Re: [gentoo-dev] Re: Lastrite: media-gfx/pngcrush

[Rich Freeman]

On Tue, Oct 11, 2011 at 12:52 PM, Markos Chandras <hwoarang@g.o> wrote:
> Seems like none of you ever bothered to read the bug about pngcrush
> and what was discussed there.

I read the entire discussion before making a single post - it would be
irresponsible not to.  Now, I can't say that I checked the cvs
histories of the package metadata.xml file, the herd memberships, or
what somebody sent to somebody else in IRC.

If somebody had posted in the bug a month ago "FYI - I plan to mask
this package on $date so if you have a problem with that let me know,"
I doubt we'd be having this conversation at all.

> It is getting a little bit of a habit to escalate minor problems to flames in Gentoo.

Obviously Matt was unhappy with how things were handled.  It would be
best if he tried to work that out in private first, and perhaps
efforts were made to do this, and perhaps not.

However, I think the reason everybody and their uncle is posting here
is that there doesn't seem to be any acknowledgment that something
non-ideal was done here in the first place.  That just means that this
whole thing could happen again.

Arguing over who did what is counterproductive in my mind - this isn't
a forum to seek justice.  However, talking about how the process
SHOULD work is productive, and I'd just ask that in the future that
people publicly disclose in advance what they plan to do when it
involves touching packages that others maintain.  If somebody can't
come to an agreement with the maintainer on how something should be
handled the solution is to escalate within the project(s) first as
appropriate, but not to just go around doing cvs commits.

I'm fine with having a bias for action, and assuming no response in
two weeks means somebody is OK with something.  However, logging a bug
reporting an issue isn't really the same as asking for consent to mask
a package.  If a package has a critical security problem then we may
not be able to wait, but nobody is going to die if a libpng
stabilization is delayed by a week or two, and if this were started a
week or two earlier there would be no debate here.  Or, at least if
there were a debate there would be less of an aggrieved feeling.

In any case, if you don't want to see flame-wars on -dev I'd recommend
starting by not doing things that are likely to tick people off - like
committing to somebody else's package without ensuring they know about
it.  It sounds like there was an irc ping in this case, but I don't
consider those a very reliable form of communication - especially if
it is not acknowledged.

Rich