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

Bug#229850: DEFAULT_HORIZ_SYNC and DEFAULT_VERT_REFRESH still ignored (updated patch)



On Mon, Sep 27, 2004 at 08:43:57PM -0400, Jay Berkenbilt wrote:
> 
> >   In re-envisioning the X server debconfage for post-sarge, I am considering
> >   a design along the lines of:
> >
> >   * Dumbing the interaction down to just the setting of parameters.
> >   * Providing fallback question defaults to address 2).
> >   * Providing user-invisible templates that "shadow" certain questions, which
> >     will exist to let autodetection tools poke values into them.
> 
> If my understanding is correct (and yours is probably much better than
> mine), XFree86 4.4.0 (yes, I know about the licensing problems) claims
> to have an autoconfiguration mechanism, and the X.Org (6.8.1?) forked
> off (again) of XFree86 pretty late in the pre 4.4.0 days.  I wonder
> whether much of this will just disappear when Debian moves to x.org.
> (Wishful thinking?)  I'm tempted to download Fedora Core 2 or Fedora
> Core 3 RC 2 (which are both using x.org) and see what their
> configuration stuff looks like.

At one point I had believed the X-Oz/XFree86-1.1 license taint had been in
the XFree86 X server autoconfiguration code from its first commit to the
XFree86 tree, but months later found out I was wrong.

The Debian X FAQ still has my wrong understanding in it, and I need to fix
that.

More to the point, I don't think there is any single magic bullet that will
solve the X configuration problem.  I look forward to having server
autoconfiguration being a tool in the toolbox, though.  Hopefully it works
out better than XFree86 4.0's "-configure" option, which was notorious for
crashing with SIGILLs.

My personal foremost goal for the X server configuration process in Debian
is:

* We damn well better not lock the user's machine.

Some people don't see to prioritize the above quite as highly as I do.
They'd rather see an install with fewer confirming strokes of the Enter
key, or simply don't care about users whose machines lock up, because their
own machines don't.  But as long as I have influence over the issue, I'm
going to uphold the above principle until and unless I am persuaded
otherwise.

This isn't to imply I've struck the best balance in my work on X
configuration to date in Debian; far from it.  xserver-xfree86.config.in
has grown tremendously warty, and the only person who seems to both
understand it and is willing to hack on it is me.  That's not really an
acceptable state of affairs.

This is why I want to make X server maintainer config scripts a little
dumber and a lot simpler, as I have written in other messages on this
subject.  Configuration and auto-configuration are really orthogonal
issues.  The former is simple, and is about transforming debconf answers
into a config file.  The latter is about sticking one's fingers into the
hardware and having long chains of deductive logic that pick the "best"
values for those debconf answers.

There are people who seem to have solved the autoconfiguration issue much
better than I, or at least taken much more popular stabs at it.  I'd like
to get xserver-xfree86.config.in out of their way.  Poke whatever values
you like into the debconf database, and the maintainer scripts will do the
rest.  Perhaps there will be several X autoconfiguration tools developed --
probably borrowed from various distributions -- and they can duke it out
for supremacy and/or the eventual blessing of the debian-installer team.

If an autoconfig tool picks the wrong values, or locks up the machine while
attempting to figure one out, I can dutifully reassign that bug to the
appropriate package, and it won't be cluttering the list of legitimate bugs
against X.

> >   Debian's XFree86 SVN repo is not hosted on alioth (fortunately?).  It's
> >   hosted at necrotic.deadbeast.net.
> >
> >   http://necrotic.deadbeast.net/xsf/XFree86/README.txt
> 
> Yeah, I eventually figured that out ;-)

Heh.

-- 
G. Branden Robinson                |      What cause deserves following if
Debian GNU/Linux                   |      its adherents must bury their
branden@debian.org                 |      opposition with lies?
http://people.debian.org/~branden/ |      -- Noel O'Connor

Attachment: signature.asc
Description: Digital signature


Reply to: