Fixes the following warning with nixos-unstable (and upcoming 25.11)
evaluation warning: The 'pie' hardening flag has been removed in
favor of enabling PIE by default in compilers and should no longer
be used. PIE can be disabled with the -no-pie compiler flag, but
this is usually not necessary as most build systems pass this if
needed. Usage of the 'pie' hardening flag will become an error in
future.
Signed-off-by: Sefa Eyeoglu <contact@scrumplex.net>
`qrcodegen` isn't available in most repositories, package discovery
scripts for it are maintained in a different repository, and they are
bugged for Windows at least. This basically forces us into vendoring it,
which isn't cool; libqrencode seems like viable alternative used by many
more apps
Signed-off-by: Seth Flynn <getchoo@tuta.io>
Helps avoid needless rebuilds where only the revision changed. Also
better conforms to Nixpkgs' version standards
Signed-off-by: Seth Flynn <getchoo@tuta.io>
After extensive (5 minutes) of testing, it seems we don't actually come
across any Nix bugs with lib.fileset! (aside from those inherit to
it...but 🤷)
This reverts commit a49a58bc45.
Signed-off-by: Seth Flynn <getchoo@tuta.io>
This reverts commit 8312713dc2.
See nix#9428. Path coercion like `"${./.}"` in flakes causes the path to
be copied to the store twice; using the `self` argument works around
this. However, as `lib.fileset` doesn't support using `self`, so we need
to go back to `nix-filter`
Signed-off-by: seth <getchoo@tuta.io>
overlays can cause issues we don't run into upstream, especially when
used on stable channels for example. preferring the pure `packages`
output also encourages the use of our binary cache by users, who
probably don't like building the launcher all too much
Signed-off-by: seth <getchoo@tuta.io>
The primary goals here include making the flake easier to contribute to
by having it follow the standard boilerplate, while also limiting the
size of our flake.lock to lower the chance of duplicate inputs for users
Signed-off-by: seth <getchoo@tuta.io>
this allows to implictly pass `prismlauncher-unwrapped`
it won't have an effect on our package set as directly inheriting from
the scope we create filters out the extra attributes usually introduced
with a scope
Signed-off-by: seth <getchoo@tuta.io>
