staging-next 2024-10-15 #348827

Applied from pull request 343905.

https://gitlab.com/smedius/desktop-icons-ng/-/commit/9079ac663bc33421a489dab2db231f990105bdd5 This unbreaks nixosTests.gnome-extensions. Applied from pull request 343905.

glib now depends on linuxPackages.systemtap.override { withStap = false; }.

…ction 1.82 https://hydra.nixos.org/build/274399381/nixlog/1

https://gitlab.gnome.org/GNOME/gupnp/-/compare/gupnp-1.6.6...gupnp-1.6.7 Changelog-reviewed-by: Bobby Rong <[email protected]>

https://hydra.nixos.org/build/274401072/nixlog/1 Ugh, I am looking at glib-2.0.pc and sysprof-capture-4 is in Requires.private so I didn't propagate that. Later I decide I will just follow 81e445e adding libsysprof-capture next to every usages of pcre2. Not sure why gtk2 is still there after b95976f, same question for pcre.

https://gitlab.gnome.org/GNOME/glib/-/merge_requests/4016

This fixes https://hydra.nixos.org/build/274421427/nixlog/1.

Hopefully actually allow using libadwaita for X11 window decorations.

https://gitlab.gnome.org/GNOME/glib/-/merge_requests/4016

As requested by Sandro.

Hopefully helps https://hydra.nixos.org/build/274566596/nixlog/2, I am not sure.

https://gitlab.gnome.org/GNOME/gnome-applets/-/compare/3.52.0...3.54.0 Changelog-reviewed-by: Bobby Rong <[email protected]>

https://gitlab.gnome.org/GNOME/gnome-bluetooth/-/compare/47.0...47.1 Changelog-reviewed-by: Bobby Rong <[email protected]>

https://gitlab.gnome.org/GNOME/gnome-flashback/-/compare/3.52.1...3.54.0 Changelog-reviewed-by: Bobby Rong <[email protected]>

https://gitlab.gnome.org/GNOME/gnome-initial-setup/-/compare/47.0...47.1 Changelog-reviewed-by: Bobby Rong <[email protected]>

https://gitlab.gnome.org/GNOME/gnome-panel/-/compare/3.52.0...3.54.0 Changelog-reviewed-by: Bobby Rong <[email protected]>

https://gitlab.gnome.org/GNOME/gnome-software/-/compare/47.0...47.1 Changelog-reviewed-by: Bobby Rong <[email protected]>

https://gitlab.gnome.org/GNOME/metacity/-/compare/3.52.0...3.54.0 Changelog-reviewed-by: Bobby Rong <[email protected]>

https://gitlab.gnome.org/GNOME/gtksourceview/-/compare/5.14.0...5.14.1 Changelog-reviewed-by: Bobby Rong <[email protected]>

https://gitlab.gnome.org/GNOME/gtranslator/-/compare/46.1...47.0 Changelog-reviewed-by: Bobby Rong <[email protected]>

https://gitlab.gnome.org/GNOME/libwnck/-/compare/43.0...43.1 Changelog-reviewed-by: Bobby Rong <[email protected]>

https://gitlab.gnome.org/GNOME/libspelling/-/compare/0.4.1...0.4.2 Changelog-reviewed-by: Bobby Rong <[email protected]>

https://hydra.nixos.org/build/274565015/nixlog/1 `pcre` is no longer needed as the comment longer applies. `glib-2.0.pc` now has `libsysprof-capture-4` in `Requires.private` field. Normally, this does not matter since we patch our `pkg-config` to ignore that outside of static builds but `entangle` was using `pkgconf`, which is not patched. Let’s switch to standard `pkg-config`. This will also allow us to clean some more dependencies. Unfortunately, this also reveals issue with an underspecified `gio-unix-2.0` dependency so we need to fix that.

Enabled in most distros and required by valent.

andyholmes/valent@v1.0.0.alpha.45...v1.0.0.alpha.46 Fixes https://hydra.nixos.org/build/274656744/nixlog/1.

https://gitlab.gnome.org/GNOME/gtk/-/compare/4.16.2...4.16.3 Changelog-reviewed-by: Bobby Rong <[email protected]>

Fixes ccec93c.

https://hydra.nixos.org/build/274565679/nixlog/2 beeware/gbulb@bd925dc 5d125ad takes care the only revdep of this package.

