place after the truncation position, caused it to
output the wrong numbers if there was any non-zero digit after.

The third major bug is that when passing parameters to functions, if an
expression included an array (not an array element) as a parameter, it was
accepted, when it should have been rejected. It is now correctly rejected.

Beyond that, this `bc` got several improvements that both sped it up, improved
the handling of signals, and improved the error handling.

First, the requirements for `bc` were pushed back to POSIX 2008. `bc` uses one
function, `strdup()`, which is not in POSIX 2001, and it is in the X/Open System
Interfaces group 2001. It is, however, in POSIX 2008, and since POSIX 2008 is
old enough to be supported anywhere that I care, that should be the requirement.

Second, the BcVm global variable was put into `bss`. This actually slightly
reduces the size of the executable from a massive code shrink, and it will stop
`bc` from allocating a large set of memory when `bc` starts.

Third, the default Karatsuba length was updated from 64 to 32 after making the
optimization changes below, since 32 is going to be better than 64 after the
changes.

Fourth, Spanish translations were added.

Fifth, the interpreter received a speedup to make performance on non-math-heavy
scripts more competitive with GNU `bc`. While improvements did, in fact, get it
much closer (see the [benchmarks][19]), it isn't quite there.

There were several things done to speed up the interpreter:

First, several small inefficiencies were removed. These inefficiencies included
calling the function `bc_vec_pop(v)` twice instead of calling
`bc_vec_npop(v, 2)`. They also included an extra function call for checking the
size of the stack and checking the size of the stack more than once on several
operations.

Second, since the current `bc` function is the one that stores constants and
strings, the program caches pointers to the current function's vectors of
constants and strings to prevent needing to grab the current function in order
to grab a constant or a string.

Third, `bc` tries to reuse `BcNum`'s (the internal representation of
arbitary-precision numbers). If a `BcNum` has the default capacity of
`BC_NUM_DEF_SIZE` (32 on 64-bit and 16 on 32-bit) when it is freed, it is added
to a list of available `BcNum`'s. And then, when a `BcNum` is allocated with a
capacity of `BC_NUM_DEF_SIZE` and any `BcNum`'s exist on the list of reusable
ones, one of those ones is grabbed instead.

In order to support these changes, the `BC_NUM_DEF_SIZE` was changed. It used to
be 16 bytes on all systems, but it was changed to more closely align with the
minimum allocation size on Linux, which is either 32 bytes (64-bit musl), 24
bytes (64-bit glibc), 16 bytes (32-bit musl), or 12 bytes (32-bit glibc). Since
these are the minimum allocation sizes, these are the sizes that would be
allocated anyway, making it worth it to just use the whole space, so the value
of `BC_NUM_DEF_SIZE` on 64-bit systems was changed to 32 bytes.

On top of that, at least on 64-bit, `BC_NUM_DEF_SIZE` supports numbers with
either 72 integer digits or 45 integer digits and 27 fractional digits. This
should be more than enough for most cases since `bc`'s default `scale` values
are 0 or 20, meaning that, by default, it has at most 20 fractional digits. And
45 integer digits are *a lot*; it's enough to calculate the amount of mass in
the Milky Way galaxy in kilograms. Also, 72 digits is enough to calculate the
diameter of the universe in Planck lengths.

(For 32-bit, these numbers are either 32 integer digits or 12 integer digits and
20 fractional digits. These are also quite big, and going much bigger on a
32-bit system seems a little pointless since 12 digits in just under a trillion
and 20 fractional digits is still enough for about any use since `10^-20` light
years is just under a millimeter.)

All of this together means that for ordinary uses, and even uses in scientific
work, the default number size will be all that is needed, which means that
nearly all, if not all, numbers will be reused, relieving pressure on the system
allocator.

I did several experiments to find the changes that had the most impact,
especially with regard to reusing `BcNum`'s. One was putting `BcNum`'s into
buckets according to their capacity in powers of 2 up to 512. That performed
worse than `bc` did in `2.7.2`. Another was putting any `BcNum` on the reuse
list that had a capacity of `BC_NUM_DEF_SIZE * 2` and reusing them for `BcNum`'s
that requested `BC_NUM_DEF_SIZE`. This did reduce the amount of time spent, but
it also spent a lot of time in the system allocator for an unknown reason. (When
using `strace`, a bunch more `brk` calls showed up.) Just reusing `BcNum`'s that
had exactly `BC_NUM_DEF_SIZE` capacity spent the smallest amount of time in both
user and system time. This makes sense, especially with the changes to make
`BC_NUM_DEF_SIZE` bigger on 64-bit systems, since the vast majority of numbers
will only ever use numbers with a size less than or equal to `BC_NUM_DEF_SIZE`.

