Bug#231862: closed by John Ferlito <johnf at inodes.org> (libvorbis0a: oggenc still dies sometimes with floating point exception)

[Helge Kreutzmann]

Hello John,
On Sat, Apr 03, 2010 at 03:33:08AM +0000, Debian Bug Tracking System wrote:
> This bug is ancient and is most likely fixed. Feel free to re-open if
> you don't think that is the case.

What kind of bug maintenance is this????

The bug was reported with a backtrace and an offer to provide the
file, and yet not a *single* reply was made to this bug, all the
years. Then you come along and state that it is "ancient" and
"most likely fixed". 

First, you could have analyzed the backtrace and handed it to upstream
to get it fixed (like the previous one, cited in this bug report). If
further debugging was needed, you could have asked for the file and
used it for debugging. By the way, there are porter machines for
Debian developers, where they can debug software. Of course, you could
also have worked with me, the reporter, to obtain further information,
try patches etc. 

And finally, if you simply hoped that a commit fixed the bug (e.g.
because you couldn't spot it and upstream was unwilling to help) then
you could have *asked* in the bug to try the latest version and see if
this bug was closed. 

Waiting like 7 years and then replying that the bug is "ancient" and
"most likely fixed" is probably the worst course of action I could
expected, even if you have no time (possibility) to debug this issue
(which would be sad for a DD). It really smells like "hope the guy is
no longer around and see that my bug count is getting down, because
many bugs let my software (me?) look bad". 

I hope this all was just a bad day and a gross misunderstanding on my
side.

For the content of the bug. I took the file, transferred it to an
alpha (not the original machine, which is dead by now, but some later
model) and tried encoding it again (unfortunately I did not note down
the exact command line options I used, so I encoded it twice with
different bit rates). In both cases, it worked. So I might have been
lucky, but I rather assume (given the error) that the general switch to
IEEE arithmetic on alpha in gcc several years ago fixed the problem.

So in essence, closing the bug was right but the way was *way wrong*.

        Helge
   

-- 
      Dr. Helge Kreutzmann                     debian at helgefjell.de
           Dipl.-Phys.                   http://www.helgefjell.de/debian.php
        64bit GNU powered                     gpg signed mail preferred
           Help keep free software "libre": http://www.ffii.de/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.alioth.debian.org/pipermail/pkg-xiph-maint/attachments/20100404/22f2ae91/attachment.pgp>