This is an attempt to fix some problems that I experience with the previous text-mode Wizard. This new one is a complete rewrite, and has a different way of "going about it".
So far I have only tested on my laptop, which is a particularly difficult case, as it has hybrid graphics, Intel and NVIDIA video hardware on the motherboard.
The default behaviour is the kernel loads both the 'i915' and 'nouveau' modules, which require KMS. However, this prevents the 'vesa' driver from working, in case I had wanted to use that.
The new script should fix that kind of problem.
If anyone wants to try it, please do -- especially if you want to dig into the script and fix any problems! It is committed to the online Woof repository, and you also have to get the latest /root/.profile and /usr/sbin/xorgwizard.
Note, I was uncertain how much effort that I should put into this, as it is not known what the future holds for Xorg. Perhaps Wayland will move in? Maybe Mir will get to a usable status?
Comments:Posted on 27 Apr 2013, 23:42 by pemasu
Upup Raring with 3.8.7 kernel and jwm-733. xorgwizard and xorgwizard-cli creates working xorg.conf. Does not cause hanging anymore after restart X due to usable xorg.conf. I changed in BIOS to use discrete ATI graphics. Radeon driver in use. The synaptics wasnt enabled and wasnt able to edit xorg.conf yet so that synaptics would have been in use. something prevents synaptics_drv.so to be loaded.
Otherwise was ok. I will test what happens with intel graphics and with synaptics next...
I used Ubuntu Raring packages of today.
Posted on 28 Apr 2013, 6:22 by technosaurus
abandoning Xorg ship
From what I can tell the xfbdev is just as good if not better than the Xorg counterpart and the kernel fb drivers are cross platform (Xvesa isn't) and easier to fix (there are more framebuffer examples and the examples are better than Xorg's) and would even be relatively simple to port to no longer use libX11 components (only xcb). The one thing holding back the fb version is the clunky vga=ask menu, but it would be straight forward to add a vga=max parameter (suitable for most users) by tweaking
I am actually planning to do this to try and eliminate libX11 and since there is no existing toolkit for xcb, I'll be creating an immediate mode gui toolkit on top of netsurf's framebuffer (which has linux-fb, xcb, wayland and sdl backends)
Posted on 28 Apr 2013, 8:20 by BarryK
I also have only testing in Raring Puppy, which has xorg-server version 1.13.0.
I don't yet know how well it will work with older xorg servers.
Posted on 28 Apr 2013, 9:06 by BarryK
Synaptics support added.
Posted on 28 Apr 2013, 14:19 by scsijon
OT cd for sale notice
just came across a line at bottom of the page, take it you still want to know when we come across them.
Posted on 28 Apr 2013, 16:56 by BarryK
Re CD for sale
Thanks, I removed that text.
Posted on 29 Apr 2013, 4:15 by pemasu
xorg.conf.new left in /root
Feedback of the new xorgwizard:
There seems to be leftover remnant in /root. Cosmetic thing mainly. Also question about nv_drv.so driver selectivity from /usr/lib/x alternate location. I cant test. I dont have nvidia graphics at all.
Synaptics touchpad started to work. Thanks.
Posted on 29 Apr 2013, 4:16 by pemasu
better link for xorgwizard
Sorry. Wrong tag:
Posted on 29 Apr 2013, 6:36 by K Godt
dialog error Lupu 511
Have tested this new xorgwizard and fixed the lines that mwould prevent it to run in Lupu .
Also added shinobar's keymap-set, which is not included in my Lupu though the data matches ( *2010* ) .
[ ... ]
echo 'exit $?' >> /tmp/xorgwizard-driverlist-choose.sh
chmod 755 /tmp/xorgwizard-driverlist-choose.sh
Error: Expected at least 6 tokens for --menu, have 4.
Use --help to list options.
Posted on 29 Apr 2013, 8:43 by K Godt
Xorg -configure segfaults
Found the problem :
Xorg -configure segfaults with not creating /root/xorg.conf.new
and /tmp/xorgwizard-driverlist0 is empty .
X.Org X Server 1.7.6
Release Date: 2010-03-17
X Protocol Version 11, Revision 0
Build Operating System: Linux 2.6.24-25-server i686 Ubuntu
Current Operating System: Linux puppy_luci_218 184.108.40.206-KRG-i586TSC-64GB-1 #9 SMP PREEMPT Fri Mar 25 03:44:43 GMT-8 2011 i686
Kernel command line: panic=56 root=/dev/sdb6 ro vga=normal panic=30 log_buf_len=512k printk.time=1
Build Date: 23 April 2010 05:11:50PM
xorg-server 2:1.7.6-2ubuntu7 (Bryce Harrington <firstname.lastname@example.org>)
[ ... ]
0: Xorg (xorg_backtrace+0x3b) [0x80e937b]
1: Xorg (0x8048000+0x61c7d) [0x80a9c7d]
2: (vdso) (__kernel_rt_sigreturn+0x0) [0xb78d0410]
3: Xorg (xf86CallDriverProbe+0x182) [0x80b82c2]
4: Xorg (DoConfigure+0x1c8) [0x816b8d8]
5: Xorg (InitOutput+0x1da) [0x80b989a]
6: Xorg (0x8048000+0x1ebbb) [0x8066bbb]
7: /lib/libc.so.6 (__libc_start_main+0xe6) [0xb7629bc6]
8: Xorg (0x8048000+0x1e961) [0x8066961]
Segmentation fault at address (nil)
Last part of my /etc/X11/xorg.conf :
#everything past here is auto-generated by Puppy's Xorg Wizard...
Funny, that xinit is able to fire up X though
Posted on 29 Apr 2013, 18:48 by BarryK
Thanks, added a line to remove it.
Posted on 29 Apr 2013, 19:03 by BarryK
xorg alternate driver location
The script xorgwizard-cli has this:
#during woof build, may have put some drivers here...
cp -a -f --remove-destination /usr/lib/x/drivers-alternate/* /usr/lib/xorg/modules/drivers/ 2> /dev/null
rm -f /usr/lib/x/drivers-alternate/* 2> /dev/null
Because I don't know if we need that mechanism anymore.
Perhaps it is still useful for the automatic startup of X, but if a user needs to run the text-mode xorgwizard-cli then it runs "Xorg -configure" which detects all drivers that may be suitable for that video hardware -- and I think it best if all drivers are present, i.e. none in the alternate-drivers folder. Then the Wizard lists the found suitable drivers, and the user chooses which one to use.
However, I am not certain. Have only tested on my laptop, with hybrid Intel and NVIDIA graphics, and the -configure only offered:
...it didn't offer 'nv'. Which is strange. Why not? It would seem to indicate that the -configure is partial to ignoring nv if nouvea is present.
But, we need other testers who have NVIDIA graphics.
Posted on 30 Apr 2013, 7:57 by don570
Is rebooting needed??
On my IBM desktop with intel graphics chip,
I chose 'Intel driver,needs KMS(requires reboot)'
but is the reboot required??
I was able to get to the desktop, then the machine
went black screen after a few minutes.
I go through the steps on Pemasu's Raring thread
Perhaps you can get rid of that 'reboot' step
in the procedure.
Posted on 30 Apr 2013, 13:16 by zygo
Stipulating driver and resolution
Does this wizard allow one to tell xorg the driver and resolution and not to probe? 3 posts starting at
Posted on 30 Apr 2013, 19:54 by BarryK