It's a useful feature for people that like to keep their environment clean. Thanks for working on this, Toni!

If you'd like this fix to be included in v2.7.2, please push it to master by Aug 4 and I'll backport to stable. Please don't forget to specify BUG, FIXED, Differential Revision tags in the commit message.

If you'd like this fix to be included in v2.7.2, please push it to master by Aug 4 and I'll backport to stable. Please don't forget to specify BUG, FIXED, Differential Revision tags in the commit message.

The new files were updated (copyright years were changed).

This is the new version of the source code, after some minor improvements, having updated it because of the changes to the source code that have happened since the last time (like in krsearchdialog.cpp), and having used daily the classes meanwhile.

It works properly now - thanks!
I checked the commit you mentioned in the bug. Before it the checkbox was still ignored but the links were always not followed.

It works and I'm fine with the proposed fix. Thanks Toni!

(optional) I'd use QFontMetrics::horizontalAdvance(A) and QFontMetrics::width(A) for additional type check.

Thanks Davide! It's cleaner and macro name is good. No issues with fonts were discovered when I tested.

No objections

Changed with macro, as suggested by @nmel; the code now is cleaner (and QFONTMETRICS_HORIZONTAL_ADVANCE seems a bit long)

Thanks for testing again!

Thanks for testing everyone!

If you'd like this fix to be included into v2.7.2, please merge by the end of the month.

If you'd like this fix to be included into v2.7.2, please merge by the end of the month.

It's an improvement. Thanks!

It works fine but please fix coding style before pushing your changes. Thanks!

I agree we can substitute the width method with the horizontalAdvance, however I propose we use compat.h to eliminate code duplication here.

Reasonable!

Thanks for checking, Davide! :)

The new code works using Kubuntu 18.04. Although the final behavior of Synchronizer is not perfect, the patch improves the current situation, therefore: thanks, Alex! (and Yuri :-) )

The new code works as expected using Kubuntu 18.04. Thanks, Alex! (and Yuri and Nikita :-) )

@yurchor : I follwed the guidelines from the documentation (https://doc.qt.io/qt-5/qfontmetrics-obsolete.html#width); I haven't tried to see where the difference really is, I think there could be some pixel overlapped or some extra space.

Isn't it be easier to use boundingRect(text).width() as a replacement of horizontalAdvance(text) which does not require Qt 5.11 and is not deprecated?

Looking at Qt, all the new methods are present in version 5.9 so that no checks with preprocessor are needed.

Tested on Fedora 30; it compiles and the bug is not present anymore; during the same test, a Krusader version without this patch was having this problem while the patched one was able to rename the file.

Reminder for git push :)

Works fine for me (3-day testing). Thanks.

Reviewed the code and tested — everything is great!

Tested to work as expected.

Forgot to add the reference to this revision in the commit message, closed with Shortcuts to move tab

New revision is also good - feel free to push to master.

Removed the single tab check in favor of a comment, like proposed by @abika

@gengisdave : A comment would be nice. Thanks.

@abika : this is intentional to define that these two methods do nothing when the panel has just one tab, I can change them with a comment remarking this logic

Looks good. Thanks for the improvement!

The new code works using Kubuntu 19.04. Other people can do their checks. Davide, using the same keys as Chromium, Firefox, Libreoffice Calc, etc. is a advantage. Thanks, Davide! (and Nikita :-) )

Thanks Davide!

Tested to compile, install and work as expected.

Thanks Nikita! Changes applied and pushed into maste.

• fixup! FileSystem: Fix using creation time from KIO as changed time (ctime)

Thanks for the fix and refactoring, Alex! Compiles and works as expected. Please check my inline comments.

Thank you, Yuri and Alex!

After rethinking this, I agree with you, Nikita. We should not impose the burden of supporting an outdated Qt version on ourself.

QOverload is only available in Qt >= 5.7, and Qt 5.9 is the earliest supported version.
See also https://bugs.kde.org/show_bug.cgi?id=405212 .
I proposed this change in the discussion of D19623 and there were no objections for a month.

Thanks! Landed (Hope this revision will close itself).

Tested with various umasks and local / remote fs - works nicely. Please don't forget to fix the build (and possibly, the commit messages) before merging. Thanks!

Alex, it looks good now - thanks! Let me test it and I'll get back to you in a few days.

You're right Nikita. I changed it now so that local files are created with QFile. Umask is applied here.

• SQUASH_ME Panel: Create new local text files directly with QFile

Refactoring code is fine, however setting 644 unconditionally introduces a security issue. The permissions for new files should follow umask setting. Otherwise users may unintentionally share their new files with other users.

Thanks for the reviews!

Thanks. Working fine.

Nice fix and improvement. Thanks!

• fixed build and formatting

Due to no response from Davide, I'm taking over. I also see lots of deprecation warnings and would like to fix it. The most recent revision doesn't compile, however. I have a fix and will update shortly.

+1
Compiles, installs and works as expected.

Kindly ping.

Hello All,

Kindly ping.

Hello Guys,

OK, so will you provide me the necessary permissions for that? I suppose, it is this repository: git@git.kde.org:krusader.git

In D16223#345641, @nmel wrote:

In D16223#344556, @mchabrecek wrote:

I can push it, I might need some help:
Is it enough to merge my branch to master, add code review link (= Differential revision https://phabricator.kde.org/D16223 ??) and push the master?
Or is there something more?

Since this is a single change and not multiple related changes, I would squash your branch into a single commit, rebase on top of the master, add the CR link and push it. Thanks!

In D16223#344556, @mchabrecek wrote:

I can push it, I might need some help:
Is it enough to merge my branch to master, add code review link (= Differential revision https://phabricator.kde.org/D16223 ??) and push the master?
Or is there something more?

In D16223#344206, @nmel wrote:

In D16223#344018, @mchabrecek wrote:

Is it approved and will be landed, or I have to send another arc diff with these 3 adjustments?

For minor tweaks like this we usually approve and let author incorporate the fixes before pushing. Just to reduce turnaround time. Of course, if someone else pushes (or lands) the change, the author should send updated diff first.

Miroslav, are you going to push this change to the repo or you want us to do it? I think it's worth getting the permission as you're going to do several consecutive changes on this refactoring and, hopefully, other improvements. Please read our commit guidelines before pushing.