Last of all, `bc`'s signal handling underwent a complete redesign. (This is the
reason that this version is `3.0.0` and not `2.8.0`.) The change was to move
from a polling approach to signal handling to an interrupt-based approach.

Previously, every single loop condition had a check for signals. I suspect that
this could be expensive when in tight loops.

Now, the signal handler just uses `longjmp()` (actually `siglongjmp()`) to start
an unwinding of the stack until it is stopped or the stack is unwound to
`main()`, which just returns. If `bc` is currently executing code that cannot be
safely interrupted (according to POSIX), then signals are "locked." The signal
handler checks if the lock is taken, and if it is, it just sets the status to
indicate that a signal arrived. Later, when the signal lock is released, the
status is checked to see if a signal came in. If so, the stack unwinding starts.

This design eliminates polling in favor of maintaining a stack of `jmp_buf`'s.
This has its own performance implications, but it gives better interaction. And
the cost of pushing and popping a `jmp_buf` in a function is paid at most twice.
Most functions do not pay that price, and most of the rest only pay it once.
(There are only some 3 functions in `bc` that push and pop a `jmp_buf` twice.)

As a side effect of this change, I had to eliminate the use of `stdio.h` in `bc`
because `stdio` does not play nice with signals and `longjmp()`. I implemented
custom I/O buffer code that takes a fraction of the size. This means that static
builds will be smaller, but non-static builds will be bigger, though they will
have less linking time.

This change is also good because my history implementation was already bypassing
`stdio` for good reasons, and unifying the architecture was a win.

Another reason for this change is that my `bc` should *always* behave correctly
in the presence of signals like `SIGINT`, `SIGTERM`, and `SIGQUIT`. With the
addition of my own I/O buffering, I needed to also make sure that the buffers
were correctly flushed even when such signals happened.

For this reason, I **removed the option to build without signal support**.

As a nice side effect of this change, the error handling code could be changed
to take advantage of the stack unwinding that signals used. This means that
signals and error handling use the same code paths, which means that the stack
unwinding is well-tested. (Errors are tested heavily in the test suite.)

It also means that functions do not need to return a status code that
***every*** caller needs to check. This eliminated over 100 branches that simply
checked return codes and then passed that return code up the stack if necessary.
The code bloat savings from this is at least 1700 bytes on `x86_64`, *before*
taking into account the extra code from removing `stdio.h`.

## 2.7.2

This is a production release with one major bug fix.

The `length()` built-in function can take either a number or an array. If it
takes an array, it returns the length of the array. Arrays can be passed by
reference. The bug is that the `length()` function would not properly
dereference arrays that were references. This is a bug that affects all users.

**ALL USERS SHOULD UPDATE `bc`**.

## 2.7.1

This is a production release with fixes for new locales and fixes for compiler
warnings on FreeBSD.

## 2.7.0

This is a production release with a bug fix for Linux, new translations, and new
features.

Bug fixes:

* Option parsing in `BC_ENV_ARGS` was broken on Linux in 2.6.1 because `glibc`'s
  `getopt_long()` is broken. To get around that, and to support long options on
  every platform, an adapted version of [`optparse`][17] was added. Now, `bc`
  does not even use `getopt()`.
* Parsing `BC_ENV_ARGS` with quotes now works. It isn't the smartest, but it
  does the job if there are spaces in file names.

The following new languages are supported:

* Dutch
* Polish
* Russian
* Japanes
* Simplified Chinese

All of these translations were generated using [DeepL][18], so improvements are
welcome.

There is only one new feature: **`bc` now has a built-in pseudo-random number
generator** (PRNG).

The PRNG is seeded, making it useful for applications where
`/dev/urandom` does not work because output needs to be reproducible. However,
it also uses `/dev/urandom` to seed itself by default, so it will start with a
good seed by default.

