Suggestion: boot code 'xvesa=max'

[maro]

I've started to use a USB pendrive to boot TC from two different notebooks that support two different maximum screen resolutions. I was therefore wondering whether the 'xvesa' boot code could be amended by allowing in addition to a fixed "XxYxD" string the alternative value to be 'max'. A proper "XxYxD" string could then be determined automatically from the 'Xvesa -listmodes' output using the numercially "highest" value.

I'd imagine that using something like

Code: [Select]

[ "$XVESA" = "max" ] && which Xvesa > /dev/null \
    && XVESA=$( Xvesa -listmodes 2>&1 | sort -k 2n | cut -d ' ' -f 2 | tail -1 )
in '/etc/init.d/tc-config' (just before the '[ -n "$XVESA" ]  && sed -i ...' line) would do the trick.

Please note that I have not tested this code snippet properly, but I'm reasonably confident that the "calculation" should work as intended.


EDIT: When I finally went to put my own suggestion into action I realised that the command sequence was lacking a crucial element at the very end of the chain, hence it has now been amended.

[tinypoodle]

Very interesting approach which could improve the nomadic capacities of TC even more   

[gerald_clark]

Much easier is to modify .xsession or xorg.conf from bootlocal.sh

[roberts]

I had considered this way back (DSL days). However, the listmodes of Xvesa is reporting from the graphics card. The monitor is not necessarily able to display such a max mode. In fact, I rarely see such capability. Perhaps it might be useful on matched systems (monitor + graphics card), e.g., laptops, but even netbooks typically need 915resolution. I would need more convincing that such a boot option would be useful.
Would it really work on many systems?

[gutmensch]

FWIW I put the read-edid.tcz extension into the repo, which tries to read monitor modes and timings. Since its usage is a little strange I also put a script into the extension called monitor_modes, which returns Modes in a "Xvesa -listmodes likely" way. It's far from stable, so just testing ;-)

[curaga]

The *edid and dd*probe apps do not work on 64-bit, so there's a bit of a downside to them.

[gutmensch]

The author states read-edid 2.0 works on other architectures like x64 (depending on libx86), but I didn't try it out ;-)

Test 1 with VMWare: Doesn't work.
Test 2 with real 24" LCD Display and real ATI X1250 (*urgh*): Works.