xref: /freebsd/contrib/bc/NEWS.md (revision f32672272f37100dfa23f1f01c16d6388543ba64)
1# News
2
3## 3.2.3
4
5This is a production release that fixes a bug in `gen/strgen.sh`. I recently
6changed `gen/strgen.c`, but I did not change `gen/strgen.sh`.
7
8Users that do not use `gen/strgen.sh` do not need to upgrade.
9
10## 3.2.2
11
12This is a production release that fixes a portability bug in `configure.sh`. The
13bug was using the GNU `find` extension `-wholename`.
14
15## 3.2.1
16
17This is a production release that has one fix for `bcl(3)`. It is technically
18not a bug fix since the behavior is undefined, but the `BclNumber`s that
19`bcl_divmod()` returns will be set to `BCL_ERROR_INVALID_NUM` if there is an
20error. Previously, they were not set.
21
22## 3.2.0
23
24This is a production release that has one bug fix and a major addition.
25
26The bug fix was a missing `auto` variable in the bessel `j()` function in the
27math library.
28
29The major addition is a way to build a version of `bc`'s math code as a library.
30This is done with the `-a` option to `configure.sh`. The API for the library can
31be read in `./manuals/bcl.3.md` or `man bcl` once the library is installed with
32`make install`.
33
34This library was requested by developers before I even finished version 1.0, but
35I could not figure out how to do it until now.
36
37If the library has API breaking changes, the major version of `bc` will be
38incremented.
39
40## 3.1.6
41
42This is a production release that fixes a new warning from Clang 12 for FreeBSD
43and also removes some possible undefined behavior found by UBSan that compilers
44did not seem to take advantage of.
45
46Users do ***NOT*** need to upgrade, if they do not want to.
47
48## 3.1.5
49
50This is a production release that fixes the Chinese locales (which caused `bc`
51to crash) and a crash caused by `bc` executing code when it should not have been
52able to.
53
54***ALL USERS SHOULD UPGRADE.***
55
56## 3.1.4
57
58This is a production release that fixes one bug, changes two behaviors, and
59removes one environment variable.
60
61The bug is like the one in the last release except it applies if files are being
62executed. I also made the fix more general.
63
64The behavior that was changed is that `bc` now exits when given `-e`, `-f`,
65`--expression` or `--file`. However, if the last one of those is `-f-` (using
66`stdin` as the file), `bc` does not exit. If `-f-` exists and is not the last of
67the `-e` and `-f` options (and equivalents), `bc` gives a fatal error and exits.
68
69Next, I removed the `BC_EXPR_EXIT` and `DC_EXPR_EXIT` environment variables
70since their use is not needed with the behavior change.
71
72Finally, I made it so `bc` does not print the header, though the `-q` and
73`--quiet` options were kept for compatibility with GNU `bc`.
74
75## 3.1.3
76
77This is a production release that fixes one minor bug: if `bc` was invoked like
78the following, it would error:
79
80```
81echo "if (1 < 3) 1" | bc
82```
83
84Unless users run into this bug, they do not need to upgrade, but it is suggested
85that they do.
86
87## 3.1.2
88
89This is a production release that adds a way to install *all* locales. Users do
90***NOT*** need to upgrade.
91
92For package maintainers wishing to make use of the change, just pass `-l` to
93`configure.sh`.
94
95## 3.1.1
96
97This is a production release that adds two Spanish locales. Users do ***NOT***
98need to upgrade, unless they want those locales.
99
100## 3.1.0
101
102This is a production release that adjusts one behavior, fixes eight bugs, and
103improves manpages for FreeBSD. Because this release fixes bugs, **users and
104package maintainers should update to this version as soon as possible**.
105
106The behavior that was adjusted was how code from the `-e` and `-f` arguments
107(and equivalents) were executed. They used to be executed as one big chunk, but
108in this release, they are now executed line-by-line.
109
110The first bug fix in how output to `stdout` was handled in `SIGINT`. If a
111`SIGINT` came in, the `stdout` buffer was not correctly flushed. In fact, a
112clean-up function was not getting called. This release fixes that bug.
113
114The second bug is in how `dc` handled input from `stdin`. This affected `bc` as
115well since it was a mishandling of the `stdin` buffer.
116
117The third fixed bug was that `bc` and `dc` could `abort()` (in debug mode) when
118receiving a `SIGTERM`. This one was a race condition with pushing and popping
119items onto and out of vectors.
120
121The fourth bug fixed was that `bc` could leave extra items on the stack and
122thus, not properly clean up some memory. (The memory would still get
123`free()`'ed, but it would not be `free()`'ed when it could have been.)
124
125The next two bugs were bugs in `bc`'s parser that caused crashes when executing
126the resulting code.
127
128The last two bugs were crashes in `dc` that resulted from mishandling of
129strings.
130
131The manpage improvement was done by switching from [ronn][20] to [Pandoc][21] to
132generate manpages. Pandoc generates much cleaner manpages and doesn't leave
133blank lines where they shouldn't be.
