This is to resolve RH bug #458803 --- imgcreate/yuminst.py | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/imgcreate/yuminst.py b/imgcreate/yuminst.py index a7b04b6..628b0cd 100644 --- a/imgcreate/yuminst.py +++ b/imgcreate/yuminst.py @@ -103,8 +103,8 @@ class LiveCDYum(yum.YumBase): # dict so that things don't get pulled back in as a result # of them. yes, this is ugly. conditionals should die. for req, pkgs in self.tsInfo.conditionals.iteritems(): - if x in pkgs: - pkgs.remove(x) + if x.name in pkgs: + pkgs.remove(x.name) self.tsInfo.conditionals[req] = pkgs else: logging.warn("No such package %s to remove" %(pkg,))
On Tue, 2008-08-12 at 08:48 -0400, Jesse Keating wrote:
This is to resolve RH bug #458803
This ends up breaking compatibility with older yum, which isn't really acceptable. Also, the right thing is to fix yum to *not* change its API/ABI -- not to adjust every caller to handle the fallout
Jeremy
On Tue, 2008-08-12 at 09:42 -0400, Jeremy Katz wrote:
This ends up breaking compatibility with older yum, which isn't really acceptable. Also, the right thing is to fix yum to *not* change its API/ABI -- not to adjust every caller to handle the fallout
I don't disagree, was just working with what I got :/
On Tue, 2008-08-12 at 10:28 -0400, Jesse Keating wrote:
On Tue, 2008-08-12 at 09:42 -0400, Jeremy Katz wrote:
This ends up breaking compatibility with older yum, which isn't really acceptable. Also, the right thing is to fix yum to *not* change its API/ABI -- not to adjust every caller to handle the fallout
I don't disagree, was just working with what I got :/
You have the yum source too :-)
But fix worked up for yum and pushed by Seth about 20 minutes ago, so we should be good
Jeremy
livecd@lists.fedoraproject.org