Re: Uploading bug reports

[Franz Sirl <Franz dot Sirl-kernel at lauterbach dot com>]

Am Sun, 23 Aug 1998 schrieb Martin von Loewis:
>In many reports, no preprocessor output is included because people
>feel that it is too large for email. Sometimes, this is certainly the
>case.

Yes, but even worse it seems that Jeff and other people manually check all
messages >40k. First time I wanted to post a bigger bugreport I didn't remember
the exact message size limit and thought 'hey, it will be rejected by the
server if it's too big, so don't care for the moment'. I was really surprised
when Jeff personally replied to this message.


>It would be nice if we could point these people to a place where they 
>can ftp-upload their reports, and then send a message to egcs-bugs 
>where the file is located.  

I would like such a facility, but there's one big problem with this, abuse. So
the incoming dir would have to be write-only, which means that nobody can look
at the source of a report until the upload is confirmed by someone and moved to
a different directory.

>While this upload facility would be preferably on egcs.cygnus.com, any
>place would do for the moment. In the past, I had quite good
>experience with downloading such files from the Web (hi Franz), so
>anybody who could offer such a service, please step forward.

I think most people reporting here should have some public web/ftp space where
they can put the bug reports. Disadvantage here is that the sources will be
most probably be removed a while after they have been solved (like mine, cause
I only have limited web space), but being solved each report should have a
corresponding testcase now.

Nevertheless I think there should be such an upload facility as a last resort
for posting bugs.
And I think egcs should be "bloated" with the full instructions on how to
handle a full bug report on every internal compiler error. Lots of people
simply won't look into the docs and therefor send totally unsatisfactory "full"
reports.

BTW, a option like "-save-temps-and-strip-out-all-not-needed-for-function xyz"
would cut down report size too ;-).

Franz.