134
135## 3.0.3
136
137This is a production release that adds one new feature: specific manpages.
138
139Before this release, `bc` and `dc` only used one manpage each that referred to
140various build options. This release changes it so there is one manpage set per
141relevant build type. Each manual only has information about its particular
142build, and `configure.sh` selects the correct set for install.
143
144## 3.0.2
145
146This is a production release that adds `utf8` locale symlinks and removes an
147unused `auto` variable from the `ceil()` function in the [extended math
148library][16].
149
150Users do ***NOT*** need to update unless they want the locales.
151
152## 3.0.1
153
154This is a production release with two small changes. Users do ***NOT*** need to
155upgrade to this release; however, if they haven't upgraded to `3.0.0` yet, it
156may be worthwhile to upgrade to this release.
157
158The first change is fixing a compiler warning on FreeBSD with strict warnings
159on.
160
161The second change is to make the new implementation of `ceil()` in `lib2.bc`
162much more efficient.
163
164## 3.0.0
165
166*Notes for package maintainers:*
167
168*First, the `2.7.0` release series saw a change in the option parsing. This made
169me change one error message and add a few others. The error message that was
170changed removed one format specifier. This means that `printf()` will seqfault
171on old locale files. Unfortunately, `bc` cannot use any locale files except the
172global ones that are already installed, so it will use the previous ones while
173running tests during install. **If `bc` segfaults while running arg tests when
174updating, it is because the global locale files have not been replaced. Make
175sure to either prevent the test suite from running on update or remove the old
176locale files before updating.** (Removing the locale files can be done with
177`make uninstall` or by running the `locale_uninstall.sh` script.) Once this is
178done, `bc` should install without problems.*
179
180*Second, **the option to build without signal support has been removed**. See
181below for the reasons why.*
182
183This is a production release with some small bug fixes, a few improvements,
184three major bug fixes, and a complete redesign of `bc`'s error and signal
185handling. **Users and package maintainers should update to this version as soon
186as possible.**
187
188The first major bug fix was in how `bc` executed files. Previously, a whole file
189was parsed before it was executed, but if a function is defined *after* code,
190especially if the function definition was actually a redefinition, and the code
191before the definition referred to the previous function, this `bc` would replace
192the function before executing any code. The fix was to make sure that all code
193that existed before a function definition was executed.
194
195The second major bug fix was in `bc`'s `lib2.bc`. The `ceil()` function had a
196bug where a `0` in the decimal place after the truncation position, caused it to
197output the wrong numbers if there was any non-zero digit after.
198
199The third major bug is that when passing parameters to functions, if an
200expression included an array (not an array element) as a parameter, it was
201accepted, when it should have been rejected. It is now correctly rejected.
202
203Beyond that, this `bc` got several improvements that both sped it up, improved
204the handling of signals, and improved the error handling.
205
206First, the requirements for `bc` were pushed back to POSIX 2008. `bc` uses one
207function, `strdup()`, which is not in POSIX 2001, and it is in the X/Open System
208Interfaces group 2001. It is, however, in POSIX 2008, and since POSIX 2008 is
209old enough to be supported anywhere that I care, that should be the requirement.
210
211Second, the BcVm global variable was put into `bss`. This actually slightly
212reduces the size of the executable from a massive code shrink, and it will stop
213`bc` from allocating a large set of memory when `bc` starts.
214
215Third, the default Karatsuba length was updated from 64 to 32 after making the
216optimization changes below, since 32 is going to be better than 64 after the
217changes.
218
219Fourth, Spanish translations were added.
220
221Fifth, the interpreter received a speedup to make performance on non-math-heavy
222scripts more competitive with GNU `bc`. While improvements did, in fact, get it
223much closer (see the [benchmarks][19]), it isn't quite there.
224
225There were several things done to speed up the interpreter:
226
227First, several small inefficiencies were removed. These inefficiencies included
228calling the function `bc_vec_pop(v)` twice instead of calling
229`bc_vec_npop(v, 2)`. They also included an extra function call for checking the
230size of the stack and checking the size of the stack more than once on several
231operations.
232
233Second, since the current `bc` function is the one that stores constants and
234strings, the program caches pointers to the current function's vectors of
235constants and strings to prevent needing to grab the current function in order
236to grab a constant or a string.
237
238Third, `bc` tries to reuse `BcNum`'s (the internal representation of
239arbitary-precision numbers). If a `BcNum` has the default capacity of
240`BC_NUM_DEF_SIZE` (32 on 64-bit and 16 on 32-bit) when it is freed, it is added
241to a list of available `BcNum`'s. And then, when a `BcNum` is allocated with a
242capacity of `BC_NUM_DEF_SIZE` and any `BcNum`'s exist on the list of reusable
243ones, one of those ones is grabbed instead.
244
245In order to support these changes, the `BC_NUM_DEF_SIZE` was changed. It used to
246be 16 bytes on all systems, but it was changed to more closely align with the
247minimum allocation size on Linux, which is either 32 bytes (64-bit musl), 24
248bytes (64-bit glibc), 16 bytes (32-bit musl), or 12 bytes (32-bit glibc). Since
249these are the minimum allocation sizes, these are the sizes that would be
250allocated anyway, making it worth it to just use the whole space, so the value
251of `BC_NUM_DEF_SIZE` on 64-bit systems was changed to 32 bytes.
252
253On top of that, at least on 64-bit, `BC_NUM_DEF_SIZE` supports numbers with
254either 72 integer digits or 45 integer digits and 27 fractional digits. This
255should be more than enough for most cases since `bc`'s default `scale` values
256are 0 or 20, meaning that, by default, it has at most 20 fractional digits. And
25745 integer digits are *a lot*; it's enough to calculate the amount of mass in
258the Milky Way galaxy in kilograms. Also, 72 digits is enough to calculate the
259diameter of the universe in Planck lengths.
260
261(For 32-bit, these numbers are either 32 integer digits or 12 integer digits and
26220 fractional digits. These are also quite big, and going much bigger on a
26332-bit system seems a little pointless since 12 digits in just under a trillion
264and 20 fractional digits is still enough for about any use since `10^-20` light
265years is just under a millimeter.)
266
267All of this together means that for ordinary uses, and even uses in scientific
268work, the default number size will be all that is needed, which means that
269nearly all, if not all, numbers will be reused, relieving pressure on the system
270allocator.
271
272I did several experiments to find the changes that had the most impact,
273especially with regard to reusing `BcNum`'s. One was putting `BcNum`'s into
274buckets according to their capacity in powers of 2 up to 512. That performed
275worse than `bc` did in `2.7.2`. Another was putting any `BcNum` on the reuse
276list that had a capacity of `BC_NUM_DEF_SIZE * 2` and reusing them for `BcNum`'s
277that requested `BC_NUM_DEF_SIZE`. This did reduce the amount of time spent, but
278it also spent a lot of time in the system allocator for an unknown reason. (When
279using `strace`, a bunch more `brk` calls showed up.) Just reusing `BcNum`'s that
280had exactly `BC_NUM_DEF_SIZE` capacity spent the smallest amount of time in both
281user and system time. This makes sense, especially with the changes to make
282`BC_NUM_DEF_SIZE` bigger on 64-bit systems, since the vast majority of numbers
283will only ever use numbers with a size less than or equal to `BC_NUM_DEF_SIZE`.
284
285Last of all, `bc`'s signal handling underwent a complete redesign. (This is the
286reason that this version is `3.0.0` and not `2.8.0`.) The change was to move
287from a polling approach to signal handling to an interrupt-based approach.
288
289Previously, every single loop condition had a check for signals. I suspect that
290this could be expensive when in tight loops.
291
292Now, the signal handler just uses `longjmp()` (actually `siglongjmp()`) to start
293an unwinding of the stack until it is stopped or the stack is unwound to
294`main()`, which just returns. If `bc` is currently executing code that cannot be
295safely interrupted (according to POSIX), then signals are "locked." The signal
296handler checks if the lock is taken, and if it is, it just sets the status to
297indicate that a signal arrived. Later, when the signal lock is released, the
298status is checked to see if a signal came in. If so, the stack unwinding starts.
299
300This design eliminates polling in favor of maintaining a stack of `jmp_buf`'s.
301This has its own performance implications, but it gives better interaction. And
302the cost of pushing and popping a `jmp_buf` in a function is paid at most twice.
303Most functions do not pay that price, and most of the rest only pay it once.
304(There are only some 3 functions in `bc` that push and pop a `jmp_buf` twice.)
305
306As a side effect of this change, I had to eliminate the use of `stdio.h` in `bc`
307because `stdio` does not play nice with signals and `longjmp()`. I implemented
308custom I/O buffer code that takes a fraction of the size. This means that static
309builds will be smaller, but non-static builds will be bigger, though they will
310have less linking time.
311
312This change is also good because my history implementation was already bypassing
313`stdio` for good reasons, and unifying the architecture was a win.
314
315Another reason for this change is that my `bc` should *always* behave correctly
316in the presence of signals like `SIGINT`, `SIGTERM`, and `SIGQUIT`. With the
317addition of my own I/O buffering, I needed to also make sure that the buffers
318were correctly flushed even when such signals happened.
319
320For this reason, I **removed the option to build without signal support**.
321
322As a nice side effect of this change, the error handling code could be changed
323to take advantage of the stack unwinding that signals used. This means that
324signals and error handling use the same code paths, which means that the stack
325unwinding is well-tested. (Errors are tested heavily in the test suite.)
326
327It also means that functions do not need to return a status code that
328***every*** caller needs to check. This eliminated over 100 branches that simply
329checked return codes and then passed that return code up the stack if necessary.
330The code bloat savings from this is at least 1700 bytes on `x86_64`, *before*
331taking into account the extra code from removing `stdio.h`.
