thanks.

@z3ntu - do you have push access to push this?

These are leftovers from when all constructor was public.

I'm not sure I like the implementation. The current AbstractBarcode can also be implemented, at least theoretically, by software providing other barcodes that forever reason isn't available here. I have at least done that for some PoC ports of other things.

Regarding discount. I think the bettetr path moving forward is to port to newer Qt that has markdown handling stuff in QTextDocument

In D23866#553108, @mlaurent wrote:

Ping ? What is the status ?:)

In D23866#535611, @mlaurent wrote:

So what is the status of this patch ?

In D23866#529334, @ltoscano wrote:

Thanks, you have been fast. Would it be possible to use the QWebEngine-based KPart currently in konqueror.git?

Adding Montel because he has experience with locking QtWE down for e.g. mail usage.

Forgot some removed files

Seems to rely quite a bit on the khtml/kio integration and custom url handlers. Maybe there is a way for QtWe to have custom url handling and content serving.

Is the manslave still usefull without the manpart ?

there is a kchmviewer that uses qtwebengine directly for rendering. I don't know how that fits in the okular way of doing things.

Okular uses KHTML for the chm backend.

According to the RFC, the server should ignore case:

After reading the code a bit, I wonder

• can't some of the parsing code be made testable (and tests added)
  • This likely also requires to get rid of the global static.
• rather than poll /proc/mounts every second, I was wondering if one of the filesystem notification things (QFileSystemWatcher, KDirWatch) could tell when to re-parse it

In D20312#445122, @ahmedbilal wrote:

I think unit tests would be nice. Including unit tests documenting the behavior for multiple users with the same full name.

I'm not sure why it as such is needed, but it isn't wrong.

Coding style: it looks like most places in this file uses {} also around one line ifs.

This is the limit of my perl knowledge.

Who sets BUILD_SHARED_LIBS ? I don't see anything in this cmake stuff doing it.

doesn't add_library default to STATIC if not specified? Else we should add that keyword there rather than install the libs. They are just a convenience bundling of related bits in the source tree, not actual libraries.

Let's try poke @mlaurent and see if he has any comments.

In D17152#366536, @ngraham wrote:

What I'm saying is that I think it makes more sense to communicate your preferred email workflow to KMail rather then to Plasma. I'm open to being convinced out of this point of view, of course. :)

In D17152#366525, @ngraham wrote:

In D17152#366495, @anthonyfieroni wrote:

So i like second one, it can still be fixed to 10K+ because it does not matter above.

I'll try with + for a couple of days.

@ngraham I use kmail ... but I have hundreds of thousands unread emails without counting my spam folders. (commit lists, bug lists, ...)

@ngraham It might be that many notifications != no notifications in some cases, but at least in the case of emails, with many unread emails basically doesn't make a difference.

Very much an improvement. But we need a license file.

Did you imagine something like what I changed in D16137#343060 (the latest change) ?

from a source & binary compatibility point of view, this looks great.

What's the origin/license of these files ? How are they created ?

In D15715#330743, @astippich wrote:

since I'm still unsure about those things: the q2t function was not declared static, but was never exported. It is still safe to remove, right?

Looks good to me. A nice and simple way to drastically increase test coverage.

@fvogt mostly a unit test that ensures that all these 3 codepaths in slotData works. I'm not sure that the current unit tests ensures that. I might be wrong though.

Though unit tests would be nice. It looks like something that quite easy could break if next author is not careful.