It also outputs 32 bits on 32-bit platforms and 64 bits on 64-bit platforms, far
better than the 15 bits of C's `rand()` and `bash`'s `$RANDOM`.

In addition, the PRNG can take a bound, and when it gets a bound, it
automatically adjusts to remove bias. It can also generate numbers of arbitrary
size. (As of the time of release, the largest pseudo-random number generated by
this `bc` was generated with a bound of `2^(2^20)`.)

***IMPORTANT: read the [`bc` manual][9] and the [`dc` manual][10] to find out
exactly what guarantees the PRNG provides. The underlying implementation is not
guaranteed to stay the same, but the guarantees that it provides are guaranteed
to stay the same regardless of the implementation.***

On top of that, four functions were added to `bc`'s [extended math library][16]
to make using the PRNG easier:

* `frand(p)`: Generates a number between `[0,1)` to `p` decimal places.
* `ifrand(i, p)`: Generates an integer with bound `i` and adds it to `frand(p)`.
* `srand(x)`: Randomizes the sign of `x`. In other words, it flips the sign of
  `x` with probability `0.5`.
* `brand()`: Returns a random boolean value (either `0` or `1`).

## 2.6.1

This is a production release with a bug fix for FreeBSD.

The bug was that when `bc` was built without long options, it would give a fatal
error on every run. This was caused by a mishandling of `optind`.

## 2.6.0

This release is a production release ***with no bugfixes***. If you do not want
to upgrade, you don't have to.

No source code changed; the only thing that changed was `lib2.bc`.

This release adds one function to the [extended math library][16]: `p(x, y)`,
which calculates `x` to the power of `y`, whether or not `y` is an integer. (The
`^` operator can only accept integer powers.)

This release also includes a couple of small tweaks to the [extended math
library][16], mostly to fix returning numbers with too high of `scale`.

## 2.5.3

This release is a production release which addresses inconsistencies in the
Portuguese locales. No `bc` code was changed.

The issues were that the ISO files used different naming, and also that the
files that should have been symlinks were not. I did not catch that because
GitHub rendered them the exact same way.

## 2.5.2

This release is a production release.

No code was changed, but the build system was changed to allow `CFLAGS` to be
given to `CC`, like this:

```
CC="gcc -O3 -march=native" ./configure.sh
```

If this happens, the flags are automatically put into `CFLAGS`, and the compiler
is set appropriately. In the example above this means that `CC` will be "gcc"
and `CFLAGS` will be "-O3 -march=native".

This behavior was added to conform to GNU autotools practices.

## 2.5.1

This is a production release which addresses portability concerns discovered
in the `bc` build system. No `bc` code was changed.

* Support for Solaris SPARC and AIX were added.
* Minor documentations edits were performed.
* An option for `configure.sh` was added to disable long options if
  `getopt_long()` is missing.

## 2.5.0

This is a production release with new translations. No code changed.

The translations were contributed by [bugcrazy][15], and they are for
Portuguese, both Portugal and Brazil locales.

## 2.4.0

This is a production release primarily aimed at improving `dc`.

* A couple of copy and paste errors in the [`dc` manual][10] were fixed.
* `dc` startup was optimized by making sure it didn't have to set up `bc`-only
  things.
* The `bc` `&&` and `||` operators were made available to `dc` through the `M`
  and `m` commands, respectively.
* `dc` macros were changed to be tail call-optimized.

The last item, tail call optimization, means that if the last thing in a macro
is a call to another macro, then the old macro is popped before executing the
new macro. This change was made to stop `dc` from consuming more and more memory
as macros are executed in a loop.

The `q` and `Q` commands still respect the "hidden" macros by way of recording
how many macros were removed by tail call optimization.

## 2.3.2

This is a production release meant to fix warnings in the Gentoo `ebuild` by
making it possible to disable binary stripping. Other users do *not* need to
upgrade.

## 2.3.1

This is a production release. It fixes a bug that caused `-1000000000 < -1` to
return `0`. This only happened with negative numbers and only if the value on
the left was more negative by a certain amount. That said, this bug *is* a bad
bug, and needs to be fixed.

**ALL USERS SHOULD UPDATE `bc`**.

## 2.3.0

This is a production release with changes to the build system.

## 2.2.0

This release is a production release. It only has new features and performance
improvements.

