![]() |
Hi Martin, At 11:42 19/11/98 -0500, Martin Renters wrote: >The start_msg and stop_msg flags are used as user-friendly labels >when the system starts up. It executes all the start_msg commands >produces a list on the screen and then goes through that list one >line at a time and tells you whether it was executed OK, had errors >or was not applicable. I will modify my patch so that it adds these two (start_msg and stop_msg) to the init script, they will not cause problems for any other platform, they just won't ever be run except on HP. I found a document on this stuff: /usr/share/doc/filesys.txt which explains all this stuff. It mentions that for public applications a number should be specifically reserved from HP. I guess i will email them and ask for a number. >From the doc i think we should do something like: ln -s /sbin/init.d/hylafax /sbin/init.d/rc2.d/K150hylafax ln -s /sbin/init.d/hylafax /sbin/init.d/rc3.d/S850hylafax or whatever the reserved number turns out to be. >> > p.s. hylafax runs as a client only on my HPs because their serial ports >> > suck too. >> >> Ahh. So can we pitch HP as a server altogether? Or should we try to include >> that? > >This is wrong. HPs work fine as servers. Try setting them up to use >hardware flow control if you're having problems and remember that some of >the HP server class machines need HP serial cables (or custom cables with >HP's non-standard pinouts - look on their support web site for more >info) HP's work fine as servers - the cable pinouts are annoying but i think they have fixed this on more modern machines. There's another annoying problem with the serial lines that always need some sort of getty or other process running on them to work properly - otherwise the kernel sets the line back to 300 baud. This tends to break the faxaddmodem script and other scripts unless there is a getty running in the background at the time the script is run. - Robert, who knows now for *certain* why he was putting this off...