This reverts commit b32a057425,
which breaks even the most straightforward uses of srcOnly:
nix-repl> srcOnly guile
error: anonymous function at /home/src/nixpkgs/pkgs/build-support/src-only/default.nix:1:1 called with unexpected argument 'drvPath', at /home/src/nixpkgs/lib/customisation.nix:69:16
nix-repl> srcOnly hello
error: anonymous function at /home/src/nixpkgs/pkgs/build-support/src-only/default.nix:1:1 called with unexpected argument 'drvPath', at /home/src/nixpkgs/lib/customisation.nix:69:16
Link: https://github.com/NixOS/nixpkgs/pull/80903#issuecomment-617172927
The syntax is ${parameter:-word} (i.e. previously this used
"latestTag" instead of the actual value).
(Fixes a regression from #85278.)
Also: Even though getting the latest tag isn't really security critical
(as long as Git itself is secure against untrusted input), I'd prefer to
switch from the Git to the HTTPS protocol (for authentication of the
server and encryption + uses a standard port).
What's happening now is that both cri-o and podman are creating
/etc/containers/policy.json.
By splitting out the creation of configuration files we can make the
podman module leaner & compose better with other container software.
This is less brittle and breaks loud if the code changes.
Also remove the /usr/bin/file patch. It is not really required
for the build to work, the generated warning is harmless.
* Replace LD_LIBRARY_PATH with OS-specific name (e.g. DYLD_LIBRARY_PATH
on macOS).
* Disable Python tests on macOS, because they use gpg, which fails due
to a very long socket path (https://github.com/NixOS/nix/pull/1085).
The former should be fixed upstream. The latter is a Nix-specific issue,
but it can be worked-around upstream by making Python tests respect
--disable-gpg-test.
* pythonPackages.matplotlib: add veprbl to maintainers
* pythonPackages.matplotlib: remove an old darwin hack
* pythonPackages.matplotlib: remove python and stdenv from buildInputs
* python3Packages.matplotlib: remove a hack for python33
* pythonPackages.matplotlib: remove outdated checkPhase
Downloading the baseline_images is not a real issue, building against
older freetype (local_freetype = True) is, perhaps, not what we
want. The good news is that the tests would pass (tested on 3.2.1) if
we were to enable them:
========== 6684 passed, 1332 skipped, 10 xfailed in 228.64s (0:03:48) ==========
* pythonPackages.matplotlib: provide setup.cfg from a file instead of a patch
* python3Packages.matplotlib: 3.1.3 -> 3.2.1
