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

Daniel Eckl <[email protected]> writes:

> Hi all!
>
> So the OS changes the resolving after switching network and glibc has
> to be notified of this.

Basically, yes.

> For me that sounds if it would not be operas job to do this. I'd more
> say, the OS needs to do this, and since this function is missing, one
> could write a little script and call it via if-up or write a
> (scripted) daemon that watches resolv.conf and then triggers glibc.
>
> Then as well opera as every other program using glibc resolving would
> benefit.

Solutions with if-up or daemons won't work.  The name resolver is part
of libc, which is loaded into each process.  So there are variables in
each process (maybe even each thread) that need to be updated with the
new values.

It may seem obvious that libc should automatically detect configuration
changes and reload if necessary.  But I hear that the glibc maintainer
does not want to do that.  I haven't looked into his reasons, so I can't
say whether those reasons are good.

> Best regards,
> Daniel
>
> 2010/7/23 Kenneth Crudup <[email protected]>:
>>
>>> > What about running "iwatch" on /etc/resolv.conf
>>
>> On Fri, 23 Jul 2010, Eirik Byrkjeflot Anonsen wrote:
>>
>>> I don't think we want to depend on iwatch being installed...
>>
>> Not the package- the system call.

Ah, ok.  Yes, using one of the system calls for monitoring file changes
where availabe could help.  Though we would still need to do something
else on systems that does not have such a system call.

We would also still have to handle cases where the name resolution
configuration is changed through different means than modifying one of
the files we know about.  Maybe that can't happen.  But I don't feel
like trusting such an assumption (yes, I'm paranoid that way...)

>> The reason I asked this in the first place is 'cause it seems like the
>> pages would finally load if Opera would just wait a while longer for
>> the DNS resolution (as evidenced by the fact that hitting "reload" can
>> usually make it work, but that doesn't work for pages that use form
>> data(?) and the like).

That is rather surprising.  I wouldn't expect that to happen.  Maybe
this problem is a bit different after all...

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