On 11/21/2011 10:38 PM, Balcaen John wrote:
Le mardi 22 novembre 2011 00:11:49, Frank Griffin a écrit :

Please understand that I have *never* modified any Mageia-installed NM
configuration file.
Ok so in fact you did not follow my suggestion here
https://bugs.mageia.org/show_bug.cgi?id=865#c1 ?

You know very well I did, as documented in https://bugs.mageia.org/show_bug.cgi?id=865#c2 .

To be more precise, the problem originally occurred with configuration files as installed by Mageia, unmodified by me. I was simply making the point that the problem scenario wasn't some custom arrangement that I might have screwed up.

hum ?
The only comment here was that :
1) ifcfg-mdv plugin was broken (&  now in fact dead in mageia 2+)

2) try ifcfg-rh to see if it's also affected  (
https://bugs.mageia.org/show_bug.cgi?id=865#c6 )
you did not try arguing that the maintainer should report it upstream.
  ( https://bugs.mageia.org/show_bug.cgi?id=865#c8 ) which is quite difficult if
the maintainer (well in this case there is no maintainer so it's even more
difficult) is not able to reproduce simply because for example he does not have
the same wifi card. That's why i suggest you to report it upstream since you
should be able to answer questions from upstream if needed.

This entire discussion makes it evident why I wasn't the best one to report it upstream. Far more about how NM was introduced into Mageia has come out in this thread than was ever posted in the bugs I mentioned. The googled references to other people having the problem were not hardware-specific, and didn't involve the hardware I was using, so the knowledge of whoever implemented NM in Mageia would be far more valuable to upstream than access to my hardware. I posted detailed logs, and I would have been following the upstream bug in any case.


3) keyfile is not affected by the hostname bug in #865

However when i suggested to switch to keyfile (on irc indeed not on mageia
mailing list) instead of the ifcfg-rh plugin or (any others non native
plugins) i was told that we need to support ifcfg (which is right indeed)

So in summary we have a working plugin (keyfile) but we're not able to use it
because we need to support ifcfg- files so we're stick with the redhat plugin
for now until we found someone able to fix the missing functionnalities in this
plugin.

This really is not about saying you (or anyone else) did anything wrong here. I'm just making the point that Cauldron shouldn't be a one-way street. When something gets dropped in and has serious enough problems, and the situation (for whatever reason) prevents us from being able to fix it, then the appropriate thing to do is withdraw it until we can.

Reply via email to