[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Stopping X from dying on font issues (Was: Debian runlevel policy?)

On Sat, Mar 06, 1999 at 10:03:13PM -0500, Daniel Martin wrote:
> Then use an xset in /etc/X11/Xsession or in /etc/X11/xdm/Xsetup_0.
> This is what I do with other X options that I could put in the XF86
> config. file but don't for lack of knowing what the relevant option is
> called (for example, I do a 'xset +dpms' in these startup files; I
> also get fonts from a network fontserver which occasionally has
> issues, and an xset instead of a FontPath means I can still use my
> machine when the font server is down or unreachable).

I've no idea why the hell this did not come to mind but I will be making
this alteration shortly, thanks for smacking me in the head with the
obvious...  xdm would not start before the fontserver, though the X
server running it might.  Even if it did, if it's not there when the user
logs in to xdm no biggie---though it would CERTAINLY be up by then unless
it was broken.

> Actually, I might go so far as to say that the XF86Config file should
> not contain anything that can be set later with xset and which can be
> left out of the config. file without breaking it.  Hrm... this gives
> me an idea - is there some file which we can guarantee is read each
> time an X server is started (in the normal course of events)?
> (Xsession isn't read by an xdm-initiated server until someone logs in, 
> unfortunately)  If there were such a file, we could put all xset-able
> and potentially server-killing stuff into that file, and simply log
> all errors instead of having X go bouncing up and down.

Yes, look at how xbanner works.

> Of course, what I am willing to bet is the most common misconfiguration 
> - incorrect/inconsistent video settings - will still cause X to
> bounce, but anything that will make a few X configurations more stable
> (without, I think, causing too much chaos) is a good thing. (TM)

Indeed!  Of course, big-cursor font needs to still be there and stuff. 
Hmm, actually now you give ME an idea by answering your idea about the
big-cursor package...  it can be made much more useful than it is now
with a little work.

Joseph Carter <knghtbrd@debian.org>            Debian GNU/Linux developer
PGP: E8D68481E3A8BB778EE22996C9445FBE             The Source Comes First!
To boldly go where no bunch of geeks have gone before :)   --Joel Klecker

Attachment: pgpbvAjFlrxxi.pgp
Description: PGP signature

Reply to: