@peterhoeg

I woopsed that PR, so here is clean one. Already squashed commits.

@GrahamcOfBorg eval

@GrahamcOfBorg eval

Guys, if you can believe it, I with Hoeg already talked and tested it in #69046. Hoeg asked to make cosmetic changes, those changes triggered a rebuild of package, and since that - also a sub-graph, including all versions of ImageMagic which have big numbers of successors.

There is why moved to staging, but while switching - I borked that thread.

P.S.
I promised him to further continue work on this package. Since I took on responsibility of maintaining - I would separate {lib, bin}, by creating the real dvjulibre library package (which does not exist in nixpkgs at the moment), and then go through the dependencies and look which require bins, and which can benefit from lib, and which require both. But that is a relatively big question, so I am separating concerns. I ask to merge this cosmetics first.

The plan was to avoid another big rebuild when the split goes in.

Well, sorry, I perceived the discussion as i wrote.
I said to you my view in that thread already.

The code is clear aslo. I don't want to see how people waste their time trying to evaluate cosmetics that already was evaluated and reviewed.

I want to do atomic PRs. I would not turn completed cosmetic coffee-break devoted PR into recursive yak shaving PR. I also wrote a big hearty letter about why I do not do that, but I put it the table. Life thought me that. But it is also obvious why big PRs should not happen. As soon as work done and reviewed - I would send/accept what is improved, and do another atomic change.

I hope you all have a calm time. It feels that the community is tired. NixPkgs achieved the max 12.7 warp speed, while 10 warp - is a speed on Linux Kernel changes. Thank you.

I wrote a bug-report screen play in the issue tracker, to explain the issue with priority'ies and it's default values. And send-in the description. Also I bug-tracked your issue today. I hope that counts for that I am honest.

There is a mass-rebuild in staging anyway. So it does not matter in the end.