Conversation
|
At least As if |
|
Filed https://savannah.gnu.org/support/index.php?110990 to ask if |
|
Ended up being at least
|
trofi
force-pushed
the
autoconf-update
branch
from
c20326e to
2cfa666
Compare
github-actions
bot
added
the
6.topic: python
|
I think I fixed all the breakages I could find. Ready for review. We should be able to flip back individual packages to Build-tested successfully quite a few packages against Full package list
|
Some package will require significant rework to support autoconf-2.72. Example are `firebird` and `python3Packages.awslambdaric`. Let's leave them on previous version until it's sorted upstream.
Without the change `python3Packages.awslambdaric` fails the build
against `autoconf-2.72` as:
python3.11-awslambdaric> checking curl version... 7.83.1
python3.11-awslambdaric> ./configure: line 6673: syntax error near unexpected token `;;'
python3.11-awslambdaric> ./configure: line 6673: ` ;;'
`python3Packages.awslambdaric` embeds tarballs of external dependencies
and is a bit tricky to patch. Let's pin it to a working version.
Without the change `firebird` builds fail against `autconf-2.72` as:
checking for cc_r... gcc
./configure: line 5894: syntax error near unexpected token `;;'
./configure: line 5894: `printf "%s\n" "$as_me: WARNING: --with-system-editline specified, not found. Using bundled editline" >&2;} ;;'
Upstream switched to `cmake` for new releases. Let's pin older ones to
`autoconf-2.71`.
trofi
force-pushed
the
autoconf-update
branch
from
2cfa666 to
62abc91
Compare
| @@ -0,0 +1,69 @@ | |||
| { lib, stdenv, fetchurl, m4, perl, texinfo }: | |||
There was a problem hiding this comment.
Should we take this opportunity to add a common.nix file and share some of this code across versions?
There was a problem hiding this comment.
I personally find it slightly easier to reason about the packages when they don't reuse parts of their implementation. A few pros from the top of my head:
- deletion becomes just file (or directory) deletion
- package update touches a single version, makes it easier to reason about the impact for packages that are stdenv dependencies
- change history in
default.nixis linear and does not get cluttered by occasional backports to older versions - ofborg gets sad to assign reviewers until
posattributes (and extra pending fixes: nix: pointnix editandofborgat the originalversionattribute #276892 ) are applied
It feels to me like autoconf does not contain much code to share and does not expect much churn in previous versions.
But if deduplication is a strong requirement I can try to reshuffle things a bit.
There was a problem hiding this comment.
I mean we can keep it like this, no strong feelings on my side.
ofborg
bot
added
10.rebuild-darwin-stdenv
3 participants
Changes: https://lists.gnu.org/archive/html/info-gnu/2023-12/msg00002.html
Description of changes
Things done
nix.conf? (See Nix manual)sandbox = relaxedsandbox = truenix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage./result/bin/)Add a 👍 reaction to pull requests you find important.