On Thursday 03 February 2005 23:56, Ciaran McCreesh wrote:
(B> On Thu, 3 Feb 2005 23:40:45 +0900 Jason Stubbs <[EMAIL PROTECTED]>
(B>
(B> wrote:
(B> | [13:21] <jstubbs> seem to be expected to have unit tests and
(B> | regressions tests for everything and make sure everything is bug free
(B> | before release
(B>
(B> Yes, I do expect portage releases to be bug free when it comes to
(B> commit-affecting operations. Are you asking us to use untested code when
(B> committing to the live tree?
(B
(BYes. You've been doing it all along anyway.
(B
(B> Would you be upset if we ended up with a few hundred core packages with,
(B> say, broken manifests as a result?
(B
(BYes. That's why I'm forewarning you and asking for testing before putting it
(Binto a release.
(B
(BNote, what you are postulating is crazyness. I said in the first email that
(Ball I wanted was to confirm that there aren't going to be any tracebacks.
(BWhat you're trying to turn this into is about the equivalent of refusing to
(Btest a patch that changes the colour of USE flags for fear that it might
(B"rm -rf /".
(B
(BThe patch does what I said it does in the original email. I have no doubt that
(Bit doesn't do that. What I do doubt is whether I may have mispelled or left
(Bout a variable here or there that leads to a traceback. Nothing more.
(B
(BRegards,
(BJason Stubbs
(B
(B--
([email protected] mailing list
