Akim Demaille <[EMAIL PROTECTED]> écrit:
> | Then, it would be easy for testers like myself to run such a script
> | blindly, and mail you the resulting archive that contains all the
> | info you need. That way we won't waste too much time b/t us: you ask
> | "can you try X and mail me the output of Y" and we have to re-run tests.
> Tom's scripts might help, indeed. Actually, this issue is so generic
> that it is surprising that there are no tools already developed. I
> wouldn't be surprised that François has something automated?
Tom Tromey <[EMAIL PROTECTED]> écrit:
> As far as I know nobody has done much here.
> My scripts have some nice features (I can run my tests on a number of
> machines at once -- I usually test 11 architectures or so), but nice
> error reporting isn't one of them.
I once thought about automatic reporting, but I quickly excluded this from my
mind, as many people would be extremely annoyed if we were forcing something
out of their system without their full control. Yet, between extremes, a
compromise would be welcome. It might be useful, indeed, that an all-ready
message be automatically prepared, and be left for the installer to send.
On the other hand, we should keep a way for the installer to easily try
to stab the bug himself, without necessarily having to unpack a message.
In a word, we should find ways so it is as easy and immediate as possible
for the installer to work on the problem if he feels like it (so encouraging
people to help us when they know how :-), as well as it should be easy
for the installer to wrap up everything in a reviewable report, and send it.
Instead of packaging a `tar' file within a message, forcing an unavoidable
unpacking step at the other extremity, I would favour a ready-to-send MIME
multipart message, more clear than Base64 if possible. That should ease
quick review before sending, as well as quick perusal on receipt, accompanied
by easy extraction of needed parts. This might often be sufficient.
--
François Pinard http://www.iro.umontreal.ca/~pinard
