Changes:
* argument sanity checks:
  * `PKGNAME[-PKGVER] | SRCURL`
    * check if missing
    * check if specified more than once (see below)
  * specifying more than one buildtype flag
  * `-n` (set pkgname) without using SRCURL as last argument
  * `-s` (sourceforge source) without using PKGNAME as last argument
* Typo fix: exist -> exists
* `usage()`:
  * always print PKGNAME and PKGDESC (instead of NAME and DESC,
    NAME was used in one place and PKGNAME in another)
  * link to <https://spdx.org/licenses/>
  * `-m` (meson) flag was missing in short usage line at the top
  * indicate that the buildtypes are exclusive
  * `-c` flag: remove "to new directory" wording to make the
    message shorter (this should be obvious)
  * remove empty line at the end

NOTE: Before this commit, the `PKGNAME[-PKGVER] | SRCURL` was allowed
to be specified more than once, and the code looped over the arguments.
But this was not documented in `usage()` and had unexpected results:

```
$ newapkbuild first second third
$ tree
.
___ first
    ___ APKBUILD
    ___ first
    ___ ___ APKBUILD
    ___ ___ first
    ___ ___ ___ APKBUILD
    ___ ___ ___ src
    ___ ___ src
    ___ src
```

scanelf may pick up tempfiles created by strip or setfattr since it runs
in spearate process and pipes the out to a subshell. This causes a race
and may lead to the while loop attempt to strip seomthing that no longer
exists.

We fix that by test if file exists before try manipulate it. We could
have written he file list to a temp file first, but this way we benefit
from multiple cores working in parallel.

unpack will no longer unpack without a checksum, even with -f.  This
means that newapkbuild will not be able to deduce what kind of build
system is contained within, so the templates for CMake, Perl, etc are
never used.

This patch ensures checksumming is done right after fetch, so that
unpack works properly.

abuild package misleadingly stated that it will create packages in
$REPODIR (suggesting that it actually creates apks).

`git describe` by default looks for tags, but `git clone` does not clone
tags by default which causes failures on travis currently.

Also redirect `git describe` errors to /dev/null while being here.