[Asterisk-Dev] bug reporting howto?

[John Todd]

At 9:24 AM +0300 on 7/27/05, Tzafrir Cohen wrote:
>Here is a simple advice: if you want to report an issue with Asterisk
>and want bug marshals to actually read your reports, don't attach a
>patch to them.
>
>I have reached that sad conclusion from two bug reports:
>
>http://bugs.digium.com/view.php?id=4637
>http://bugs.digium.com/view.php?id=4734
>
>In both I have reported some easy-to-reproduce bugs. The first one seems
>to have been closed this morning due to the fact that the employees of
>Digium do not want to read a patch.
>
>Howevr that copyrights issue does not apply to problem descriptions in
>the bug reports. I have submitted a trivial patch in hope that it would
>help demonstrating the problem. However that issue can still be easily
>reproduced and discussed. In fact, I believe anyone familiar enough with
>the source of Asterisk could have easily traced the problem even without
>that patch.
>
>Current status of that bug is "closed'. I read this as "if we ignore the
>problem, it will go away". If you want to use that attitude, please do.
>The problem is fixed in the version I distribute, but aparantly not yet
>in the trunk. Such an attitude makes me avoiding reporting problems I
>have found altogether.
>
>
>So do I need to submit a disclaimer simply for having a bug marshal look
>at my bug reports? Or should have I not mentioned the patch to fix it?
>How should I report bugs?
>
>--
>Tzafrir Cohen     icq#16849755  +972-50-7952406
>tzafrir.cohen at xorcom.com  http://www.xorcom.com


Not having any visibility into this problem other than what was 
written in the tickets, I would suggest the following:

Do not submit patches that are not disclaimed.  Even though these 
patches may be only to describe further the problem, they confuse the 
single-minded.  In addition, if you submit a patch to which there is 
no disclaimer, and the patch that is independently developed by 
someone else looks very similar, then there is a potential conflict 
later when you might potentially claim someone "stole"  your patch 
code.  This may never be the case of course, but then that potential 
eventuality is the Fear that led to the whole point of the 
disclaimer, isn't it?

You may think that submitting a patch that is not disclaimed is more 
helpful than letting people figure it out, but it is easier to figure 
out the creation of a new patch than it is to wade through the 
political BS about why you may or may not have disclaimed something. 
Therefore, I would suggest that it is not more helpful to include 
patches until such time as you're ready/able to disclaim them.

Perhaps a more immediately helpful method would be to describe the 
patch, but not actually submit code.  You did so in #4637 but then 
the knee-jerk reactions about disclaimers distracted and got the best 
of everyone and the ticket was closed without actually discussing the 
problem.

I appreciate your work and discovery of these bugs, and hope that 
they can be re-opened and resolved in a suitable manner.

JT