Conversation
|
Can we enable libdeflate support?
|
|
Good idea! Done. I have noticed that none of the propagatedBuildInputs need to be propagated (their headers are not included by installed libtiff headers, and libtiff does not install static libraries), but unpropagating them may reduce the set of libraries available to dependent projects and possible break their build, so I have not moved them to buildInputs in this PR. |
There was a problem hiding this comment.
Please undo this change. Having the inputs one on each line is better for diff and merges.
There was a problem hiding this comment.
I undid this change because it reduces the diff in this PR, but having args on one line is not better for review since GitHub displays line edits well (highlighting the changed part), and has no meaningful impact on merges since conflicts are rare and are not likely caused just by the changed args. The formatting of the args affects readability, and for me (and I think in general) one arg per line makes it worse by drawing attention from the important parts of the definition to the trivial (and duplicated in its body).
There was a problem hiding this comment.
and has no meaningful impact on merges since conflicts are rare and are not likely caused just by the changed args.
Often merging the master and staging branches I have a different experience here. Its very annoying having to check the lists of arguments when there are many arguments. Here there aren't many, but solving merge conflicts when you have a whole block of arguments is not fun.
|
Fixes two security vulnerabilities in libtiff before 4.2.0. We therefore need to backport those fixes.
|
|
Fixes backported in #116280 |
Motivation for this change
Regular update. Changelog: http://www.simplesystems.org/libtiff/v4.2.0.html
Things done
sandboxinnix.confon non-NixOS linux)nix-shell -p nixpkgs-review --run "nixpkgs-review wip"./result/bin/)nix path-info -Sbefore and after)