332
333## 2.7.2
334
335This is a production release with one major bug fix.
336
337The `length()` built-in function can take either a number or an array. If it
338takes an array, it returns the length of the array. Arrays can be passed by
339reference. The bug is that the `length()` function would not properly
340dereference arrays that were references. This is a bug that affects all users.
341
342**ALL USERS SHOULD UPDATE `bc`**.
343
344## 2.7.1
345
346This is a production release with fixes for new locales and fixes for compiler
347warnings on FreeBSD.
348
349## 2.7.0
350
351This is a production release with a bug fix for Linux, new translations, and new
352features.
353
354Bug fixes:
355
356* Option parsing in `BC_ENV_ARGS` was broken on Linux in 2.6.1 because `glibc`'s
357  `getopt_long()` is broken. To get around that, and to support long options on
358  every platform, an adapted version of [`optparse`][17] was added. Now, `bc`
359  does not even use `getopt()`.
360* Parsing `BC_ENV_ARGS` with quotes now works. It isn't the smartest, but it
361  does the job if there are spaces in file names.
362
363The following new languages are supported:
364
365* Dutch
366* Polish
367* Russian
368* Japanes
369* Simplified Chinese
370
371All of these translations were generated using [DeepL][18], so improvements are
372welcome.
373
374There is only one new feature: **`bc` now has a built-in pseudo-random number
375generator** (PRNG).
376
377The PRNG is seeded, making it useful for applications where
378`/dev/urandom` does not work because output needs to be reproducible. However,
379it also uses `/dev/urandom` to seed itself by default, so it will start with a
380good seed by default.
381
382It also outputs 32 bits on 32-bit platforms and 64 bits on 64-bit platforms, far
383better than the 15 bits of C's `rand()` and `bash`'s `$RANDOM`.
384
385In addition, the PRNG can take a bound, and when it gets a bound, it
386automatically adjusts to remove bias. It can also generate numbers of arbitrary
387size. (As of the time of release, the largest pseudo-random number generated by
388this `bc` was generated with a bound of `2^(2^20)`.)
389
390***IMPORTANT: read the [`bc` manual][9] and the [`dc` manual][10] to find out
391exactly what guarantees the PRNG provides. The underlying implementation is not
392guaranteed to stay the same, but the guarantees that it provides are guaranteed
393to stay the same regardless of the implementation.***
394
395On top of that, four functions were added to `bc`'s [extended math library][16]
396to make using the PRNG easier:
397
398* `frand(p)`: Generates a number between `[0,1)` to `p` decimal places.
399* `ifrand(i, p)`: Generates an integer with bound `i` and adds it to `frand(p)`.
400* `srand(x)`: Randomizes the sign of `x`. In other words, it flips the sign of
401  `x` with probability `0.5`.
402* `brand()`: Returns a random boolean value (either `0` or `1`).
403
404## 2.6.1
405
406This is a production release with a bug fix for FreeBSD.
407
408The bug was that when `bc` was built without long options, it would give a fatal
409error on every run. This was caused by a mishandling of `optind`.
410
411## 2.6.0
412
413This release is a production release ***with no bugfixes***. If you do not want
414to upgrade, you don't have to.
415
416No source code changed; the only thing that changed was `lib2.bc`.
417
418This release adds one function to the [extended math library][16]: `p(x, y)`,
419which calculates `x` to the power of `y`, whether or not `y` is an integer. (The
420`^` operator can only accept integer powers.)
421
422This release also includes a couple of small tweaks to the [extended math
423library][16], mostly to fix returning numbers with too high of `scale`.
424
425## 2.5.3
426
427This release is a production release which addresses inconsistencies in the
428Portuguese locales. No `bc` code was changed.
429
430The issues were that the ISO files used different naming, and also that the
431files that should have been symlinks were not. I did not catch that because
432GitHub rendered them the exact same way.
433
434## 2.5.2
435
436This release is a production release.
437
438No code was changed, but the build system was changed to allow `CFLAGS` to be
439given to `CC`, like this:
440
441```
442CC="gcc -O3 -march=native" ./configure.sh
443```
444
445If this happens, the flags are automatically put into `CFLAGS`, and the compiler
446is set appropriately. In the example above this means that `CC` will be "gcc"
447and `CFLAGS` will be "-O3 -march=native".
448
449This behavior was added to conform to GNU autotools practices.
450
451## 2.5.1
452
453This is a production release which addresses portability concerns discovered
454in the `bc` build system. No `bc` code was changed.
455
456* Support for Solaris SPARC and AIX were added.
457* Minor documentations edits were performed.
458* An option for `configure.sh` was added to disable long options if
459  `getopt_long()` is missing.
460
461## 2.5.0
462
463This is a production release with new translations. No code changed.
464
465The translations were contributed by [bugcrazy][15], and they are for
466Portuguese, both Portugal and Brazil locales.
467
468## 2.4.0
469
470This is a production release primarily aimed at improving `dc`.
