trying test-driven bugfixes in plasma-framework

[Marco Martin]

Hi all,
Since we are trying gerrit, let's try to push it even further:

Right now, we are freezed, so that means in plasma-framework, plasma-workspace 
et al will happen only bugfixes for a month or so.

Having Gerrit on plasma-framework, I hope that makes easier for me keeping 
track of changes in plasma-framework: this is an experimental period, so is 
still not sure if it will make things easier or harder for everybody, let's 
try.
For that would need reviews not to stay hanged there for more than one day or 
so, but on the other hand to be actually reviewed before being pushed.
here is the list of open things, let's see if it can perform better than the 
reviewboard one
https://gerrit.vesnicky.cesnet.cz/r/#/q/project:plasma-framework

It's a good moment now, since being mostly bugfixes, it should be mostly 
atomic commits without big changes (I still have to decide if the Dialog 
refactor can fully be considered bugfix or if is better to wait next 
frameworks release, usual problem it's hard to autotest x11-related, 
windowmanager-related stuff).

One thing I would love for bugfixes at least in plasma-framework it to have 
them accompained by a test, that reproduces the bug.
I know it will not be possible for all kinds of bugs: not for crashers, not 
for most pure aestetical problems.

But at least some of the behavior-related bugs that are caused by some piece 
of API not working as advertised under some particular conditions, that can 
have and should have a new autotest.

It was relatively easy to add those kinds of tests for the startup related bug 
Alex reported on friday, so I'm moderately optimist about this (even bugs of 
qml components stuff, there's a test framework for that(tm)).

If we get coverage enabled, that would be even better.


-- 
Marco Martin