Enrico Weigelt wrote:
* Jakub Moc <[EMAIL PROTECTED]> schrieb:
<snip>
That somehow looks like the guided file-a-new-bug form we had some time
ago.
It's still there, just not linked from homepage (and needs a few touches
here and there)
http://bugs.gentoo.org/enter_bug.cgi?format=guided
http://bugs.gentoo.org/enter_bug.cgi?product=Gentoo%20Linux&format=guided
hmm, looks like an good start, but for me doesn't go far enough.
The user still has to type in too much manually.
I'd suggest an tool, where the user *first* gives the package name,
and the tool gathers any necessary information (version, useflags,
make config, ...). Dependend on the package, this tool may ask specific
questions or gather specific information (ie. on web applications,
we'd be interested in webserver, browser, certain network configs, ...).
The usage should be as simple as possible for normal users, so they
have great interest in using it. For example, asking the user to
look for similar bug is probably good for devs/maintainers, but
uncomfortable for plain users. But we shouldn't forget, that all
the plain users are (in their mass) also an important contributor.
(they find and report bugs, which probably never would be found
by the maintainers).
cu
I think this is an excellent idea. I wrote an automated trouble
ticketing system years ago for the mainframe and it's an excellent way
to get quality information into the system. The one drawback was that
we ended up having a lot more bugs entered into the system. Once you
have the shell it wouldn't be hard to hook it into batch processes and
have emerge automatically send bug reports when it fails. When this
happens you end up with multiple, identical bug reports because users
tend to re-create the error a few times before they head to the forums
or bugzilla for help. On the plus side it would be much easier to find
duplicate bugs since the titles would be uniform.
Glide
--
[email protected] mailing list