1.	The performance of `sqrt(x)` was improved.
2.	The new function `root(x, n)` was added to the extended math library to
	calculate `n`th roots.
3.	The new function `cbrt(x)` was added to the extended math library to
	calculate cube roots.

## 2.1.3

This is a non-critical release; it just changes the build system, and in
non-breaking ways:

1.	Linked locale files were changed to link to their sources with a relative
	link.
2.	A bug in `configure.sh` that caused long option parsing to fail under `bash`
	was fixed.

## 2.1.2

This release is not a critical release.

1.	A few codes were added to history.
2.	Multiplication was optimized a bit more.
3.	Addition and subtraction were both optimized a bit more.

## 2.1.1

This release contains a fix for the test suite made for Linux from Scratch: now
the test suite prints `pass` when a test is passed.

Other than that, there is no change in this release, so distros and other users
do not need to upgrade.

## 2.1.0

This release is a production release.

The following bugs were fixed:

1.	A `dc` bug that caused stack mishandling was fixed.
2.	A warning on OpenBSD was fixed.
3.	Bugs in `ctrl+arrow` operations in history were fixed.
4.	The ability to paste multiple lines in history was added.
5.	A `bc` bug, mishandling of array arguments to functions, was fixed.
6.	A crash caused by freeing the wrong pointer was fixed.
7.	A `dc` bug where strings, in a rare case, were mishandled in parsing was
	fixed.

In addition, the following changes were made:

1.	Division was slightly optimized.
2.	An option was added to the build to disable printing of prompts.
3.	The special case of empty arguments is now handled. This is to prevent
	errors in scripts that end up passing empty arguments.
4.	A harmless bug was fixed. This bug was that, with the pop instructions
	(mostly) removed (see below), `bc` would leave extra values on its stack for
	`void` functions and in a few other cases. These extra items would not
	affect anything put on the stack and would not cause any sort of crash or
	even buggy behavior, but they would cause `bc` to take more memory than it
	needed.

On top of the above changes, the following optimizations were added:

1.	The need for pop instructions in `bc` was removed.
2.	Extra tests on every iteration of the interpreter loop were removed.
3.	Updating function and code pointers on every iteration of the interpreter
	loop was changed to only updating them when necessary.
4.	Extra assignments to pointers were removed.

Altogether, these changes sped up the interpreter by around 2x.

***NOTE***: This is the last release with new features because this `bc` is now
considered complete. From now on, only bug fixes and new translations will be
added to this `bc`.

## 2.0.3

This is a production, bug-fix release.

Two bugs were fixed in this release:

1.	A rare and subtle signal handling bug was fixed.
2.	A misbehavior on `0` to a negative power was fixed.

The last bug bears some mentioning.

When I originally wrote power, I did not thoroughly check its error cases;
instead, I had it check if the first number was `0` and then if so, just return
`0`. However, `0` to a negative power means that `1` will be divided by `0`,
which is an error.

I caught this, but only after I stopped being cocky. You see, sometime later, I
had noticed that GNU `bc` returned an error, correctly, but I thought it was
wrong simply because that's not what my `bc` did. I saw it again later and had a
double take. I checked for real, finally, and found out that my `bc` was wrong
all along.

That was bad on me. But the bug was easy to fix, so it is fixed now.

There are two other things in this release:

1.	Subtraction was optimized by [Stefan Eßer][14].
2.	Division was also optimized, also by Stefan Eßer.

## 2.0.2

This release contains a fix for a possible overflow in the signal handling. I
would be surprised if any users ran into it because it would only happen after 2
billion (`2^31-1`) `SIGINT`'s, but I saw it and had to fix it.

## 2.0.1

This release contains very few things that will apply to any users.

1.	A slight bug in `dc`'s interactive mode was fixed.
2.	A bug in the test suite that was only triggered on NetBSD was fixed.
3.	**The `-P`/`--no-prompt` option** was added for users that do not want a
	prompt.
4.	A `make check` target was added as an alias for `make test`.
5.	`dc` got its own read prompt: `?> `.

## 2.0.0

This release is a production release.

This release is also a little different from previous releases. From here on
out, I do not plan on adding any more features to this `bc`; I believe that it
is complete. However, there may be bug fix releases in the future, if I or any
others manage to find bugs.