This may have been relevant 10 years ago ([DOUBT]), but it sure as hell isn't now, and recursive symlinks are evil. May it burn in hell, as that is where it belongs.

Previously, logind functionality would be disabled if systemd was not found. Since 1.2.5, PipeWire also supports building with elogind. To minimize variance between systemd enabled and not, let's use that for non-systemd builds.

Diff: python-poetry/poetry-core@1.9.0...1.9.1 Changelog: https://github.com/python-poetry/poetry-core/blob/1.9.1/CHANGELOG.md

This commit partially reverts ccec93c. This will break a lot of packages in haskellPackages (due to a missing webkitgtk), but at least hackage-packages.nix is reproducible with the generation shell script again. Unfortunately, it seems the generated nature of hackage-packages.nix was ignored when preparing the webkitgtk change: #345611 (comment)

Unfortunately, this is the newest GHC revision we can update to. After https://gitlab.haskell.org/ghc/ghc/-/commit/145a6477854d4003a07573d5e7ffa0c9a64ae29c, it becomes necessary to bootstrap with an in tree Cabal version. Supporting this in nixpkgs is a waste of time since it wouldn't be necessary for any release version.

Patch is upstream. Fixes: ee5127c ("iproute2: 6.10.0 -> 6.11.0")

https://github.com/MagicStack/uvloop/releases/tag/v0.21.0

While the Swift wrapper should take care of this, not everything that links these libraries goes through the wrapper.

…ift overlay" This reverts commit 8641252.

Without the change `gobject-introspection` fails the build on `staging-next` as: applying patch /nix/store/5w1icwq865fhjs9w2s0vb6xqn5irzxv7-setuptools-74.0-compat.patch patching file giscanner/ccompiler.py Reversed (or previously applied) patch detected! Assume -R? [n] Apply anyway? [n] Skipping patch. 4 out of 4 hunks ignored -- saving rejects to file giscanner/ccompiler.py.rej patching file giscanner/msvccompiler.py Reversed (or previously applied) patch detected! Assume -R? [n] Apply anyway? [n] Skipping patch. 1 out of 1 hunk ignored -- saving rejects to file giscanner/msvccompiler.py.rej

Without the change `dmenu` fails to build on `staging-next` as: $ nix build --no-link -f. dmenu -L ... dmenu> build flags: SHELL=/nix/store/mm0pa3z7kk6jh1i9rkxqxjqmd8h1qpxf-bash-5.2p37/bin/bash CC:=\$$CC$ dmenu> cp config.def.h config.h dmenu> gcc -c -std=c99 -pedantic -Wall -Os -I/usr/X11R6/include -I/usr/include/freetype2 -D_DEFAULT_SOURCE -D_BSD_SOURCE -D_XOPEN_SOURCE=700 -D_POSIX_C_SOURCE=200809L -DVERSION=\"5.3\" -DXINERAMA dmenu.c dmenu> In file included from dmenu.c:17: dmenu> /nix/store/2rxphcg0qabc3a8c4lvy610sm03bh72y-libXft-2.3.8-dev/include/X11/Xft/Xft.h:40:10: fatal error: ft2build.h: No such file or directory dmenu> 40 | #include <ft2build.h> dmenu> | ^~~~~~~~~~~~ dmenu> compilation terminated. dmenu> make: *** [Makefile:12: dmenu.o] Error 1 The change uses `pkg-config` to discover library dependencies and lib paths.

currently fails with: ``` error: builder for '/nix/store/7g13ladvphijvrhswzl6zdwx373ks6ac-rust-cbindgen-0.27.0.drv' failed with exit code 101; last 25 log lines: > left: ["release", "x86_64-unknown-linux-gnu"] > right: ["release"] > note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace > > ---- bin_explicit_debug_build stdout ---- > thread 'bin_explicit_debug_build' panicked at tests/profile.rs:93:5: > assertion `left == right` failed > left: ["debug", "x86_64-unknown-linux-gnu"] > right: ["debug"] > > ---- bin_default_uses_debug_build stdout ---- > thread 'bin_default_uses_debug_build' panicked at tests/profile.rs:87:5: > assertion `left == right` failed > left: ["debug", "x86_64-unknown-linux-gnu"] > right: ["debug"] > > > failures: > bin_default_uses_debug_build > bin_explicit_debug_build > bin_explicit_release_build > > test result: FAILED. 0 passed; 3 failed; 0 ignored; 0 measured; 3 filtered out; finished in 0.42s > > error: test failed, to rerun pass `--test profile` For full logs, run 'nix log /nix/store/7g13ladvphijvrhswzl6zdwx373ks6ac-rust-cbindgen-0.27.0.drv'. ```` Related to: #348031

