On 7/15/2012 20:14, Greg Troxel wrote:
The patches are just allowing it to build on Dragonfly, changes are obvious. The only remarkable thing is that MAXNAMELEN isn't defined on DragonFly. I cheated by putting a MAXNAMELEN definition in config.h, otherwise I would have had to patch several other files that referenced it. It isn't obvious that they don't break anything else. Please add comments and feed upstream. As the maintainer, that's what I would have asked (that you follow the pkgsrc norms) had you asked me before making changes.
I don't even have time to push patches upstream for packages that I maintain and care about (FreeBSD and pkgsrc) -- see GCC where I have several dozen patches that need to be submitted.
My POV: the package maintainer should be responsible for that kind of interaction as they would know who / how / etc. Unfortunately, it appears that the majority of pkgsrc packages don't even have a maintainer, so that approach can't work. (The fact of maintainerless-packages is a separate issue, let's not explore further).
Let's just say this then:1) I've seen several postings about comments on patches, both to me and others. I'm aware patches *can* have comments. I find them burdensome, so I have only been maintaining comments where they already exist or when I truly feel a comment is warranted. I don't think it's a "norm" that 100% of patches are commented. That might be the desire, but I'd vote against it if asked.
2) Ideally I have nothing against trying to push patches upstream. I just can't do it for the sheer number of packages that need fixing, not enough time. I certainly am not inclined to do it for a package that is very old compared to it's latest upstream release (I tried once, everything was obsolete including the dev website). So I guess I'd do it for packages I care about.
I don't care about this one. I spent more than an hour fixing it already, and I didn't even check to see if the package was maintained. It was, and if the commit bothers you, I'll just revert it.
John