This release has only a few new features:

1.	`atan2(y, x)` was added to the extended math library as both `a2(y, x)` and
	`atan2(y, x)`.
2.	Locales were fixed.
3.	A **POSIX shell-compatible script was added as an alternative to compiling
	`gen/strgen.c`** on a host machine. More details about making the choice
	between the two can be found by running `./configure.sh --help` or reading
	the [build manual][13].
4.	Multiplication was optimized by using **diagonal multiplication**, rather
	than straight brute force.
5.	The `locale_install.sh` script was fixed.
6.	`dc` was given the ability to **use the environment variable
	`DC_ENV_ARGS`**.
7.	`dc` was also given the ability to **use the `-i` or `--interactive`**
	options.
8.	Printing the prompt was fixed so that it did not print when it shouldn't.
9.	Signal handling was fixed.
10.	**Handling of `SIGTERM` and `SIGQUIT`** was fixed.
11.	The **built-in functions `maxibase()`, `maxobase()`, and `maxscale()`** (the
	commands `T`, `U`, `V` in `dc`, respectively) were added to allow scripts to
	query for the max allowable values of those globals.
12.	Some incompatibilities with POSIX were fixed.

In addition, this release is `2.0.0` for a big reason: the internal format for
numbers changed. They used to be a `char` array. Now, they are an array of
larger integers, packing more decimal digits into each integer. This has
delivered ***HUGE*** performance improvements, especially for multiplication,
division, and power.

This `bc` should now be the fastest `bc` available, but I may be wrong.

## 1.2.8

This release contains a fix for a harmless bug (it is harmless in that it still
works, but it just copies extra data) in the [`locale_install.sh`][12] script.

## 1.2.7

This version contains fixes for the build on Arch Linux.

## 1.2.6

This release removes the use of `local` in shell scripts because it's not POSIX
shell-compatible, and also updates a man page that should have been updated a
long time ago but was missed.

## 1.2.5

This release contains some missing locale `*.msg` files.

## 1.2.4

This release contains a few bug fixes and new French translations.

## 1.2.3

This release contains a fix for a bug: use of uninitialized data. Such data was
only used when outputting an error message, but I am striving for perfection. As
Michelangelo said, "Trifles make perfection, and perfection is no trifle."

## 1.2.2

This release contains fixes for OpenBSD.

## 1.2.1

This release contains bug fixes for some rare bugs.

## 1.2.0

This is a production release.

There have been several changes since `1.1.0`:

1.	The build system had some changes.
2.	Locale support has been added. (Patches welcome for translations.)
3.	**The ability to turn `ibase`, `obase`, and `scale` into stacks** was added
	with the `-g` command-line option. (See the [`bc` manual][9] for more
	details.)
4.	Support for compiling on Mac OSX out of the box was added.
5.	The extended math library got `t(x)`, `ceil(x)`, and some aliases.
6.	The extended math library also got `r2d(x)` (for converting from radians to
	degrees) and `d2r(x)` (for converting from degrees to radians). This is to
	allow using degrees with the standard library.
7.	Both calculators now accept numbers in **scientific notation**. See the
	[`bc` manual][9] and the [`dc` manual][10] for details.
8.	Both calculators can **output in either scientific or engineering
	notation**. See the [`bc` manual][9] and the [`dc` manual][10] for details.
9.	Some inefficiencies were removed.
10.	Some bugs were fixed.
11.	Some bugs in the extended library were fixed.
12.	Some defects from [Coverity Scan][11] were fixed.

## 1.1.4

This release contains a fix to the build system that allows it to build on older
versions of `glibc`.

## 1.1.3

This release contains a fix for a bug in the test suite where `bc` tests and
`dc` tests could not be run in parallel.

## 1.1.2

This release has a fix for a history bug; the down arrow did not work.

## 1.1.1

This release fixes a bug in the `1.1.0` build system. The source is exactly the
same.

The bug that was fixed was a failure to install if no `EXECSUFFIX` was used.

## 1.1.0

This is a production release. However, many new features were added since `1.0`.