#337635

Upstream do not plan to support this version (see <#347484 (comment)>), so we should not package a version that will surely accumulate CVEs from V8 etc. in 24.11. As this package was only added yesterday, I don’t think there’s any need for a compatibility alias.

It’s not clear to me what SDK version this actually ideally wants, but `apple-sdk_11` is a conservative choice for now.

... specifically, it's the `CARGO_BUILD_TARGET` env var introduced in commit: 683f97e

Backports an upstream patch.

Backports an upstream patch to not require /etc/os-release at build time.

Without the change the build of `xorg.libXfont` fails as: > xtrans-1.5.1/include/X11/Xtrans/Xtrans.c: In function '_FontTransSelectTransport': > xtrans-1.5.1/include/X11/Xtrans/Xtrans.c:165:10: error: 'i' undeclared (first use in this function) > 165 | for (i = 0; i < PROTOBUFSIZE && protobuf[i] != '\0'; i++) > | ^

…-> 0.32.2 (#348852)

Diff: https://gitlab.freedesktop.org/mesa/mesa/-/compare/mesa-24.2.4...mesa-24.2.5 Changelog: https://docs.mesa3d.org/relnotes/24.2.5.html

https://gitlab.gnome.org/GNOME/epiphany/-/compare/47.0...47.1

https://gitlab.gnome.org/GNOME/gnome-maps/-/compare/v47.0...v47.1

https://gitlab.gnome.org/GNOME/rygel/-/compare/rygel-0.44.0...rygel-0.44.1

https://gitlab.gnome.org/GNOME/libadwaita/-/compare/1.6.0...1.6.1

This reverts commit 6f7657b. Causes a libjxl test to fail. Bisected to openexr commit e4e6cf2 ("Initial conversion of input file to use core"), which doesn't look like an intentional behavior change, so let's revert this until that's figured out. I've reported it to libjxl, and will report it to openexr once they confirm it's not a problem on their end. Link: libjxl/libjxl#3901

This would have caught the regression introduced by 6f7657b ("openexr_3: 3.2.4 -> 3.3.0"), and reverted in the parent commit.

Without the change the build fails as: $ nix build --impure --expr 'with import ./. {}; procps.override { withSystemd = false; watchOnly = true; }' -L ... procps> build flags: -j16 SHELL=.../bash usrbin_execdir=\$$out$/bin watch PKG_LDFLAGS= procps> make: *** No rule to make target 'watch'. Stop. While at it dropped unused `PKG_LDFLAGS=` flag.

This works since <#346950>.

…349274)

…sthru.tests.static` on `x86_64-darwin` (#348023)

The upstream issue for liquidhaskell is resolved. liquidhaskell is marked broken for an unrelated issue in one of its dependencies.

haskellPackages.brick_0_70_1: remove at 0.70.1 taskell's upstream repository has been archived after three years of inactivity. There also doesn't seem to be an obvious fork that provides a suitable, maintained replacement. No changelog entry since taskell has already been marked as broken in 24.05 from the beginning.

Seems like Stackage recently has reintroduced ghc-syntax-highlighter to its set at version 0.0.10.0 which is just what we need.

Stackage already specifies == 0.5.*

These were used by the now removed pkgs.petrinizer.

stylish-cabal does no longer build for GHC 8.6.5 (binary) for unrelated reasons. Doesn't really make sense to invest further into it.

https://gitlab.gnome.org/GNOME/snapshot/-/compare/47.0.1...47.1

Some library dependency must have stopped re-exporting 'void', so now it needs an extra import line. This package doesn't have any VCS repo.

