There was a ton of duplication which is not maintainable, so refactor
everything into a single test which still covers the differences.
This in turn revealed some issues in the unstowing logic:
- We shouldn't conflict if we find a file which isn't a link or a
directory; we can just skip over it.
- Unstowing with `--dotfiles` was using the wrong variable to obtain
the package path, and as a result having to perform an unnecessary
call to `adjust_dotfile()`.
So fix those at the same time.
Remove the dependency on the ancient and unmaintained texi2html, which
was difficult to get running on most distros other than openSUSE.
There are two more modern alternative approaches which can replace
this:
- Use texi2any
- Use makeinfo --html --no-split
The latter seems to be the standard way these days, so we switch to
that; however we keep Makefile rules for all three, and a phony
meta-rule 'manual-single-html-all' to allow quick comparison between
them. Make tweaks accordingly to minimise the differences and improve
the output.
The rules for the older two approaches do not get triggered by
default.
Fixes#21: https://github.com/aspiers/stow/issues/21
* Use docker to locally run tests on multiple perl versions
* perlbrew used to install multiple perls
* perl environments bootstrapped when image built
* Based on travis configuration
* Adds docker folder to MANIFEST.SKIP
* Running image runs all tests.
* If exit with no error, then all tests pass.
* ./build-docker.sh builds stow testing image.
* ./test-docker.sh runs stow testing image.
* Assumes built with options in build-docker.sh
* *-docker.sh scripts added to MANIFEST.SKIP
http://article.gmane.org/gmane.comp.sysutils.automake.general/13192
This means that it is still included in the distribution and
installed, but is only automatically (re)built when making a
new distribution - if stow.texi changes during development, it
has to be rebuilt manually via 'make pdf'.
