View Single Post
Posts: 114 | Thanked: 298 times | Joined on Jan 2011 @ Berlin
#70
Originally Posted by marmistrz View Post
1/ You might look at MeeGo CE rather than nemo, as Nemo can use mlite (a subset of meegotouch)
2/ And what about putting the library into /opt/xcb/ or so?
3/ What about, as a next goal, upgrading x11 as a whole? Some packages are already in -devel. (x11proto-input-meego-dev & etc. - the names are changed).
2) Putting them to /opt/xcb only helps if we can force the fully qualified soname or use LD_LIBRARY_PATH, I'd like to avoid both. But I've no problem putting the renamed libXc11.so.6 (and the xcb part) to another directory and change the paths for ldconfig, if you think this is better than optify.

3) I thought, libxau and other cause severe problems. However if they only cause problems on the stock kernel we might consider a power kernel only (and thumb only) build.

Note however that the way I've currently solved this is not a big burden for the future as a simple symlink from libXc11* to the new X11 libs can solve this (if and only if we can use them as a drop in replacement of the properly named old libs).

I'd like to make a first usable (for developers) version of Qt5 as soon as possible, because porting X11, porting Qt5 itself and porting and/or creating application with Qt5 are things that can be done in parallel by different groups. And the possible users of Qt5 (esp. QtQuick2) outnumber any other group. So if they can program slightly buggy but working programs with a (slightly) buggy Qt5 on an old (and buggy?) X11, we'll have actually stable, usable programs if we switch the back-ends to less buggy ones.

TL;DR: This works for now, helps bringing Qt5 to devs and we can change it later (without too much effort).
 

The Following 2 Users Say Thank You to frafI For This Useful Post: