[rekonq] Comments on IRC Chat 2009-04-18

[Paweł Prażak]

On Sun, Apr 19, 2009 at 7:52 AM, DeKay <dekaymail at gmail.com> wrote:

> OK.  After some discussions in IRC, I put a bunch of new issues into
> the google bug tracker.  I know Andrea wanted to stick with sf, but
> nobody else can stand it.  So in to google they went.
>
> It seems that the project is going to need better communication
> between the members if it is to keep making progress.  People seem to
> be taking the project in different directions.  There has been the
> removal of some functionality that seemed to have a good reason for
> being there.  I hope we can find a way to communicate the various
> ideas, show some flexibility, and agree on a common goal going
> forward.


Yes, good communication is crucial/critical. (I'm not sure what word is
better)


> So here are a few of my thoughts to kick things off, in no particular
> order:
> - don't put out a new release until the browser becomes more stable.
> There are a lot of bugs, crashes, and even some dangerous behavior
> right now (eg http://code.google.com/p/rekonq/issues/detail?id=25).
> The last thing you want is to give people a bad impresssion early on,
> because you'll lose many of those people forever.  I still think of
> Nautilus as crap even though I haven't used it for many years  :-)


I totally agree.

- improve communication.  This mailing list needs more activity.
> - pick a new name.


Maybe reKon ;D

- settle on a bug tracker


Google code / KDE bugzilla (if available)

- (most importantly) come up with some kind of roadmap.  The TODO list
> is relatively short term.  Stuff like this needs to be in there so we
> can aim for longer term goals:
>  * What needs refactoring and what doesn't?


Downloads've been refactored, they still need some love thou.

MainWindow and MainView are the biggest classes with lots of methods and
initialization.
We could break them into pieces, e.g.:
- make new class MainWindowPrivate with all the boring initialization and
stuff and leave in MainWindow only the necessary logic.
- divide MainView to get new class(es) WebTab (and if needed
MainViewPrivate)
WebTab is most important to make class responsibilities clearer, make
testing easier and prepare for future enhancements (like bookmarks:/
protocol widget or Chrome-like "home page" with recent pages and stuff,
etc...) - I have initial work donw in my experimental branch and I've
already backported some of the stuff.

UrlBar has been rewriten byt this is only initial stuff, we need to
reimplement KCompletion and KCompletionBox for starters an than move forward
to awsomebar-like funcionality (bookmark menu, tags, etc)

History needs some love (we can try to port latest Qt demo browsers history
manager), there are lots of bugs there.

 * Multiple window support.  Will it be added back or is it removed forever?


I think we need multiple windows support (is there any browser that don't
have it?)

 * Plugin / extension system


Plugins as in shared libraries are fairly easy to do (in fact rekonq in my
and avaddons branch is a shared library to make testing possible and to
allow prefetching by preload daemon)
As I said earlier,side panel should (will) be made as shared library.

But for extensions, IMHO the ideal thing would be to share common extensions
system with konqueror (they already have basic adblock), but I'm not sure
how it's implemented.


>
>  * Bookmarks


Bookmarks needs some further changes as well (new pixmap provider among the
others) and an BookamarkModel for the bookmark sidepanel and UrlBar
awsome-bar-like funcionality.
Bookmarks taging is essential IMO (with ability to search by tag from
urlbar)

We need RSS support as well, this is basic feature (integration with
Akregator and Google Reader is a must)


And lots of small things like:
Drag'n'drop doesn't work.
301 is not supported.
ctrl+f5 cache flush and reload
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.kde.org/pipermail/rekonq/attachments/20090419/67f5f7fa/attachment.htm