[Yum-devel] regets in urlgrabber

[seth vidal]

On Thu, 2005-01-20 at 01:22 -0500, seth vidal wrote:
> Hey Guys,
>  i've not done much work debugging this yet but I have an extremely
> simple test case.
> 
> In repos.py line 404, reget is commented out.
> 
> When I uncomment it and try to do anything that gets files it will
> magically fail that mirror and move along to the next, where it will
> fail, until it runs out of mirrors.
> 
> Thoughts on what I should be looking at to fix this one?
> 

Never mind - I think I figured it out.

It seems like byte ranges + reget = 'simple' == not so happy.
So I worked around that. If you'd like to test it give it a whirl.

I found one, very minor, bug in urlgrabber - grabber.py - line 904:

except socket.error:

it needs to read:
except socket.error, e:

I've checked in the change to the one inside yum - figured you'd want it
for cvs as well.

I think regets is working reasonably as expected, i've tested the error
conditions and I think it's right.

Michael, Ryan: If either of you could get the chance and check the code
in yum/repos.py Repository.get() and Repository.setupGrab() and see if
the retries functionality is working like you'd think it should, I'd
appreciate it. For some reason I'm thinking it's not b/c of how quickly
it fails over to the next mirror.

Thanks,
-sv