armv8l is always 32-bit and should implement the armv7 ISA, so
just use the same filename as for armv7.
This avoids wrongly building a 64-bit binary because of an incorrect assumption
of what 'armv8l' is (a 32-bit system).
Obviously, we should not then build a 64-bit (arm64) binary. Especially
given armv8l is often used in arm32-on-arm64 chroots to build stuff for
weaker-powered arm32 devices.
Signed-off-by: Sam James <sam@gentoo.org>
This commit speeds up the parsing/processing of ANSI escape codes by
roughly 7.5x. The speedup is mostly accomplished by replacing the regex
with dedicated parsing logic (nextAnsiEscapeSequence()) and reducing the
number of allocations in extractColor().
#### Benchmarks
```
name old time/op new time/op delta
ExtractColor-16 4.89µs ± 5% 0.64µs ± 2% -86.87% (p=0.000 n=9+9)
name old speed new speed delta
ExtractColor-16 25.6MB/s ± 5% 194.6MB/s ± 2% +661.43% (p=0.000 n=9+9)
name old alloc/op new alloc/op delta
ExtractColor-16 1.37kB ± 0% 0.31kB ± 0% -77.31% (p=0.000 n=10+10)
name old allocs/op new allocs/op delta
ExtractColor-16 48.0 ± 0% 4.0 ± 0% -91.67% (p=0.000 n=10+10)
```
* Update .travis.yml and use stages
This updates the .travis.yml configuration to use separate stages for
unittests and CLI tests. The output is now clearer, since for unittests
and CLI tests separate web pages are available.
* Use go modules and simplify build
* Add ppc64le support to Makefile
* Update crypt libs to fix tty ioctls on ppc64le
The hardcoded tty ioctl commands in the terminal package were not
correct for ppc64le and caused the ioctls to return ENOTTY for
commands like TCGETS and so on. The bug is fixed in later versions.
