Task-number: QTCREATORBUG-3141

Always nice to know which part failed.

Preparing the introduction of tooltips for the new
CDB engine. Fix some breakpoint states in CDB.

QFontMetricsF::averageCharWidth seems to always return a fractional size
(at least on Linux/GNOME with DejaVu Mono), even when the font is not
actually drawn at sub-pixels. This caused the visible wrap column to
appear in the wrong place.

As a workaround, use the width of the character 'x', which is pretty
average and does align to pixels when needed.

Task-number: QTCREATORBUG-2746
Reviewed-by: con

Run some commands synchronously. This avoids issues with the UI
reacting to file changes done by the VCS on windows which can lead
to crashes.

Task-number: QTCREATORBUG-3021
Reviewed-by: Tobias Hunger

Introduce GDBMI-based 'stack' extension command instead of the
builtin 'k' as this does not print the correct instruction pointer.

Task-Number: QTCREATORBUG-3157

Reviewed-by: Tobias Hunger

Do not drop stack index in resetLocation (which is called
when switching). Do not switch to source unless we have
the file.

Reviewed-by: hjk

I needed to find out how this works the hard way, so add
at least some hints on how to work with code folding:-)

Allow to do codefolding on files and hunks in the diff editor

Task-Number: QTCREATORBUG-3100
Reviewed-by: Tobias Hunger

Reviewed-by: Tobias Hunger

Task-number: QTCREATORBUG-3136
Reviewed-by: Erik Verbruggen

The template .pri file of the "Qt Quick Application" and
"Qt Mobile Application" wizards linked against eiksrv
when locking the screen orientation. That is however not
necessary.

The real problem is that eiksrv does not get shipped with
the Symbian^1 SDK in Nokia Qt SDK 1.0/1.1. Therefore, the
innecessary link attempt even fails.

This patch removes the -leiksrv and bumps the version for
the template.

Cherrypicking into qtcreator/2.1 is highly recommended.

Task-Number: QTCREATORBUG-3142
Task-Number: QTSDK-194

Well, this is just a temporary hack :)

Task-number: QTCREATORBUG-2313
Reviewed-by: Roberto Raggi

Formatting of disassembler view took too much time.

This was used to enhance the user experience in the gdb/MI (non-python)
days as the location marker was shown faster. This does not seem to
make much of a difference anymore. Removal also avoids the "show disassembly
first, source next" flicker.