vapier/pcalc

command line util for doing hex/dec/oct/bin math quickly

pcalc: The Programmer's calculator

There was always a loophole when it came to a need to covert between hexadecimal / decimal / octal / binary.

Especially if it involved an operation like 0x1234 + 0x20. It took a lot of hard work, and mostly a good pocket calculator.

Features

• Full math parser, parentheses, add, sub, mult, div, exponential
• Automatic conversion between HEX DEC OCT BIN numbers
• Mixing different bases in one expression
• Definable variables
• Math constants (E PI ...)
• Built in math functions (sin/cos/sqrt ...)

Examples

Here are bunch of common examples.

See the SYNTAX.md file for the full syntax. You can also run pcalc for a local reference.

$ pcalc 0x300
        768                     0x300                   0y1100000000

# Note: Shell expansion on the '*' char requires the single quotes.
$ pcalc '0x300 + 3 * 3'
        777                     0x309                   0y1100001001

$ pcalc sqrt 2
        1.414213562373095       0x1                     0y1

# Note: The single quotes are to keep the shell happy with the () characters.
$ pcalc 'sqrt(2)'
        1.414213562373095       0x1                     0y1

$ pcalc 0x12 or 23
        23                      0x17                    0y10111

# To execute a script, use the `@` prefix on the filename.
$ pcalc @pcalc.001

Building & Installing

To compile the project, simply use make:

$ make

Note: Some platforms might not provide flex or -lfl. You will need to adjust the Makefile accordingly.

To install it, you can do:

$ sudo make install

Testing

To run the unittests, use:

$ make check

Note: Depending on the precision of floating point math in your system, some unittests might fail. It's OK to ignore those specific failures.

Feedback

Please use the GH issue tracker for any updates/ports/suggestions. Feedback always welcome!

https://github.com/vapier/pcalc/issues

License

This project is licensed under GPLv2 (or later). The full text can be found in the COPYING file.

pcalc's Issues

linking on macOS fails: lexer lib is -ll instead of -lfl

Just thought I'd let you know it compiles fine on OS X if you change line 14 of the makefile from:
LDLIBS += -lm lfl
to:
LDLIBS += -lm ll

`./testsuite` does not support `--jobs` under OpenBSD

The generated testsuite file does not support an --jobs flag under OpenBSD. Consider the terminal session below as an example.

Simply setting TESTSUITEFLAGS = appears to resolve this issue.

Tested under OpenBSD 6.4 on an AMD64 based system.

cad@nessus:~/D/pcalc > gmake MAKE=gmake check test
cc -ggdb -Wall -DVERSION='"4"'  -c -o pcalc.o pcalc.c
cc -ggdb -Wall -DVERSION='"4"'  -c -o pcalcl.o pcalcl.c
pcalcl.c:1705:17: warning: unused function 'yyunput' [-Wunused-function]
    static void yyunput (int c, char * yy_bp )
                ^
pcalcl.c:1746:16: warning: function 'input' is not needed and will not be emitted
      [-Wunneeded-internal-declaration]
    static int input  (void)
               ^
2 warnings generated.
cc -ggdb -Wall -DVERSION='"4"'  -c -o funct.o funct.c
cc -ggdb -Wall -DVERSION='"4"'  -c -o math.o math.c
cc -ggdb -Wall -DVERSION='"4"'  -c -o symbol.o symbol.c
cc -ggdb -Wall -DVERSION='"4"'  -c -o help.o help.c
cc -ggdb -Wall -DVERSION='"4"'  -c -o store.o store.c
cc -ggdb -Wall -DVERSION='"4"'  -c -o print.o print.c
cc -ggdb -Wall -DVERSION='"4"'  -c -o str.o str.c
cc -ggdb -Wall -DVERSION='"4"'  -c -o convert.o convert.c
cc -ggdb -Wall  pcalc.o pcalcl.o funct.o math.o symbol.o help.o store.o print.o str.o con
ert.o  -lm -lfl -o pcalc
symbol.o: In function `install_sym':
/home/cad/Downloads/pcalc/symbol.c:171: warning: strcpy() is almost always misused, pleas
 use strlcpy()
store.o: In function `store':
/home/cad/Downloads/pcalc/store.c:93: warning: sprintf() is often misused, please use snp
intf()
gmake -C ptest check
gmake[1]: Entering directory '/home/cad/Downloads/pcalc/ptest'
./testsuite --jobs=`getconf _NPROCESSORS_ONLN || echo 1`
getconf: _NPROCESSORS_ONLN: unknown variable
testsuite: error: invalid variable name: --jobs
gmake[1]: *** [Makefile:39: check] Error 1
gmake[1]: Leaving directory '/home/cad/Downloads/pcalc/ptest'
gmake: *** [Makefile:39: check] Error 2
cad@nessus:~/D/pcalc > ./ptest/testsuite --help
Usage: ./ptest/testsuite [OPTION]... [VARIABLE=VALUE]... [TESTS]

Run all the tests, or the selected TESTS, and save a detailed log file.
Upon failure, create debugging scripts.

