in that case the reply to write simply is zero, and errno then gives the filesystem full number.
I guess we're missing a select() loop for openBSE compatibility here anyways.
since thats the temp directory which ran full [core dump :-( ] its running full independent of the other filesystems.
so... while the real errormessage is a must have (I hope I did that correctly, plus added matching http error codes in webcit) locking up inbound smtp etc. in that case would be a BIG DON'T, since it tends to be a completely different filesystem.
The last fs full error I caught was with the netconfigs dumping. it would probably a good thing to halt networking jobs in a disk full scenario, plus signal it to somebody. (I don't think aide messages are enough for that)
I don't know whats happening once the BDB files run full, probably a good thing to test + then implement smtp 450 + a total lock on all write operations.
something completely differnt (well at least most)
which way do you add images to the blogmode? attachments to the message? or files api?
I guess the first could become a _big_ performance issue once the number of views raises.