471
472* A couple of copy and paste errors in the [`dc` manual][10] were fixed.
473* `dc` startup was optimized by making sure it didn't have to set up `bc`-only
474  things.
475* The `bc` `&&` and `||` operators were made available to `dc` through the `M`
476  and `m` commands, respectively.
477* `dc` macros were changed to be tail call-optimized.
478
479The last item, tail call optimization, means that if the last thing in a macro
480is a call to another macro, then the old macro is popped before executing the
481new macro. This change was made to stop `dc` from consuming more and more memory
482as macros are executed in a loop.
483
484The `q` and `Q` commands still respect the "hidden" macros by way of recording
485how many macros were removed by tail call optimization.
486
487## 2.3.2
488
489This is a production release meant to fix warnings in the Gentoo `ebuild` by
490making it possible to disable binary stripping. Other users do *not* need to
491upgrade.
492
493## 2.3.1
494
495This is a production release. It fixes a bug that caused `-1000000000 < -1` to
496return `0`. This only happened with negative numbers and only if the value on
497the left was more negative by a certain amount. That said, this bug *is* a bad
498bug, and needs to be fixed.
499
500**ALL USERS SHOULD UPDATE `bc`**.
501
502## 2.3.0
503
504This is a production release with changes to the build system.
505
506## 2.2.0
507
508This release is a production release. It only has new features and performance
509improvements.
510
5111.	The performance of `sqrt(x)` was improved.
5122.	The new function `root(x, n)` was added to the extended math library to
513	calculate `n`th roots.
5143.	The new function `cbrt(x)` was added to the extended math library to
515	calculate cube roots.
516
517## 2.1.3
518
519This is a non-critical release; it just changes the build system, and in
520non-breaking ways:
521
5221.	Linked locale files were changed to link to their sources with a relative
523	link.
5242.	A bug in `configure.sh` that caused long option parsing to fail under `bash`
525	was fixed.
526
527## 2.1.2
528
529This release is not a critical release.
530
5311.	A few codes were added to history.
5322.	Multiplication was optimized a bit more.
5333.	Addition and subtraction were both optimized a bit more.
534
535## 2.1.1
536
537This release contains a fix for the test suite made for Linux from Scratch: now
538the test suite prints `pass` when a test is passed.
539
540Other than that, there is no change in this release, so distros and other users
541do not need to upgrade.
542
543## 2.1.0
544
545This release is a production release.
546
547The following bugs were fixed:
548
5491.	A `dc` bug that caused stack mishandling was fixed.
5502.	A warning on OpenBSD was fixed.
5513.	Bugs in `ctrl+arrow` operations in history were fixed.
5524.	The ability to paste multiple lines in history was added.
5535.	A `bc` bug, mishandling of array arguments to functions, was fixed.
5546.	A crash caused by freeing the wrong pointer was fixed.
5557.	A `dc` bug where strings, in a rare case, were mishandled in parsing was
556	fixed.
557
558In addition, the following changes were made:
559
5601.	Division was slightly optimized.
5612.	An option was added to the build to disable printing of prompts.
5623.	The special case of empty arguments is now handled. This is to prevent
563	errors in scripts that end up passing empty arguments.
5644.	A harmless bug was fixed. This bug was that, with the pop instructions
565	(mostly) removed (see below), `bc` would leave extra values on its stack for
566	`void` functions and in a few other cases. These extra items would not
567	affect anything put on the stack and would not cause any sort of crash or
568	even buggy behavior, but they would cause `bc` to take more memory than it
569	needed.
570
571On top of the above changes, the following optimizations were added:
572
5731.	The need for pop instructions in `bc` was removed.
5742.	Extra tests on every iteration of the interpreter loop were removed.
5753.	Updating function and code pointers on every iteration of the interpreter
576	loop was changed to only updating them when necessary.
5774.	Extra assignments to pointers were removed.
578
579Altogether, these changes sped up the interpreter by around 2x.
580
581***NOTE***: This is the last release with new features because this `bc` is now
582considered complete. From now on, only bug fixes and new translations will be
583added to this `bc`.
584
585## 2.0.3
586
587This is a production, bug-fix release.
588
589Two bugs were fixed in this release:
590
5911.	A rare and subtle signal handling bug was fixed.
5922.	A misbehavior on `0` to a negative power was fixed.
593
594The last bug bears some mentioning.
595
596When I originally wrote power, I did not thoroughly check its error cases;
597instead, I had it check if the first number was `0` and then if so, just return
598`0`. However, `0` to a negative power means that `1` will be divided by `0`,
599which is an error.
600
601I caught this, but only after I stopped being cocky. You see, sometime later, I
602had noticed that GNU `bc` returned an error, correctly, but I thought it was
603wrong simply because that's not what my `bc` did. I saw it again later and had a
604double take. I checked for real, finally, and found out that my `bc` was wrong
605all along.
606
607That was bad on me. But the bug was easy to fix, so it is fixed now.
608
609There are two other things in this release:
610
6111.	Subtraction was optimized by [Stefan Eßer][14].
6122.	Division was also optimized, also by Stefan Eßer.
613
614## 2.0.2
615
616This release contains a fix for a possible overflow in the signal handling. I
617would be surprised if any users ran into it because it would only happen after 2
618billion (`2^31-1`) `SIGINT`'s, but I saw it and had to fix it.
619
620## 2.0.1
621
622This release contains very few things that will apply to any users.
623
6241.	A slight bug in `dc`'s interactive mode was fixed.
6252.	A bug in the test suite that was only triggered on NetBSD was fixed.
6263.	**The `-P`/`--no-prompt` option** was added for users that do not want a
627	prompt.
6284.	A `make check` target was added as an alias for `make test`.
6295.	`dc` got its own read prompt: `?> `.
630
631## 2.0.0
632
633This release is a production release.
634
635This release is also a little different from previous releases. From here on
636out, I do not plan on adding any more features to this `bc`; I believe that it
637is complete. However, there may be bug fix releases in the future, if I or any
638others manage to find bugs.
639
640This release has only a few new features:
641
6421.	`atan2(y, x)` was added to the extended math library as both `a2(y, x)` and
643	`atan2(y, x)`.
6442.	Locales were fixed.
6453.	A **POSIX shell-compatible script was added as an alternative to compiling
646	`gen/strgen.c`** on a host machine. More details about making the choice
647	between the two can be found by running `./configure.sh --help` or reading
648	the [build manual][13].
6494.	Multiplication was optimized by using **diagonal multiplication**, rather
650	than straight brute force.
6515.	The `locale_install.sh` script was fixed.
6526.	`dc` was given the ability to **use the environment variable
653	`DC_ENV_ARGS`**.
6547.	`dc` was also given the ability to **use the `-i` or `--interactive`**
655	options.
6568.	Printing the prompt was fixed so that it did not print when it shouldn't.
6579.	Signal handling was fixed.
65810.	**Handling of `SIGTERM` and `SIGQUIT`** was fixed.
65911.	The **built-in functions `maxibase()`, `maxobase()`, and `maxscale()`** (the
660	commands `T`, `U`, `V` in `dc`, respectively) were added to allow scripts to
661	query for the max allowable values of those globals.
66212.	Some incompatibilities with POSIX were fixed.
663
664In addition, this release is `2.0.0` for a big reason: the internal format for
665numbers changed. They used to be a `char` array. Now, they are an array of
666larger integers, packing more decimal digits into each integer. This has
667delivered ***HUGE*** performance improvements, especially for multiplication,
668division, and power.
669
670This `bc` should now be the fastest `bc` available, but I may be wrong.
671
672## 1.2.8
673
674This release contains a fix for a harmless bug (it is harmless in that it still
675works, but it just copies extra data) in the [`locale_install.sh`][12] script.
676
677## 1.2.7
678
679This version contains fixes for the build on Arch Linux.
680
681## 1.2.6
682
683This release removes the use of `local` in shell scripts because it's not POSIX
684shell-compatible, and also updates a man page that should have been updated a
685long time ago but was missed.
686
687## 1.2.5
688
689This release contains some missing locale `*.msg` files.
690
691## 1.2.4
692
693This release contains a few bug fixes and new French translations.
694
695## 1.2.3
696
697This release contains a fix for a bug: use of uninitialized data. Such data was
698only used when outputting an error message, but I am striving for perfection. As
699Michelangelo said, "Trifles make perfection, and perfection is no trifle."
700
701## 1.2.2
702
703This release contains fixes for OpenBSD.
704
705## 1.2.1
706
707This release contains bug fixes for some rare bugs.
708
709## 1.2.0
710
711This is a production release.
712
713There have been several changes since `1.1.0`:
714
7151.	The build system had some changes.
7162.	Locale support has been added. (Patches welcome for translations.)
7173.	**The ability to turn `ibase`, `obase`, and `scale` into stacks** was added
718	with the `-g` command-line option. (See the [`bc` manual][9] for more
719	details.)
7204.	Support for compiling on Mac OSX out of the box was added.
7215.	The extended math library got `t(x)`, `ceil(x)`, and some aliases.
7226.	The extended math library also got `r2d(x)` (for converting from radians to
723	degrees) and `d2r(x)` (for converting from degrees to radians). This is to
724	allow using degrees with the standard library.
7257.	Both calculators now accept numbers in **scientific notation**. See the
726	[`bc` manual][9] and the [`dc` manual][10] for details.
7278.	Both calculators can **output in either scientific or engineering
728	notation**. See the [`bc` manual][9] and the [`dc` manual][10] for details.
7299.	Some inefficiencies were removed.
73010.	Some bugs were fixed.
73111.	Some bugs in the extended library were fixed.
73212.	Some defects from [Coverity Scan][11] were fixed.
733
734## 1.1.4
735
736This release contains a fix to the build system that allows it to build on older
737versions of `glibc`.
