- User Since
- May 24 2016, 11:24 AM (143 w, 5 h)
Sun, Feb 17
Fri, Feb 15
Can you add context to this diff by either using arc or git diff with more context?
Sat, Feb 9
I can confirm that it fixes all border display issues that I encountered in the released version.
Fri, Feb 8
Thu, Feb 7
Wed, Feb 6
Tue, Feb 5
Landed (phab didn't notice): https://cgit.kde.org/kdeclarative.git/commit/?id=856672f370fb32c7a3c3d1a13d873fb8b767d0e8
@trmdi: Do you have push access? If not, which name <mail> should be used for the commit?
The second change fvogt mentioned is IMHO handled quite strangely here - !m_enabled || m_temporaryInhibition is the inverse of !m_enabled || m_temporaryInhibition, so doing a event->ignore() === event->setAccepted(false) prior to the first return statement would cover the ignored case, and allows the event to bubble up.
Sat, Feb 2
Thu, Jan 31
I tried to understand what this change does both by trying to reproduce the issue and reading Qt code.
Here the symptom was more drags not getting accepted at all than flipping back and forth, but this patch fixes that as well.
Mon, Jan 28
Also bail out early if image data known broken.
(Unrelated to the overflow fixes, but nice to have)
Sat, Jan 26
We've had Qt 5.12 packaged and available since the first alpha and started to build KDE software against the second alpha - would it be possible to do it similarly in the KDE CI?
Starting to build also against pre-release Qt would be helpful as currently KDE CI is lagging behind what rolling release distros actually ship and it might help preventing some issues.
Fri, Jan 25
Not tested, but looks correct to me. (Famous last words?)
Replace XDG_RUNTIME_DIR, test still passes
Mon, Jan 21
Jan 18 2019
This should probably land on Plasma/5.12 as well.
Add some style
Use a new bool instead.
Hmm, maybe then add another setter setAutomaticSocketNaming instead to switch to automatic socket name query instead. When it's not called before start it would fall back to old behavior.
Jan 17 2019
Tested with cmake 3.10.2, works.
Jan 15 2019
Jan 14 2019
Jan 13 2019
The login manager should set this already
Jan 11 2019
Jan 10 2019
Jan 6 2019
That's true, but if openSUSE run our unit tests it would be found.
Jan 5 2019
Dec 26 2018
Is this the issue this fix is intended for>
Dec 22 2018
Dec 21 2018
Dec 5 2018
Hmm, the "After" pictured depicts terrible kerning. Take a look at "Discover" in particular. Is this really the trade-off we're faced with? It feels like there are upstream Qt bugs here.
Turns out the label width rounding is enough
Dec 4 2018
Dec 3 2018
Because it doesn't need to. It makes no difference whether the object is empty or has only one of the keys I read.
Can host/mprisplugin.cpp deal with payload being null properly?
Nov 30 2018
I implemented the same hack here locally some time ago and it works perfectly fine if started with spectacle -platform xcb.
Nov 29 2018
It might work as well to just ignore the "interrupted" state completely and only react when error becomes set.
Nov 28 2018
OTOH, this patch actually causes a mismatch if one changes the cursor theme in Plasma's cursor theme KCM. (it won't affect GTK applications, they will continue to use breeze...)
Nov 27 2018
Nov 24 2018
Yeah, I had a feeling that that might end up as a better place for the fix. Someone else would need to do that patch or help me with it though. I gave it a shot but couldn't get the binding right.
IMO this workaround is generic enough that adding it into ScrollablePage.qml or even templates/private/ScrollView.qml itself is better as it hits almost every use.
Nov 23 2018
LGTM, but I have no idea about this part of JS.