In recent versions of chipsec the build scripts have changed and we
haven't built the kernel module in nixpkgs. Upstream has changed the
variable for the kernel sources.
Additional patches are needed as a workaround. Those issues should be
fixed upstream.
pandoc-citeproc is an optional dependency of apostrophe's dependency
pypandoc. However apostrophe doesn't use pandoc-citeproc as a filter
(anymore?), so we can safely remove this dependency which invites
breakage as the package has been deprecated by upstream.
* Translate all seds in postPatch into patches (for setting the static
path and skipping the test that needs network access)
* The patch for the changed pandoc heading generation was simplified:
Since we know our pandoc version is always that new, we can skip the
version check.
* Skip the test for pandoc-citeproc: pandoc-citeproc has been deprecated
in favor of pandoc --citeproc by the upstream pandoc developer.
pypandoc's testsuite doesn't reflect this yet (although it should
support --citeproc theoretically) to avoid depending on
pandoc-citeproc for the checkPhase (as we expect it to break again or
continue to be broken) we skip the test requiring pandoc-citeproc.
The breakage of pypandoc due to pandoc-citeproc was pointed out here:
https://github.com/NixOS/nixpkgs/pull/116635#issuecomment-809258707
Thank you!
I just realized that the original repository disappeared, so I decided
switch to a fork[1] for now. Also, I changed the following things:
* The project got relicensed[2] to `gpl3Plus`. Unfortunately the tags are
gone, but it seems as this has happened after the 1.0 release, so no
problem for us.
* The patch for sway 1.5[3] got merged into the fork, so it doesn't have
to be applied here.
* Decided to go for `master` here for now. First of all there are no
tags anymore, also there are a few fixes on `master` now.
* Compared the new source with the old 1.0 tarball via `diff -r`. Even
though most of the changes were written by the original author on top
of 1.0, I did a diff against everything from 1.0 and didn't find any
issues. Also, `luispabon` doesn't seem to be a throw-away account so
I'd say that using this fork for now is fine.
[1] https://github.com/luispabon/wdisplays
[2] 69564522bc
[3] 5198a9c94b
* Enable exactly one backend (Native seems like the safest choice, but
GMP also seems sane, interested to hear opinions on this!)
* Apply patch which fixes a type mismatch issue between Natural.hs and
Natural.hs-boot.
