Re: DNS caching(?) page-hangup issues and Opera 10.X

Kenneth Crudup <[email protected]> writes:

> On Fri, 9 Jul 2010, Eirik Byrkjeflot Anonsen wrote:
>
>> Interesting.  Now we know exactly where the problem appears.  That is
>> definitely helpful.
>
> So, maybe they'll have some headway in the next daily? I mean, I've been
> waiting for Opera to decide it knows how to look up a hostname that it
> has already displayed *five times* before (seriously, I only hit "refresh").

I forgot to ask:  You have reported a bug?  Do you remember anything
about that bug report that can enable me to find it?  (It seems it is
not possible to search for bugs reported by specific users.)

(The date it was reported would probably help.  Some "unique" text in
the description maybe.)

>> I would expect that this problem only occurs if a process has looked up
>> a name before the network switch and then tries to look up another name
>> after the switch.
>
> So, in the case like above where it manages to resolve a host once, but
> fails the next time (... or the time after that?).

Right.  (At least, assuming I understood the question...)

I'm thinking that one possible reason for this behaviour could be that
libc creates a socket for talking to DNS on the first lookup.  When the
network is switched, this socket is no longer valid.  But libc still
tries to use it.  That would certainly explain the EIO return value.

eirik
--
Opera-Linux: https://list.opera.com/mailman/listinfo/opera-linux
More lists: https://list.opera.com/mailman/listinfo/
Unsubscribe: mailto:[email protected]?subject=unsubscribe