738
739## 1.1.3
740
741This release contains a fix for a bug in the test suite where `bc` tests and
742`dc` tests could not be run in parallel.
743
744## 1.1.2
745
746This release has a fix for a history bug; the down arrow did not work.
747
748## 1.1.1
749
750This release fixes a bug in the `1.1.0` build system. The source is exactly the
751same.
752
753The bug that was fixed was a failure to install if no `EXECSUFFIX` was used.
754
755## 1.1.0
756
757This is a production release. However, many new features were added since `1.0`.
758
7591.	**The build system has been changed** to use a custom, POSIX
760	shell-compatible configure script ([`configure.sh`][6]) to generate a POSIX
761	make-compatible `Makefile`, which means that `bc` and `dc` now build out of
762	the box on any POSIX-compatible system.
7632.	Out-of-memory and output errors now cause the `bc` to report the error,
764	clean up, and die, rather than just reporting and trying to continue.
7653.	**Strings and constants are now garbage collected** when possible.
7664.	Signal handling and checking has been made more simple and more thorough.
7675.	`BcGlobals` was refactored into `BcVm` and `BcVm` was made global. Some
768	procedure names were changed to reflect its difference to everything else.
7696.	Addition got a speed improvement.
7707.	Some common code for addition and multiplication was refactored into its own
771	procedure.
7728.	A bug was removed where `dc` could have been selected, but the internal
773	`#define` that returned `true` for a query about `dc` would not have
774	returned `true`.
7759.	Useless calls to `bc_num_zero()` were removed.
77610.	**History support was added.** The history support is based off of a
777	[UTF-8 aware fork][7] of [`linenoise`][8], which has been customized with
778	`bc`'s own data structures and signal handling.
77911.	Generating C source from the math library now removes tabs from the library,
780	shrinking the size of the executable.
78112.	The math library was shrunk.
78213.	Error handling and reporting was improved.
78314.	Reallocations were reduced by giving access to the request size for each
784	operation.
78515.	**`abs()` (`b` command for `dc`) was added as a builtin.**
78616.	Both calculators were tested on FreeBSD.
78717.	Many obscure parse bugs were fixed.
78818.	Markdown and man page manuals were added, and the man pages are installed by
789	`make install`.
79019.	Executable size was reduced, though the added features probably made the
791	executable end up bigger.
79220.	**GNU-style array references were added as a supported feature.**
79321.	Allocations were reduced.
79422.	**New operators were added**: `$` (`$` for `dc`), `@` (`@` for `dc`), `@=`,
795	`<<` (`H` for `dc`), `<<=`, `>>` (`h` for `dc`), and `>>=`. See the
796	[`bc` manual][9] and the [`dc` manual][10] for more details.
79723.	**An extended math library was added.** This library contains code that
798	makes it so I can replace my desktop calculator with this `bc`. See the
799	[`bc` manual][3] for more details.
80024.	Support for all capital letters as numbers was added.
80125.	**Support for GNU-style void functions was added.**
80226.	A bug fix for improper handling of function parameters was added.
80327.	Precedence for the or (`||`) operator was changed to match GNU `bc`.
80428.	`dc` was given an explicit negation command.
80529.	`dc` was changed to be able to handle strings in arrays.
806
807## 1.1 Release Candidate 3
808
809This release is the eighth release candidate for 1.1, though it is the third
810release candidate meant as a general release candidate. The new code has not
811been tested as thoroughly as it should for release.
812
813## 1.1 Release Candidate 2
814
815This release is the seventh release candidate for 1.1, though it is the second
816release candidate meant as a general release candidate. The new code has not
817been tested as thoroughly as it should for release.
818
819## 1.1 FreeBSD Beta 5
820
821This release is the sixth release candidate for 1.1, though it is the fifth
822release candidate meant specifically to test if `bc` works on FreeBSD. The new
823code has not been tested as thoroughly as it should for release.
824
825## 1.1 FreeBSD Beta 4
826
827This release is the fifth release candidate for 1.1, though it is the fourth
828release candidate meant specifically to test if `bc` works on FreeBSD. The new
829code has not been tested as thoroughly as it should for release.
830
831## 1.1 FreeBSD Beta 3
832
833This release is the fourth release candidate for 1.1, though it is the third
834release candidate meant specifically to test if `bc` works on FreeBSD. The new
835code has not been tested as thoroughly as it should for release.
836
837## 1.1 FreeBSD Beta 2
838
839This release is the third release candidate for 1.1, though it is the second
840release candidate meant specifically to test if `bc` works on FreeBSD. The new
841code has not been tested as thoroughly as it should for release.
842
843## 1.1 FreeBSD Beta 1
844
845This release is the second release candidate for 1.1, though it is meant
846specifically to test if `bc` works on FreeBSD. The new code has not been tested as
847thoroughly as it should for release.