1.	**The build system has been changed** to use a custom, POSIX
	shell-compatible configure script ([`configure.sh`][6]) to generate a POSIX
	make-compatible `Makefile`, which means that `bc` and `dc` now build out of
	the box on any POSIX-compatible system.
2.	Out-of-memory and output errors now cause the `bc` to report the error,
	clean up, and die, rather than just reporting and trying to continue.
3.	**Strings and constants are now garbage collected** when possible.
4.	Signal handling and checking has been made more simple and more thorough.
5.	`BcGlobals` was refactored into `BcVm` and `BcVm` was made global. Some
	procedure names were changed to reflect its difference to everything else.
6.	Addition got a speed improvement.
7.	Some common code for addition and multiplication was refactored into its own
	procedure.
8.	A bug was removed where `dc` could have been selected, but the internal
	`#define` that returned `true` for a query about `dc` would not have
	returned `true`.
9.	Useless calls to `bc_num_zero()` were removed.
10.	**History support was added.** The history support is based off of a
	[UTF-8 aware fork][7] of [`linenoise`][8], which has been customized with
	`bc`'s own data structures and signal handling.
11.	Generating C source from the math library now removes tabs from the library,
	shrinking the size of the executable.
12.	The math library was shrunk.
13.	Error handling and reporting was improved.
14.	Reallocations were reduced by giving access to the request size for each
	operation.
15.	**`abs()` (`b` command for `dc`) was added as a builtin.**
16.	Both calculators were tested on FreeBSD.
17.	Many obscure parse bugs were fixed.
18.	Markdown and man page manuals were added, and the man pages are installed by
	`make install`.
19.	Executable size was reduced, though the added features probably made the
	executable end up bigger.
20.	**GNU-style array references were added as a supported feature.**
21.	Allocations were reduced.
22.	**New operators were added**: `$` (`$` for `dc`), `@` (`@` for `dc`), `@=`,
	`<<` (`H` for `dc`), `<<=`, `>>` (`h` for `dc`), and `>>=`. See the
	[`bc` manual][9] and the [`dc` manual][10] for more details.
23.	**An extended math library was added.** This library contains code that
	makes it so I can replace my desktop calculator with this `bc`. See the
	[`bc` manual][3] for more details.
24.	Support for all capital letters as numbers was added.
25.	**Support for GNU-style void functions was added.**
26.	A bug fix for improper handling of function parameters was added.
27.	Precedence for the or (`||`) operator was changed to match GNU `bc`.
28.	`dc` was given an explicit negation command.
29.	`dc` was changed to be able to handle strings in arrays.

## 1.1 Release Candidate 3

This release is the eighth release candidate for 1.1, though it is the third
release candidate meant as a general release candidate. The new code has not
been tested as thoroughly as it should for release.

## 1.1 Release Candidate 2

This release is the seventh release candidate for 1.1, though it is the second
release candidate meant as a general release candidate. The new code has not
been tested as thoroughly as it should for release.

## 1.1 FreeBSD Beta 5

This release is the sixth release candidate for 1.1, though it is the fifth
release candidate meant specifically to test if `bc` works on FreeBSD. The new
code has not been tested as thoroughly as it should for release.

## 1.1 FreeBSD Beta 4

This release is the fifth release candidate for 1.1, though it is the fourth
release candidate meant specifically to test if `bc` works on FreeBSD. The new
code has not been tested as thoroughly as it should for release.

## 1.1 FreeBSD Beta 3

This release is the fourth release candidate for 1.1, though it is the third
release candidate meant specifically to test if `bc` works on FreeBSD. The new
code has not been tested as thoroughly as it should for release.

## 1.1 FreeBSD Beta 2

This release is the third release candidate for 1.1, though it is the second
release candidate meant specifically to test if `bc` works on FreeBSD. The new
code has not been tested as thoroughly as it should for release.

## 1.1 FreeBSD Beta 1

This release is the second release candidate for 1.1, though it is meant
specifically to test if `bc` works on FreeBSD. The new code has not been tested as
thoroughly as it should for release.

## 1.1 Release Candidate 1

This is the first release candidate for 1.1. The new code has not been tested as
thoroughly as it should for release.

## 1.0

This is the first non-beta release. `bc` is ready for production use.

As such, a lot has changed since 0.5.