You should not change environment variables unless explicitly passed
as command line arguments.  Set `AUTOTEST_PATH' to select the executables
to exercise.  Each relative directory is expanded as build and source
directories relatively to the top level of this distribution.  E.g.,

  $ ./ptest/testsuite AUTOTEST_PATH=bin

possibly amounts into

  PATH=/tmp/foo-1.0/bin:/src/foo-1.0/bin:$PATH

Operation modes:
  -h, --help     print the help message, then exit
  -V, --version  print version number, then exit
  -c, --clean    remove all the files this test suite might create and exit
  -l, --list     describes all the tests, or the selected TESTS

Execution tuning:
  -k, --keywords=KEYWORDS
                 select the tests matching all the comma separated KEYWORDS
                 accumulates
  -e, --errexit  abort as soon as a test fails; implies --debug
  -v, --verbose  force more detailed output
                 default for debugging scripts
  -d, --debug    inhibit clean up and debug script creation
                 default for debugging scripts
  -x, --trace    enable tests shell tracing

Report bugs to <https://github.com/vapier/pcalc/issues>.

Binary shift operators don't work

When I type

$ pcalc 20 < 1
bash: no such file or directory: 1

and

$ pcalc 20 << 1
heredoc> 

Is this a bug or is the syntax wrong?

pcalc install

I have problem in installing pcalc on my opensuse leap 42.2 machine.

Could you please give me hints ?

Regards,

MarcoC

test 12 failes on macppc

testsuite.log

As noted, test 12 fails under OpenBSD 6.4/macppc.

Further potentially useful information:

cad@montag:/u/p/p/p/p/ptest > sysctl hw
hw.machine=macppc
hw.model=7447A (Revision 0x105)
hw.ncpu=1
hw.byteorder=4321
hw.pagesize=4096
hw.disknames=wd0:c375357beab85361,cd0:
hw.diskcount=2
hw.sensors.maxtmp0.temp0=49.25 degC (Internal)
hw.sensors.maxtmp0.temp1=45.00 degC (External)
hw.cpuspeed=1499
hw.vendor=Apple Computer, Inc.
hw.product=PowerMac10,2
hw.physmem=1073741824
hw.usermem=1073729536
hw.ncpufound=1
hw.allowpowerdown=1
hw.ncpuonline=1
cad@montag:/u/p/p/p/p/ptest > uname -a
OpenBSD montag.int.cxd.guru 6.4 GENERIC#231 macppc

numbers larger than 53-bits lose precision due to double normalization

Observed behavior

Expected behavior

It seems to work with smaller hex numbers just fine (e.g. by removing the large ff prefix on these numbers:

pcalc.001 & pcalc.002 tests fail depending on accuracy of host math lib

testsuite.log

The test suite fails on some tests. This appears to be some kind of floating point rounding issue.

This was tested against both afcf2f4 and v4.

The tests were run on an OpenBSD 6.4 system on an AMD64 processor.

Note that the workaround noted in #5 had to be applied in order to run the testes.

Test regression from 4 -> 5 on sparc

Hi, the unit tests which are passing on v4 are failing on v5 on sparc.

Relevant snippet:

## ---------------------- ##
## Detailed failed tests. ##
## ---------------------- ##

#                             -*- compilation -*-
1. pcalc.001.at:1: testing pcalc.001 ...
/pcalc.001.at:1: ${top_builddir:-$at_dir/..}/pcalc @${srcdir:-$at_dir}/pcalc.001 | \
	sed \
		-e 's:[	 ][	 ]*: :g' \
		-e 's: *$::'

--- -	2023-12-20 04:54:09.304649598 +0000
+++ /var/tmp/portage/sci-calculators/pcalc-5/work/pcalc-5/ptest/testsuite.dir/at-groups/1/stdout	2023-12-20 04:54:09.293040691 +0000
@@ -15,7 +15,7 @@
 LOG:
  2.302585092994046 0x2 0y10
  0.4342944819032518 0x0 0y0
- 1 0x1 0y1
+ 1 0x0 0y0
 Fahrenheit to celsius: 80F
  26.66666666666667 0x1a 0y11010
 
1. pcalc.001.at:1: 1. pcalc.001 (pcalc.001.at:1): FAILED (pcalc.001.at:1)


#                             -*- compilation -*-
1. pcalc.001.at:1: testing pcalc.001 ...
/pcalc.001.at:1: ${top_builddir:-$at_dir/..}/pcalc @${srcdir:-$at_dir}/pcalc.001 | \
	sed \
		-e 's:[	 ][	 ]*: :g' \
		-e 's: *$::'

--- -	2023-12-20 04:54:09.304649598 +0000
+++ /var/tmp/portage/sci-calculators/pcalc-5/work/pcalc-5/ptest/testsuite.dir/at-groups/1/stdout	2023-12-20 04:54:09.293040691 +0000
@@ -15,7 +15,7 @@
 LOG:
  2.302585092994046 0x2 0y10
  0.4342944819032518 0x0 0y0
- 1 0x1 0y1
+ 1 0x0 0y0
 Fahrenheit to celsius: 80F
  26.66666666666667 0x1a 0y11010
 
1. pcalc.001.at:1: 1. pcalc.001 (pcalc.001.at:1): FAILED (pcalc.001.at:1)

Downstream bug: https://bugs.gentoo.org/920351

Complete log: https://920351.bugs.gentoo.org/attachment.cgi?id=880026

If you need real hardware to test this on, I offer free shell access to the machine I reproduced this issue on. Thanks for taking a look!

convert number handling to use GMP

the number logic is fine as long as you stick to 64bits, but beyond that, things break down quickly. should look at leveraging gmp for larger values. https://gmplib.org/