[Yum] xinetd-problem

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Thanks for the quick reply, Seth.

seth vidal wrote:
>>I've noticed this too, just last night in fact.  yum seems to have some 
>>issues with pkgpolicy=last not always doing what is expected.  For 
>>example, if I have three repositories containing a Squid package, in 
>>this order: Red Hat, My package, My package plus NTLM auth support.  And 
>>then I run yum update on a system that already has the Red Hat Squid 
>>package (which uses an epoch, if that makes a difference here), it will 
>>not update or even download the other Squid headers.
> 
> 
> pkpolicy=last doesn't mean always override.
> 
> if two repos have the same package the way yum determines which one it
> will present as available is (by default) the highest version of a
> package (meaning e:v-r comparison) will be the one yum pays attention to
> for updates/installs.
> 
> if pkgpolicy=last then that means the "last" package yum sees of a
> certain name-arch will be the one it pays attention to for
> updates/installs.

Ok.  Well then it is behaving as you would expect.  Good, it isn't a bug 
after all.  It isn't at all how I would expect, but that's my problem, 
not yours.  ;-)

>>Yet another issue (I guess it's a bug, but it seems almost intentional, 
>>since a sort of the serverids is required to make it work this way) with 
>>pkgpolicy=last is that 'last' is determined by serverid alphabetical 
>>ordering rather than order in the configuration file.  It took me this 
>>long to notice it, because by happy accident, my serverids have always 
>>already been in alphabetical order in yum.conf!
> 
> 
> This is the documented in the yum.conf man page
> 
> look for pkgpolicy in there.

Ah.  The docs have changed since pkgpolicy=last was introduced, or I 
have a poor memory.  Probably the latter...especially since I've 
'luckily' had alphabetized config files for the past several months, and 
so things have worked mostly the way I thought they should.

Ok, if I consider these things wrong behavior, will a patch that adds an 
additional choice that makes yum behave the way I would expect on both 
counts be accepted?  And should I call it "last-and-I-really-mean-it"?  ;-)
-- 
Joe Cooper <joe@xxxxxxxxxxxxx>
Web caching appliances and support.
http://www.swelltech.com



[Index of Archives]     [Fedora Users]     [Fedora Legacy List]     [Fedora Maintainers]     [Fedora Desktop]     [Fedora SELinux]     [Big List of Linux Books]     [Yosemite News]     [KDE Users]

  Powered by Linux