On Wednesday 28 February 2007 01:19, Hamish Moffatt wrote: > In all fairness, you didn't seem to comment on his response to your > suggestion.
I did comment on his suggestion. "Yes, it should start before NFS because its not just mounting a file system like NFS. It needs to make the block devices available." Not to mention there was a thread on debian-devel at the time where I thought he was MIA because he was so unresponsive. > (http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=387552;msg=30;att=0). > This seems to be more of a case of not liking his solution than having a > legitimate grievance? The solution flat out does not work. I mean, as his scripts are written, the machine hangs on shutdown. Here's the basic happenings of that bug. Sept. 14 - I send initial bug report. Oct. 2 - After 3.5 weeks the maintainer releases a new package, claiming it fixed the bug. I then attempt to point out that the script is not working. Oct. 3 - I attach my implementation of the init script that at least lets the filesystems mount. Oct. 9 - I try to get the maintainer to do something about letting it propagate to testing as I think the "fix" is not a fix. I do not escalate the bug myself because I am not the maintainer. Oct. 12 - The maintainer replies only because a thread on debian-devel[1] calls into question his ability to maintain the package. I then respond with a comment on the only thing relevant in the email (the order of start scripts and how AOE is different from NFS). I provide a link to Coraid's (the implementer of aoe in the kernel) documentation about setting up this hardware. Nov. 8 - I document why the new revision of the package doesn't work even for the mounting case, and I document that udev may be the right way to completely avoid the race condition. I'd like to point out that the above occurred over about a two month period. One response from the maintainer that was prompted by pinging debian-devel is ridiculous. As you can see there is only one human response in there, and I did respond to the critical question of his email. He seems to only respond when things get emotionally charged or when I publically tried to see if some other developer could comment on my changes. The maintainer's behavior made him very discouraging to work with. I just stopped dealing with him after a while. It hasn't stopped me from filing other bug reports. You may think I'm a whiner, but the fact of the matter is that my hardware flat out does not work with his scripts, and he was unresponsive to my suggestions for fixing them. He didn't even want to consider the use case that someone might want to treat a block device like a block device instead of a filesystem container. If you don't want to consider it non-responsive, it was at least really unproductive and frustrating. [1]http://lists.debian.org/debian-devel/2006/10/msg00451.html wt -- Warren Turkal, Research Associate III/Systems Administrator Colorado State University, Dept. of Atmospheric Science -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
