This ensures that xfce and most others DMs are preferred over
xterm. Previously slim used directory order, which is undefined.
Of course, it's just lucky that xfce < xterm lexicographically, but
that also applies to the other display managers, AFAIK. We should have
a way to specify a DM order, but that can be accomodated by generating
desktop entries like "<NN>-session.desktop".
Fixes#4300. Parenthetical to #12516.
First of all this fixes an evaluation error I introduced in ae466ba,
which wasn't triggered by any of my own tests against the change because
there are usually no NixOS options that are declared outside of the
<nixpkgs> tree. I renamed the attribute name from "fn" to "fileName"
first and later to "fullPath" but forgot one still occuring "filename".
Thanks to @vcunat for noticing this.
Another thing that he pointed out was that the "stripPrefix" function
can be factored away entirely, because it's very similar to
"removePrefix" in <nixpkgs/lib>.
Unfortunately we can't use "removePrefix" as is, because we need to
account for the final shlash.
So instead of removing it twice and/or retaining "stripPrefix", let's
append a shlash on every "prefixesToStrip" and we can use "removePrefix"
as is.
Tested with:
taalo-build nixos/release.nix -A tests.installer.simple.x86_64-linux
And:
w3m -dump "$(
nix-build nixos/release.nix -A manual.x86_64-linux
)/share/doc/nixos/options.html"
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Cc: @vcunat
is extended using patches http://tools.suckless.org/dmenu/patches/ this
argument makes it easier to extend dmenu by doing:
dmenu.overwritte {
patches = [
./dmenu-4.4-follow-focus.diff";
];
};
Poezio doesn't install the files required for running the tests, but
also while building with Nix we don't end up getting valid shared
objects within the temporary build directory.
So we now running "make test" (which does "py.test -v test/") with a
PYTHONPATH that adds the /poezio directory of the path in site-packages
to make sure that the test runner is able to import the shared objects.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Cc: @lancelotsix
Let's use a simple (unflipped) fold and break out the actual core
stripPrefix function from stripAnyPrefixes (I personally love
point-less^H^H^H^Hfree style but if I'd be anal I'd even go further and
factor away the "fn:").
Also, let's use path as a better name for "fn" (filename), because
that's what it is and also cannot be confused with "fn" meaning
"function".
We now toString all of the prefixes, so there shouldn't be any need to
implicily toString the extraSources anymore.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Regression introduced by e6cd147ae7.
This broke all of the installer tests, because they needed to rebuild
the manual within the test machine, while it only has a closure of the
already pre-built system in place.
The problem here was just that the order of the arguments got mixed up
in stripAnyPrefixes, so it was actually trying to strip the path off the
prefix, not the other way around.
So in the end no prefix was stripped at all, so we ended up having full
store paths in the manual, which in turn caused the build within the VM
to fail, because the prefixes differed.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
