> D'oh! That was what I wanted to see! Which other messages were filling up > the screen?
Only the ones about set_hyper. Can you build me a version without this output? > In general, you will have a better chance to see the right messages if you > reduce and simplify the active processes depending on the filesystem that > is going to crash. For example, use a couple of telnet/rlogins only, and > punt the xterms. You are much less likely to incur extra attempts to use > the dying filesystem if you don't have random expose and focus events > waking up processes on your hurd. OK, but the "wrong" messages in this case seemed to depend only on whether any writable ext2fs's existed other than / . I'd like to get that out of the way and try again.
