[PATCH nixpkgs 0/2] Pre-merge fixes for Nixpkgs update
I'm working on merging in a new version of upstream Nixpkgs. These are a couple of small fixes that came up during that work, but can be taken care of first since they make improvements to the current version. Alyssa Ross (2): chromiumOSPackages.modemmanager-next: fix libqmi spectrumPackages.sys-vms.comp: don't pin Emacs .../linux/chromium-os/modem-manager/next.nix | 12 ++++++++++-- pkgs/os-specific/linux/spectrum/vm/comp/default.nix | 4 ++-- 2 files changed, 12 insertions(+), 4 deletions(-) base-commit: 129d329e3bcdcc972454ce284e17da16b28bd401 -- 2.31.1
We need to remove the upstream version, to replace it here with the CrOS version, and also we need to provide the CrOS version as a build input, not a native build input. --- .../linux/chromium-os/modem-manager/next.nix | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) diff --git a/pkgs/os-specific/linux/chromium-os/modem-manager/next.nix b/pkgs/os-specific/linux/chromium-os/modem-manager/next.nix index d40b2c9292e..a2239a0b9c8 100644 --- a/pkgs/os-specific/linux/chromium-os/modem-manager/next.nix +++ b/pkgs/os-specific/linux/chromium-os/modem-manager/next.nix @@ -3,7 +3,12 @@ }: modemmanager.overrideAttrs ( - { pname, nativeBuildInputs ? [], passthru ? {}, meta ? {}, ... }: + { pname + , nativeBuildInputs ? [], buildInputs ? [] + , passthru ? {}, meta ? {} + , ... + }: + { pname = "${pname}-chromiumos-next-unstable"; version = "2019-10-17"; @@ -12,7 +17,10 @@ modemmanager.overrideAttrs ( upstreamInfo.components."src/third_party/modemmanager-next"; nativeBuildInputs = nativeBuildInputs ++ - [ autoreconfHook autoconf-archive libqmi libxslt ]; + [ autoreconfHook autoconf-archive libxslt ]; + + buildInputs = [ libqmi ] + ++ lib.filter (x: x ? pname && x.pname != "libqmi") buildInputs; passthru = passthru // { updateScript = ../update.py; -- 2.31.1
On Wed Jun 30, 2021 at 2:20 AM PDT, Alyssa Ross wrote:
We need to remove the upstream version, to replace it here with the CrOS version, and also we need to provide the CrOS version as a build input, not a native build input. --- .../linux/chromium-os/modem-manager/next.nix | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-)
diff --git a/pkgs/os-specific/linux/chromium-os/modem-manager/next.nix b/pkgs/os-specific/linux/chromium-os/modem-manager/next.nix index d40b2c9292e..a2239a0b9c8 100644 --- a/pkgs/os-specific/linux/chromium-os/modem-manager/next.nix +++ b/pkgs/os-specific/linux/chromium-os/modem-manager/next.nix @@ -3,7 +3,12 @@ }:
modemmanager.overrideAttrs ( - { pname, nativeBuildInputs ? [], passthru ? {}, meta ? {}, ... }: + { pname + , nativeBuildInputs ? [], buildInputs ? [] + , passthru ? {}, meta ? {} + , ... + }: + { pname = "${pname}-chromiumos-next-unstable"; version = "2019-10-17"; @@ -12,7 +17,10 @@ modemmanager.overrideAttrs ( upstreamInfo.components."src/third_party/modemmanager-next";
nativeBuildInputs = nativeBuildInputs ++ - [ autoreconfHook autoconf-archive libqmi libxslt ]; + [ autoreconfHook autoconf-archive libxslt ]; + + buildInputs = [ libqmi ] + ++ lib.filter (x: x ? pname && x.pname != "libqmi") buildInputs;
Is having two "libqmi"s really that big of a deal? Is there a scenario where this prevents e.g. mixing versions or something? Seems like it'd be easier to just `inherit buildInputs` or `buildInputs = buildInputs` (latter for future extensibility).