I'm not sure why this needs a jailbreak. There is already a hackage revision to fix it: https://hackage.haskell.org/package/pipes-extras-1.0.15/revisions/

This commit has been generated by maintainers/scripts/haskell/regenerate-hackage-packages.sh

Exclude tests that broke with Sphinx 7.4.x. The upstream project is sadly not very well maintained these days due to funding reasons.

Backport a patch (which probably addresses a completely unrelated issue with hadrian and system libffi for GHC >= 9.2.3) that stops Cabal from handling libffi includes in libraries/ghci. This allows for cc-wrapper to freely pick the correct libffi which is important when cross compiling GHC as we need to correctly pick between build and host/target libffi each time. This fixes the build of pkgsCross.riscv64.haskell.compiler.ghc8107. Tested - haskell.packages.ghc8107.cabal2nix (x86_64-linux, aarch64-linux) - pkgsCross.riscv64.haskell.compiler.ghc8107 (x86_64-linux) Could not test darwin because rcodesign is broken at the moment (?).

This reverts commit f12c5fd. The tests were broken because paramiko couldn't be imported, which was fixed by 50d865a.

This reverts commit c3322f4. Reason for revert: extensions 0.1.0.3 doesn't support Cabal == 3.12.*. Dependency conflicts due to extensions seem annoying to deal with.

…mesa update package.yaml

Without the change the eval fails on `staging-next` as: $ nix build --no-link -f. python3Packages.dbt-adapters error: … while evaluating a branch condition at pkgs/development/interpreters/python/passthrufun.nix:27:9: 26| func = name: value: 27| if lib.isDerivation value then | ^ 28| lib.extendDerivation (valid value || throw "${name} should use `buildPythonPackage` or `toPythonModule` if it is to be part of the Python packages set.") {} value … while evaluating a branch condition at /home/slyfox/dev/git/nixpkgs-staging-next/lib/customisation.nix:263:8: 262| 263| in if missingArgs == {} | ^ 264| then makeOverridable f allArgs (stack trace truncated; use '--show-trace' to show the full, detailed trace) error: attribute 'pythonRelaxDeps' already defined at pkgs/development/python-modules/dbt-adapters/default.nix:30:3 at pkgs/development/python-modules/dbt-adapters/default.nix:34:3: 33| 34| pythonRelaxDeps = [ "protobuf" ]; | ^ 35|

Neither package support GHC 9.10 currently. Their corresponding plugins are thus disabled in HLS.

https://www.haskell.org/ghc/download_ghc_9_8_3.html

…mesa (#349186)

This reverts commit bf21377. This wasn't a big rebuild on nixpkgs master, but now on staging-next it would rebuild *-darwin stdenv, which would be too expensive at this point.

HLS can't be built yet with 9.8.3, so it doesn't make sense to make it a default yet. Unfortunately we are still missing a ghc-lib-parser release that works with ghc 9.8.3.

This commit has been generated by maintainers/scripts/haskell/mark-broken.sh based on *evaluation [1809515](https://hydra.nixos.org/eval/1809515) of nixpkgs commit [e7735af](https://github.com/NixOS/nixpkgs/commits/e7735afbdce5f32dabb78e9393792625ff6d5eaa) as of 2024-10-21 19:40 UTC* from the haskell-updates jobset on hydra under https://hydra.nixos.org/jobset/nixpkgs/haskell-updates

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

staging-next 2024-10-15 #348827

staging-next 2024-10-15 #348827

Commits on Oct 14, 2024

Commits on Oct 15, 2024

Commits on Oct 16, 2024

Commits on Oct 17, 2024

Commits on Oct 18, 2024

Commits on Oct 19, 2024

Commits on Oct 20, 2024

Commits on Oct 21, 2024

Commits on Oct 22, 2024

staging-next 2024-10-15 #348827

Are you sure you want to change the base?

staging-next 2024-10-15 #348827

Commits on Oct 14, 2024

Commits on Oct 15, 2024

Commits on Oct 16, 2024

Commits on Oct 17, 2024

Commits on Oct 18, 2024

Commits on Oct 19, 2024

Commits on Oct 20, 2024

Commits on Oct 21, 2024

Commits on Oct 22, 2024