848
849## 1.1 Release Candidate 1
850
851This is the first release candidate for 1.1. The new code has not been tested as
852thoroughly as it should for release.
853
854## 1.0
855
856This is the first non-beta release. `bc` is ready for production use.
857
858As such, a lot has changed since 0.5.
859
8601.	`dc` has been added. It has been tested even more thoroughly than `bc` was
861	for `0.5`. It does not have the `!` command, and for security reasons, it
862	never will, so it is complete.
8632.	`bc` has been more thoroughly tested. An entire section of the test suite
864	(for both programs) has been added to test for errors.
8653.	A prompt (`>>> `) has been added for interactive mode, making it easier to
866	see inputs and outputs.
8674.	Interrupt handling has been improved, including elimination of race
868	conditions (as much as possible).
8695.	MinGW and [Windows Subsystem for Linux][1] support has been added (see
870	[xstatic][2] for binaries).
8716.	Memory leaks and errors have been eliminated (as far as ASan and Valgrind
872	can tell).
8737.	Crashes have been eliminated (as far as [afl][3] can tell).
8748.	Karatsuba multiplication was added (and thoroughly) tested, speeding up
875	multiplication and power by orders of magnitude.
8769.	Performance was further enhanced by using a "divmod" function to reduce
877	redundant divisions and by removing superfluous `memset()` calls.
87810.	To switch between Karatsuba and `O(n^2)` multiplication, the config variable
879	`BC_NUM_KARATSUBA_LEN` was added. It is set to a sane default, but the
880	optimal number can be found with [`karatsuba.py`][4] (requires Python 3)
881	and then configured through `make`.
88211.	The random math test generator script was changed to Python 3 and improved.
883	`bc` and `dc` have together been run through 30+ million random tests.
88412.	All known math bugs have been fixed, including out of control memory
885	allocations in `sine` and `cosine` (that was actually a parse bug), certain
886	cases of infinite loop on square root, and slight inaccuracies (as much as
887	possible; see the [README][5]) in transcendental functions.
88813.	Parsing has been fixed as much as possible.
88914.	Test coverage was improved to 94.8%. The only paths not covered are ones
890	that happen when `malloc()` or `realloc()` fails.
89115.	An extension to get the length of an array was added.
89216.	The boolean not (`!`) had its precedence change to match negation.
89317.	Data input was hardened.
89418.	`bc` was made fully compliant with POSIX when the `-s` flag is used or
895	`POSIXLY_CORRECT` is defined.
89619.	Error handling was improved.
89720.	`bc` now checks that files it is given are not directories.
898
899## 1.0 Release Candidate 7
900
901This is the seventh release candidate for 1.0. It fixes a few bugs in 1.0
902Release Candidate 6.
903
904## 1.0 Release Candidate 6
905
906This is the sixth release candidate for 1.0. It fixes a few bugs in 1.0 Release
907Candidate 5.
908
909## 1.0 Release Candidate 5
910
911This is the fifth release candidate for 1.0. It fixes a few bugs in 1.0 Release
912Candidate 4.
913
914## 1.0 Release Candidate 4
915
916This is the fourth release candidate for 1.0. It fixes a few bugs in 1.0 Release
917Candidate 3.
918
919## 1.0 Release Candidate 3
920
921This is the third release candidate for 1.0. It fixes a few bugs in 1.0 Release
922Candidate 2.
923
924## 1.0 Release Candidate 2
925
926This is the second release candidate for 1.0. It fixes a few bugs in 1.0 Release
927Candidate 1.
928
929## 1.0 Release Candidate 1
930
931This is the first Release Candidate for 1.0. `bc` is complete, with `dc`, but it
932is not tested.
933
934## 0.5
935
936This beta release completes more features, but it is still not complete nor
937tested as thoroughly as necessary.
938
939## 0.4.1
940
941This beta release fixes a few bugs in 0.4.
942
943## 0.4
944
945This is a beta release. It does not have the complete set of features, and it is
946not thoroughly tested.
947
948[1]: https://docs.microsoft.com/en-us/windows/wsl/install-win10
949[2]: https://pkg.musl.cc/bc/
950[3]: http://lcamtuf.coredump.cx/afl/
951[4]: ./karatsuba.py
952[5]: ./README.md
953[6]: ./configure.sh
954[7]: https://github.com/rain-1/linenoise-mob
955[8]: https://github.com/antirez/linenoise
956[9]: ./manuals/bc/A.1.md
957[10]: ./manuals/dc/A.1.md
958[11]: https://scan.coverity.com/projects/gavinhoward-bc
959[12]: ./locale_install.sh
960[13]: ./manuals/build.md
961[14]: https://github.com/stesser
962[15]: https://github.com/bugcrazy
963[16]: ./manuals/bc/A.1.md#extended-library
964[17]: https://github.com/skeeto/optparse
965[18]: https://www.deepl.com/translator
966[19]: ./manuals/benchmarks.md
967[20]: https://github.com/apjanke/ronn-ng
968[21]: https://pandoc.org/
969