ESR - the magic cauldron

Bryce McKinlay bryce@mckinlay.net.nz
Tue Jul 29 04:12:00 GMT 2003


On Tuesday, Jul 29, 2003, at 08:07 Pacific/Auckland, Erik Poupaert 
wrote:
> I've regularly felt that the requirement to submit a self-contained 
> example for a bug
> report is a potent deterrent to submitting any.
>> Not only the bug reporter may not necessarily have the time or the 
> inclination to
> create this self-contained example, but it ties the role of finding 
> the bug and
> creating the example into one single person. It could perfectly be the 
> case that
> other volunteers can be found to create the self-contained example 
> required?
>> If bug reports are valuable to the project, I would consider dropping 
> the
> request for an example. Maybe ESR has a few things to say after all.

For GCJ I am not aware of any requirement that bug reports be 
accompanied by an example/test case. However from the perspective of a 
developer, there is no way a bug that cannot be easily reproduced is 
ever going to get fixed. A report along the lines of "such and such is 
broken, it makes my app crash!" is of pretty dubious value. GCC 
developers don't have the time muck around chasing this sort of thing, 
they'll move on to bugs that they can see. Test cases are absolutely 
invaluable!
Bryce.


More information about the Java mailing list

AltStyle によって変換されたページ (->オリジナル) /