passthru = passthru // { updateScript = ../update.py; -- 2.31.1
On Wed, Jun 30, 2021 at 08:32:54AM -0700, Cole Helbling wrote:
On Wed Jun 30, 2021 at 2:20 AM PDT, Alyssa Ross wrote:
We need to remove the upstream version, to replace it here with the CrOS version, and also we need to provide the CrOS version as a build input, not a native build input. --- .../linux/chromium-os/modem-manager/next.nix | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-)
diff --git a/pkgs/os-specific/linux/chromium-os/modem-manager/next.nix b/pkgs/os-specific/linux/chromium-os/modem-manager/next.nix index d40b2c9292e..a2239a0b9c8 100644 --- a/pkgs/os-specific/linux/chromium-os/modem-manager/next.nix +++ b/pkgs/os-specific/linux/chromium-os/modem-manager/next.nix @@ -3,7 +3,12 @@ }:
modemmanager.overrideAttrs ( - { pname, nativeBuildInputs ? [], passthru ? {}, meta ? {}, ... }: + { pname + , nativeBuildInputs ? [], buildInputs ? [] + , passthru ? {}, meta ? {} + , ... + }: + { pname = "${pname}-chromiumos-next-unstable"; version = "2019-10-17"; @@ -12,7 +17,10 @@ modemmanager.overrideAttrs ( upstreamInfo.components."src/third_party/modemmanager-next";
nativeBuildInputs = nativeBuildInputs ++ - [ autoreconfHook autoconf-archive libqmi libxslt ]; + [ autoreconfHook autoconf-archive libxslt ]; + + buildInputs = [ libqmi ] + ++ lib.filter (x: x ? pname && x.pname != "libqmi") buildInputs;
Is having two "libqmi"s really that big of a deal? Is there a scenario where this prevents e.g. mixing versions or something? Seems like it'd be easier to just `inherit buildInputs` or `buildInputs = buildInputs` (latter for future extensibility).
It just occurred to me that there's a better way of doing this, that prevents any ambiguity about what's happening with libqmi. What do you think about this instead? diff --git c/pkgs/os-specific/linux/chromium-os/modem-manager/next.nix w/pkgs/os-specific/linux/chromium-os/modem-manager/next.nix index d40b2c9292e..17d95c2b3bc 100644 --- c/pkgs/os-specific/linux/chromium-os/modem-manager/next.nix +++ w/pkgs/os-specific/linux/chromium-os/modem-manager/next.nix @@ -2,7 +2,7 @@ , autoconf-archive, libqmi, libxslt }: -modemmanager.overrideAttrs ( +(modemmanager.override { inherit libqmi; }).overrideAttrs ( { pname, nativeBuildInputs ? [], passthru ? {}, meta ? {}, ... }: { pname = "${pname}-chromiumos-next-unstable"; @@ -12,7 +12,7 @@ modemmanager.overrideAttrs ( upstreamInfo.components."src/third_party/modemmanager-next"; nativeBuildInputs = nativeBuildInputs ++ - [ autoreconfHook autoconf-archive libqmi libxslt ]; + [ autoreconfHook autoconf-archive libxslt ]; passthru = passthru // { updateScript = ../update.py;
--- pkgs/os-specific/linux/spectrum/vm/comp/default.nix | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/pkgs/os-specific/linux/spectrum/vm/comp/default.nix b/pkgs/os-specific/linux/spectrum/vm/comp/default.nix index 988fad5fba6..a2f068b5ec9 100644 --- a/pkgs/os-specific/linux/spectrum/vm/comp/default.nix +++ b/pkgs/os-specific/linux/spectrum/vm/comp/default.nix @@ -1,5 +1,5 @@ { lib, makeRootfs, runCommand, writeScript, writeText -, busybox, emacs26-nox, execline, gcc, linux_vm, s6, sommelier, source-code-pro +, busybox, emacs-nox, execline, gcc, linux_vm, s6, sommelier, source-code-pro , wayfire, westonLite, zsh }: @@ -7,7 +7,7 @@ runCommand "vm-comp" rec { linux = linux_vm; path = [ - busybox emacs26-nox execline gcc s6 sommelier wayfire westonLite zsh + busybox emacs-nox execline gcc s6 sommelier wayfire westonLite zsh ]; login = writeScript "login" '' -- 2.31.1
On Wed Jun 30, 2021 at 2:20 AM PDT, Alyssa Ross wrote:
--- pkgs/os-specific/linux/spectrum/vm/comp/default.nix | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/pkgs/os-specific/linux/spectrum/vm/comp/default.nix b/pkgs/os-specific/linux/spectrum/vm/comp/default.nix index 988fad5fba6..a2f068b5ec9 100644 --- a/pkgs/os-specific/linux/spectrum/vm/comp/default.nix +++ b/pkgs/os-specific/linux/spectrum/vm/comp/default.nix @@ -1,5 +1,5 @@ { lib, makeRootfs, runCommand, writeScript, writeText -, busybox, emacs26-nox, execline, gcc, linux_vm, s6, sommelier, source-code-pro +, busybox, emacs-nox, execline, gcc, linux_vm, s6, sommelier, source-code-pro , wayfire, westonLite, zsh }:
@@ -7,7 +7,7 @@ runCommand "vm-comp" rec { linux = linux_vm;
path = [ - busybox emacs26-nox execline gcc s6 sommelier wayfire westonLite zsh + busybox emacs-nox execline gcc s6 sommelier wayfire westonLite zsh ];
login = writeScript "login" '' -- 2.31.1
Reviewed-by: Cole Helbling
participants (2)
-
Alyssa Ross -
Cole Helbling