Qt build that includes commit 0c1831178540462da31fd7a4b6d2e446bc84498b

This should work.
I mean, msvc detection part...

Hm, looks ideologically similar to how it was done in SDL2...
https://hg.libsdl.org/SDL/file/792c26401463/src/power/linux/SDL_syspower.c

I could take over this revision and update the diff as required (even without adding braces around if statements), but I guess we will need to know @bgray 's email to be able to preserve authorship.

In D21267#466653, @pino wrote:

LGTM now, thanks.
Works in X11; I assume you tested it under wayland, right?

In D21267#466632, @pino wrote:

Another idea is to borrow a small bit from D20402, in particular the change in main() that wraps checkX11Capabilities() within a X11 check: as "else" case for that if, you add the error message added here.
This way, after this is done, D20402 will need to add a no-op check for wayland.

• Wrap checkX11Capabilities() in QX11Info::isPlatformX11() check

This patch does not apply on master now, needs some changes like that:

diff --git a/krfb/mainwindow.cpp b/krfb/mainwindow.cpp
index 8de7822..ad2b04c 100644
--- a/krfb/mainwindow.cpp
+++ b/krfb/mainwindow.cpp
@@ -34,6 +34,7 @@
 #include <QVector>
 #include <QSet>
 #include <QNetworkInterface>
+#include <QHostInfo>

Maybe we should land this? 😉

@pino thanks for looking at this! Fixed your comments.

• Check only for platformName and rephrase the error message.

I could test with hidpi screens, but I need to understand what was the problem and how is it fixed.

I tested, compiles and works, does not open any new windows for me, just focuses address input (I don't have any connections configured). Looks harmless overall, +1

Is it because process.h exists somewhere else?

In D18651#403466, @broulik wrote:

+1 first thing I change on a fresh setup

In D18621#402558, @ngraham wrote:

In my opinion, this sort of thing is a "help the user feel in control" issue.

I can reproduce an issue. Without patch, VCS error generated in one tab is displayed in all versioned tabs (but not in normal unversioned tabs). Also strangely error is multiplied only on the first occurence, if after first appearance of error I close all error messages in all tabs, and trigger error again, it only appears in current tab, and not in any other.

branches is a QStringList, so, we could also check that currentBranchIndex < branches.size() in condition, if it is all about bounds checking.
Anyway, the diff compiles and overall looks harmless. I'd say +1, I did not try to reproduce the crash though

To be precise, passwords are not stored as plain text, scrambled using KStringHandler::obscure(). But this is totally insecure anyway, is a wording like this OK?

• Add a warning about insecure storage in config

I tested this, it indeed fixes 390225
Withoug this patch, generated pkgconfig file was:

Probably because it is constructed globally before QCoreApplication / before translations were set up, or something like that

Works here with 2 monitors like this, tested all modes, seems OK.

I tested on current master it and it compiles and works; in several capture modes, but only on a single monitor (3840x2160). I can test later (on monday) with dual monitor setup.

same, Xiaomi Redmi Note 4X (Snapdragon) [mido] , Adreno 506, MSM8953 Snapdragon 625
https://github.com/Halium/projectmanagement/issues/71#issuecomment-443450191

In D16744#355899, @cfeck wrote:

Would a switch be more readable?

• Fix typo "in" -> "on"

1. What if in the future another, horizontal spacer component will become needed? Will you add a HorizontalSpacer.qml then? 😆 Maybe have just a Spacer with horizontal property?
2. Why Rectangle, it could be just Item, it does not cause rendering (Rectangle does)

Fine, @utecht do you have a developer account to land this?

Tested, works fine otherwise.

Well, ok

@pino 👌 , so what would be the preferred way to fix those small but eventually-needed changes?

Spectacle code is far from perfect, I'm suspect there are many other places to add const or fix Q_FOREACH ...

Compiles, works as described, no questions to those 3-6 lines of code modified 😉

I'll test tomorrow, going to sleep >_<

Resolve confilcts between <<<<<<< and >>>>>>> in several places?
Pretty sure this won't compile 😆

In D10099#342564, @ngraham wrote:

Hmm, why would the user do that? If there is no template, the behavior is unclear and undefined.

D10099 Use single number as filename for screenshot with empty filename template

What happens if user deletes filename template in Spectacle settings? Does new token help?

Original source. I also slightly changed formatting and added const keyword where found possible.

We (me) could take over this patch, if original author is no longer available 😕

https://cgit.kde.org/plasma-nm.git/diff/vpn/openconnect/CMakeLists.txt?id=a44b7f38c548d756b6cb822c047ef71cf67645f5
Probably this commit could remove dependency on QtXml (I'm not sure, QtDbus probably depends on QtXml and QtDbus was removed)
If it fixes build error .../vpn/openconnect/openconnectauth.cpp:38:10: fatal error: QDomDocument: No such file or directory then fine...

How can I reproduce broken build? Any test plan?

Maybe add it to .pro file too then?

Why isn't repository for this review set? Which repo should it be?

Looks harmless and correct

From all of my installed build tools: 19.1.0, 20.0.0, 21.1.2, 22.0.1, 23.0.1, 23.0.2, 23.0.3, 24.0.0, 26.0.1 it selects _default_sdk as 19.1.0, is it fine? 😄
I mean, which one should be used ideally?

But Currently, libcanberra is tested on Linux only. 😢 Plasma-pa and plasma-whatever can use anything and be Linux-centric, KNotifications is a framework, isn't it?
++vote for QtMultimedia option...

Don't know how much offtopic will be what I'm going to say, but I recently ran into an issue while packaging QQC2 application in flatpak with org.kde.Platform//5.9 runtime. My app was functioning perfectly fine in system, but under flatpak it failed to start with the following message: QWidget: Cannot create a QWidget without QApplication. Of course, I had only QGuiApplication. But I wasn't using widgets! The only class from QtWidgets was QSystemTrayIcon for a nice integration with desktop. My guess was that was caused by platform style used in flatpak's KDE runtime may use widgets. But this was unexpected, without flatpak all was working. In the end I was forced to use QApplication, and visually saw the style used was ogr.kde.desktop.

This is probably for BUG: 396214 ?

BUG: 396214 or better BUG: 364746 ?