Re: Xserver (Xfree 3.3.6) complaining about a bad font server

[ppetrakis alphalinux org]

Hi Kirk,

The path mentioned there is simply a placeholder generated by RH's
tools. If you look for the directory it points to it won't be there. I
believe the reason it's there is so when folks go and do install TTF
fonts they make their directory to the path already specified so
everyone could have them in the place. Why didnt RH just create the dir?
Well that would be too easy ;-) 


Now to your other questions.

There are a few tools on a RH system that will configure X for you. One
of them is Xconfigurator, a tool of RH. Running that tool should
generate an XF86Config like when the system was first installed. It
probably references the font server. Honestly , I simply pull out the
reference to the FS and code the paths in the XF86Config file. Don't
have to worry about the xfs daemon dying. Now are you using RH-7.0 or
something earlier? Reason being there is a switch passed to
Xconfigurator that tells it which config file to write.
The flags are --preferxf3 and --preferxf4 .

Regarding your type 1 fonts. It seems that the font index files may have
been destroyed somehow. I know on my system it likes to eat my index
files for my TTF fonts. You can try this and it may help. 

mkfontdir /usr/X11R6/lib/X11/fonts/Type1

The font server is compilied with support for fonts. Same thing with
XF-3.3.6 . In XF-4.0 you have to load a module per each type of font you
want to display. That is if you're loading the fonts from the Xserver
not the font server. BTW TTF fonts work with XF-4.0 you just need to
load the 'xtt' backend instead of the 'truetype'. 

Peter


KirkE@paccessglobal.com wrote:
> 
> Greetings all,
> 
> Whilst tampering with my Xserver configuration files (XF86Conf) in an
> attempt to use the Matrox drivers instead of the default Super VGA drivers,
> I managed somehow to damage the settings for several of Xserv's dependent
> daemons, including the font server.  Strangely, it complains about a front
> server based on True Type fonts during automatic system startup or manually
> invoking X, but continues to load.  Having searched the directory tree, it
> looks like I never had a True Type directory, only Type 1 PostScript fonts.
> 
> I compared the original configuration file created by Anaconda during
> install, but it references the same TType directories, though I do not
> recall any errors, what gives?
> 
> Also, I noticed several tools for configuring the Xserver which I presume to
> have been partly X generic, Red Hat derived and XFree derived.  There was
> XF86Configurator, XF86Config, X86Cfg and a few others.  Each of these seems
> to write the XF86Conf file differently as though different versions of XFree
> were being targetted (though it does not give any warning of this).
> 
> Perhaps If I understood which tool was the correct one, I could
> simultaneously fix the font server problem (it complians, but still loads
> the X session, albeit with odd looking screen fonts) as well as tune my
> configuration file.
> 
> Right now, I have a hand built XF86Conf file with bits from each of the X
> configuration tools, though when I passed it the directory to the Type 1
> Post Script fonts that Red Hat's original installer created, it refuses to
> recognize it as a legitimate directory.  Perhaps the font server that is
> being loaded is not Type 1 PS aware but True Type?
> 
> I was tempted to see if I could install Metro or the other pay-ware Xservers
> just to see what it's configuration looked like and then de-install the
> demo. I've read through the XFree documentation, but do not see anything
> specific to this issue.  I thought that saving the original configuration
> file would be sufficient, but clearly some other configuration file has been
> touched.
> 
> Your assistance is appreciated.
> 
> --KE
> 
> -----Original Message-----
> From: Ivan Kokshaysky [mailto:ink@jurassic.park.msu.ru]
> Sent: Thursday, March 01, 2001 10:29 AM
> To: axp-list@redhat.com
> Subject: Re: No valid MMIO address in PCI config space
> 
> On Wed, Feb 28, 2001 at 03:40:48PM -0500, Jay Estabrook wrote:
> > Maybe we should use both? ;-}
> 
> No, my hack isn't needed, but ignoring disabled ROMs makes a sense
> in any case.
> 
> > And what about the other OSes that XFree runs on?
> >
> > I think the X folks took the "direct" approach, rather than have to
> > fiddle around with different facilities in different OSes and even
> > different versions, changing over time, etc, etc...
> 
> Looking at 4.0 code I have an impression that they try to use OS
> facilities if possible...
> Interesting, XFree86 on linux *does* read /proc/bus/pci/devices
> (Xserver/hw/xfree86/os-support/linux/lnx_pci.c:xf86GetPciSizeFromOS()),
> but obviously it fails sometimes. So we just need to fix this
> function (or kernel, whichever is broken ;-)
> I'll investigate this tomorrow.
> 
> Ivan.
> 
> _______________________________________________
> Axp-list mailing list
> Axp-list@redhat.com
> https://listman.redhat.com/mailman/listinfo/axp-list
> 
> _______________________________________________
> Axp-list mailing list
> Axp-list@redhat.com
> https://listman.redhat.com/mailman/listinfo/axp-list