On Sat, Jun 19, 2004 at 03:28:22AM +0200, Adrian Bunk wrote: > > Is screen linked with termcap or curses? screen prefers the curses > > library (though that shouldn't make much difference). > > It's linked with the NetBSD curses library (recompiling screen with > GNU ncurses doesn't make a difference). That's a little useful to know: Inspecting the termcap, I didn't notice much except that the enacs terminfo string was missing. Reading screen's code, I see that it has some logic for inferring things related to enacs. (But I don't read it often, so I'm not familiar with the logic). > > Normally (using only the information in the termcap), once line-drawing > > has been done in the given window, it'll be initialized as it was before > > the bug-fix. From the information so far, I'm not sure whether it's > > not getting the initialization string, or if something is using direct > > escape sequences to turn the graphcs set off. > > Is there any way I can help to trace the problem down? What I generally do to get some insight is to run 'script' around the whole session and analyze the 'typescript' file. That would at least answer the question of whether screen or ssh was sending some string that turned line-drawing off, or whether none of them turned it on properly. (How that will help me reproduce the problem, I'm not sure). -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net
Attachment:
pgppGqfqX5AhK.pgp
Description: PGP signature