Long story short, I unbundled libblas from dev-libs/linear without talking to the maintainer. This lead to a small bug because of quoting problem which was promptly fixed and even extendend to another quoting fix in the ebuild. Hours later the maintainer revert my fix, which certainly enhances QA, and also mgornys later enhancement with Commit message: "Undo blas crap" My wrong doing if you like to call it so, was Didn't talk to the maintainer, before enhancing his last commit. But that doesn't justify naming my work to be "crap" forever in the commit history. Also it is breaking at least common sense in how we should structure commit messages. And secondly he is putting QA violations directly back into place although already fixed. Neglecting the QA issues, which should be already enough, I am asking because of the commit message for a formal suspension of his commit rights.
I am not asking as a comrel member.
This is also very inappropriate commit message in a public medium. I'd give +1 for 48h suspension with my QA hat but I guess I ended up being directly involved.
So yesterday I got people complaining about the -r1 you committed without consent. The most verbal complaint wasn't about the build failure you caused, but about a minimalistic linear algebra library suddenly pulling in a huge BLAS implementation it was designed to replace. If you had read the shipped documentation, you would have known that the build system *allows* you to link against a *different* BLAS if you happen to have one installed. What you did in -r1 was *forcing* that linkage. Also, you didn't discuss this at all. Maybe you were thinking with your QA hat after all? As for "crap": I've seen a lot of discussion (in the context of this project) about how appropriate that word is, and general agreement says you shouldn't feel offended. I could understand that you feel offended because your work was annulled, though.
OK, with bug #561874 closed as CANTFIX by the reporter, let's continue. The original story was: 1) I committed a QA violation by bundling a library in liblinear. 2) jlec corrected that QA violation but committed another one. 3) The other QA violation was corrected. 4) I undid all of the work in -r1, calling it "crap". 5) I was called out for violating some QA policy of not undoing work by QA people and for calling that work "crap". But now it looks like this: 1) jlec committed an ebuild that purportedly unbundled a library in liblinear. 2) People yelled at me all day for needlessly and forcibly pulling in virtual/blas and because jlec's ebuild caused bug #561826. 3) mgorny fixed bug #561826. 4) I removed the -r1 ebuild because it was completely unnecessary, calling it "crap". 5) jlec files bug #561874 and then this bug. 6) bug #561874 ends up closed as CANTFIX by the reporter.
I am pulling back my request. Sorry for the noise.
