What if you have Project, CTags and the new Zeal plugins active? You need to differentiate them somehow.

Let's take that one :)

I think the text could be "Cannot run command: %1\nWork path %2 does not exist."

I think this is a fix to the bug 347311

(Remove the unrelated change in src/view/kateview.cpp)

Also disable highlighting in selections for lines that are longe than 1024

Note: this is especially bad when the "on the fly spellchecking" is enabled as visibleRange() is used to "optimize" and only highlight what is currently visible. It is called multiple times.

Use EditingTransaction in stead of just disabling highlighting

If we get it working properly we could make the limit much higher. I tried with 30000 characters and it still worked very good. (except when selecting the line)

I tried it and it improves two things first the multiple replace is much faster and second you get the whole replace operation in one undo. I'll do an update when I have cleaned up the experiments a bit :)

The new connections are better because you get an error just for these problems. Explicitly failing is better than silently doing so on run-time.

Adding a function is not breaking backwards compatibility. You can still use Skanlite compiled against an older version with this new version.

I'm a bit confused... I thought KDevelop had much more advanced plugins/features for this? What does KDevelop need this old plugin for? Non-C/C++ languages or what?

I would like to give a -1 for removing the plugin I think that the feature is really something that has potential to be great!

+1

+1

Hmm... File Browser works badly on Windows and Terminal Tool View not at all.

If I understand the comments correct it assumes that if we use a libc that already has what we need, we don't need to use a separate library for it.

Looks good to me...

Please commit!

Looks good.

I was going to ask what the reason was for having a separate dialog, but your next review explains it.

I like it :) Please commit!

I like the direction!

use fromTheme("foo", fromTheme("fall-back"));

Sorry to hear that you are not interested any more :(

OK, Cool :)

It would add a some code complexity, but I'm not sure it adds so much value that it would be worth it.... If you want the tree-mode for one file type, the probability is quite high that you want it for the other types too. I don̈́'t remember any bug report that would request that feature... Are you asking for it? ;)

I was just about to commit this change when I noticed that:

Adding a member-variable to avoid one getter cal is just adding complexity and saves you almost nothing.

But now I'm curious which part of my text was so persuasive.

Except for the actionTriggered() function I'm ready to accept this patch.

In D14620#303899, @loh.tar wrote:

Do you miss the update in D14518 too?

The changes in the settings page is permanent, while the changes in the popup-menu is temporary -> I don't want to get rid of the settings page.

As said, it can/will be saved automatic. Then it is permanent and that page useless.

Otherwise I'm ready to commit.

I had missed this one....

I think it is ok to add these to the popup.

Of this one I'm not convinced ;)

While you update the string maybe you could remove the type (-1) and the comment about it?

I'd say go for 1)

Wonder if this should be a markdown parsing plugin ;)

We need to use i18n() for translations. the tr() calls will not be translated by the current translation system.

The plugin is available, but in the 5 minutes I tried I did not find a windows CLI application that reads the standard input to produce some output... ;)

How do you prepare your diffs?

Sorry!! forgot to add you as the author when I pushed :(

We can have the slotRefreshSymbol() merging in a different patch.

I guess the tree flickering comes from the fact that when you edit the document you change the cursor position and the current item might changed even if it shouldn't. Adding a 2s delay only slows down the flickering, but the actual problem, that the wrong item is being activated is now even worse.

What I still would like to have in this patch is the preventing of the m_currItemTimer from being started while m_updateTimer is active.

You cannot measure the CPU consumption of this particular function by looking at the "System Activity" while scrolling up and down the file. There are a million other code paths that get executed while you scroll the page. System Activity gives you an average CPU load of the application during 2000ms. This less than 1ms function only gets called once when you start scrolling and once when you stop.

It definitely does not mean "The CPU waste does not matter".