1.	`dc` has been added. It has been tested even more thoroughly than `bc` was
	for `0.5`. It does not have the `!` command, and for security reasons, it
	never will, so it is complete.
2.	`bc` has been more thoroughly tested. An entire section of the test suite
	(for both programs) has been added to test for errors.
3.	A prompt (`>>> `) has been added for interactive mode, making it easier to
	see inputs and outputs.
4.	Interrupt handling has been improved, including elimination of race
	conditions (as much as possible).
5.	MinGW and [Windows Subsystem for Linux][1] support has been added (see
	[xstatic][2] for binaries).
6.	Memory leaks and errors have been eliminated (as far as ASan and Valgrind
	can tell).
7.	Crashes have been eliminated (as far as [afl][3] can tell).
8.	Karatsuba multiplication was added (and thoroughly) tested, speeding up
	multiplication and power by orders of magnitude.
9.	Performance was further enhanced by using a "divmod" function to reduce
	redundant divisions and by removing superfluous `memset()` calls.
10.	To switch between Karatsuba and `O(n^2)` multiplication, the config variable
	`BC_NUM_KARATSUBA_LEN` was added. It is set to a sane default, but the
	optimal number can be found with [`karatsuba.py`][4] (requires Python 3)
	and then configured through `make`.
11.	The random math test generator script was changed to Python 3 and improved.
	`bc` and `dc` have together been run through 30+ million random tests.
12.	All known math bugs have been fixed, including out of control memory
	allocations in `sine` and `cosine` (that was actually a parse bug), certain
	cases of infinite loop on square root, and slight inaccuracies (as much as
	possible; see the [README][5]) in transcendental functions.
13.	Parsing has been fixed as much as possible.
14.	Test coverage was improved to 94.8%. The only paths not covered are ones
	that happen when `malloc()` or `realloc()` fails.
15.	An extension to get the length of an array was added.
16.	The boolean not (`!`) had its precedence change to match negation.
17.	Data input was hardened.
18.	`bc` was made fully compliant with POSIX when the `-s` flag is used or
	`POSIXLY_CORRECT` is defined.
19.	Error handling was improved.
20.	`bc` now checks that files it is given are not directories.

## 1.0 Release Candidate 7

This is the seventh release candidate for 1.0. It fixes a few bugs in 1.0
Release Candidate 6.

## 1.0 Release Candidate 6

This is the sixth release candidate for 1.0. It fixes a few bugs in 1.0 Release
Candidate 5.

## 1.0 Release Candidate 5

This is the fifth release candidate for 1.0. It fixes a few bugs in 1.0 Release
Candidate 4.

## 1.0 Release Candidate 4

This is the fourth release candidate for 1.0. It fixes a few bugs in 1.0 Release
Candidate 3.

## 1.0 Release Candidate 3

This is the third release candidate for 1.0. It fixes a few bugs in 1.0 Release
Candidate 2.

## 1.0 Release Candidate 2

This is the second release candidate for 1.0. It fixes a few bugs in 1.0 Release
Candidate 1.

## 1.0 Release Candidate 1

This is the first Release Candidate for 1.0. `bc` is complete, with `dc`, but it
is not tested.

## 0.5

This beta release completes more features, but it is still not complete nor
tested as thoroughly as necessary.

## 0.4.1

This beta release fixes a few bugs in 0.4.

## 0.4

This is a beta release. It does not have the complete set of features, and it is
not thoroughly tested.

[1]: https://docs.microsoft.com/en-us/windows/wsl/install-win10
[2]: https://pkg.musl.cc/bc/
[3]: http://lcamtuf.coredump.cx/afl/
[4]: ./karatsuba.py
[5]: ./README.md
[6]: ./configure.sh
[7]: https://github.com/rain-1/linenoise-mob
[8]: https://github.com/antirez/linenoise
[9]: ./manuals/bc/A.1.md
[10]: ./manuals/dc/A.1.md
[11]: https://scan.coverity.com/projects/gavinhoward-bc
[12]: ./locale_install.sh
[13]: ./manuals/build.md
[14]: https://github.com/stesser
[15]: https://github.com/bugcrazy
[16]: ./manuals/bc/A.1.md#extended-library
[17]: https://github.com/skeeto/optparse
[18]: https://www.deepl.com/translator
[19]: ./manuals/benchmarks.md
[20]: https://github.com/apjanke/ronn-ng
[21]: https://pandoc.org/