The previous implementation of plugin-support for the kakoune derivation
was based on generating, at build time, a `plugins.kak` file that would
source all .kak files in the list of plugins, and wrap the `kak` binary
in a script that would add some command-line arguments so that this
file gets loaded on start-up. The main problem with this approach
is that the plugins' code get executed *after* the user's configuration
file is loaded, so effectively one cannot automatically activate/configure
these plugins.
The idiomatic way of loading plugins is ensuring they end up installed
somwhere under `share/kak/autoload`. Because plugins are already being
packaged to have their code in `share/kak/autoload/plugins/<name-of-plugin>`,
we can obtain a derivation that includes the plugins simply by doing a
`symlinkJoin` of `kakoune-unwrapped` and all the requested plugins.
For this to work, we need to fix two issues:
1. By default, kakoune makes `share/kak/autoload` a symbolic link to
`share/kak/rc`, which contains all builtin definitions. We need
to patch this to put the symlink under `share/kak/autoload/rc`, so that
the join works.
2. By default kakoune expects the `autoload` directory to be in
`../share/kak/autoload` relative to the location of the `kak` binary.
We need to set the `KAKOUNE_RUNTIME` to point the symlinked
share/kak for this to work.
Use `buildInputs` instead of `propagatedBuildInputs` - should reduce
closure size a bit. Don't wrap executables twice. Move `kdoctools` to
native. Use libsForQt5 - default version (works).
Co-authored-by: Frederik Rietdijk <fridh@fridh.nl>
Currently lxqt is a desktop environment that's compiled against qt514.
To avoid possible issues (#101369), we (hopefully) use the same qt
version as the desktop environment at hand. LXQT should move to qt515,
and for the long term the correct qt version should be inherited by the
sddm module.
Intro:
Part of #101369: Every attribute from kdeApplications and kdeFrameworks
can be built with a few different qt5 versions. It's hard to tell the
difference between an application and a library and some applications
rely on inputs from kdeApplications and libsForQt5 alike.
Before this change, some applications that were defined with
`libsForQt5.callPackage` used libraries from the kde* sets compiled with
a specific qt5 version,
Due to `inherit (kde*) <lib or app>;` used in the widest scope, we had
issues with packages that depended on packages defined by this
`inherit`. This led to mismatched qt versions used in the same inputs,
or the inputs of inputs etc.
Hence, we added to all libsForQt5* sets, packages that will be used from
the correct libsForQt5 set, in accordance to the
`libsForQt5*.callPackage` used. All `inherit (kdeApplications) <pkgs>`
and similar inheritance was moved out of all-packages.nix to aliases.nix
only for backwards compatibility.
Now some KDE applications show up in the attribute sets `libsForQt5*`
which didn't show up there previously. This is sort of misleading, as
these are not necessary libraries, but they show up in the wider scope
thanks to them in aliases.nix. Hence it's the best that can be done
considering the circumstances and the urgency of the issue.
Since it's a qt library, we can't guarantee every package using it as a
dependency will need it compiled with the same qt version, we put it in
all `libsForQt5*` scopes (#101369).
Part of #101369: In order to avoid packages using the default `kdesu`
always built with qt515, we put it in scope only for packages defined
with a `libsForQt5`, to avoid incompatible qt versions used together in
inputs of a package.
Fixes#100707 - Since KDE's maintainers don't allow non official kde
applications under the `kdeApplications` path, many packages that depend
on KDE libraries and applications are defined in all-packages.nix.
Overriding all kdeApplications to use qt514 while all other
`libsForQt5.callPackage` in all-packages.nix are using qt515, causes
collisions.
Update pkgs/tools/networking/bacnet-stack/default.nix
Co-authored-by: Daniel Løvbrøtte Olsen <daniel.olsen99+GitHub@gmail.com>
bacnet-stack: Add maintainer and use the original GitHub repo
bacnet-stack: Alphabetize
Update pkgs/top-level/all-packages.nix
Co-authored-by: Jon <jonringer@users.noreply.github.com>
Current nixpkgs always wraps neovim with the "-u" which has sideeffects as explained in https://github.com/NixOS/nixpkgs/issues/55376 :
1. vim won't set the variable $MYVIMRC as explained #34215
2. vim skips loading folder-specific .vimrc / .nvimrc
I wanted to provide a way for users to better control what flags are used to wrap neovim. This is achived by introducing wrapNeovimUnstable et neovimUtils, utilities to help with that. We provide a compatibility layer so that wrapNeovim still works and to let us experiment with wrapNeovimUnstable to better control neovim configuration, plugin dependencies, haskell environment etc so that it becomes easier to generate per-project neovim config.
With this commit, it's possible for instance for home-manager to wrap neovim without the `-u` and just write the config in the
expected $XDG_CONFIG_HOME/nvim/init.vim .
Expect wrapNeovimUnstable interface to evolve in the upcoming months.
This is a mostly cosmetical commit, in the sense it doesn't change the contents
of any package, but reorganizes the overall Nixpkgs expressions.
Terminal emulators are an ubiquitous tool for any Unix user; even the beginners
are routinely familiarized to it. And, manifestly, there are many
implementations of terminal emulators out there, from those traditionally made
in C and C++ to those written in Haskell and Go.
Terminal emulators deserve more highlight. This commit does that by creating a
category for them.
I made a mistake merge. Reverting it in c778945806 undid the state
on master, but now I realize it crippled the git merge mechanism.
As the merge contained a mix of commits from `master..staging-next`
and other commits from `staging-next..staging`, it got the
`staging-next` branch into a state that was difficult to recover.
I reconstructed the "desired" state of staging-next tree by:
- checking out the last commit of the problematic range: 4effe769e2
- `git rebase -i --preserve-merges a8a018ddc0` - dropping the mistaken
merge commit and its revert from that range (while keeping
reapplication from 4effe769e2)
- merging the last unaffected staging-next commit (803ca85c20)
- fortunately no other commits have been pushed to staging-next yet
- applying a diff on staging-next to get it into that state
This should NOT be backported to 20.09!
When 21.03 is released, the DB changes are about a year old and
operators had two release cycles for the upgrade. At this point it
should be fair to remove the compat layer to reduce the complexity of
the module itself.
