Re: [Yum-devel] repo priorities

2007-09-12 Thread Florian Festi
seth vidal wrote: On Thu, 2007-09-06 at 15:37 -0400, seth vidal wrote: On Thu, 2007-09-06 at 08:26 +0200, Tim Lauridsen wrote: +1, for inclusion in core. something like cost=5 in the repo file. the default value for cost should be '10' or something like that. So here's an interesting

Re: [Yum-devel] repo priorities

2007-09-12 Thread Florian Festi
Florian Festi wrote: with the current code this is only partially true. In most cases pkg objects are reused now. IIRC there are a few cases left where we create new pkg objects. Instead of creating some parallel data structures if would be easier and much cleaner to fix that last few places

[Yum-devel] Re: [yum-cvs] 2 commits - yum/__init__.py yum/rpmtrans.py

2007-09-12 Thread Tim Lauridsen
Jeremy Katz wrote: yum/__init__.py |2 +- yum/rpmtrans.py |5 - 2 files changed, 5 insertions(+), 2 deletions(-) New commits: commit ac87ad138493036b6dbfec4cbfabf9daa3313f81 Author: Jeremy Katz [EMAIL PROTECTED] Date: Wed Sep 12 10:41:20 2007 -0400 event() only takes one

[Yum-devel] RFC: Capturing scriptlet output during transaction

2007-09-12 Thread Jeremy Katz
With the new transaction handling stuff, we can actually start to get to where grabbing the output of scriptlets is pretty doable so that it can be more easily consumed by tools sitting on top of yum. anaconda has been redirecting the bits to a file roughly forever (and thus, that's the API

Re: [Yum-devel] RFC: Capturing scriptlet output during transaction

2007-09-12 Thread Jeremy Katz
On Wed, 2007-09-12 at 16:38 -0400, Jeremy Katz wrote: +def _shutdownOutputLogging(self): +try: +del self.base.ts +except: +pass And actually, this bit isn't needed when we plumb things into the transaction handler -- it was only needed when I'm