[rekonq] rekonq 0.1alpha

[Paweł Prażak]

On Wed, Apr 22, 2009 at 2:20 AM, Andrea Diamantini <adjam7 at gmail.com> wrote:

> Here we are, as decided some days ago, I merged on mainline/master a lot of
> your commits.
> I really like to thank you all for your incredible work on. Anyway, without
> further ado, let's start seeing news:
>
> 1.
> As we are saying that these are technological previews I fixed compilation
> on
> Qt 4.5 removing all 4.4 workaround. That's because 4.5 is an huge step
> forward
> against 4.4, especially on the classes we are using at most (tabwidget,
> tabbar, webkit and so on..). I'm sure quite all are using 4.5, so it's a
> small
> issue..


Yes, this is good idea. It will clean up some parts of the code :)

>
>
> 2.
> There are also kde 4.2 derived bugs. (e.g. tab text) I leaved unaltered
> code
> that works with KDE trunk, but not (well) with 4.2


I think it would be better to leave it as we (me and avaddon) coded it,
because otherwise you'll get a bug on 4.2 or 4.3.

>
> 3.
> new download system and modifications in the Application class to better
> support it. QPointer(s) are really a great idea :)
> I also modified a bit download class and removed QTimer notification. It's
> my
> fault. If someone is interested in, he can re-enable that.


Timer is in place to make download progress bar possible.
QPointers helped to track the deletion ofo objects when going to background
download mode and back.


>
> 4.
> Side Panel. As requested rekonq 0.1 will have it. We sure have to work
> about
> the bookmarks panel (before releasing 0.1) and put it on.
>
> 5.
> Bookmark classes improvements. All apart from one method I didn't
> understand
> and I never seen used.


Some stuff there is in place to make future features possible.


> 6.
> Contextual menu. Here I did the major changes (in the code). I removed the
> unuseful (for me) QMap actions and provided a KActionCollection for all
> webActions.


We don't really need action collection (and it's heavier than QMap) and it
makes code cleaner and easier to read. There is no point in changing working
code just to use KDE class instead of generic container.

7.
> moveable tabs, that is gratis, via Qt 4.5 functions.
>
>
> I hope I'm remembering ALL changes I did. Please from now on, sync one your
> branch (master?) with mainline/master, test, code, suggest, do small
> themable
> commits and fill merge requests ;).
> I'd like to release 0.1alpha on next tuesday (28 apr) and from there on
> going
> in bugfixing and translation mode until 0.1 release (when? I don't know).
>
> From now on, if you don't have different ideas, we need to work on two
> essential aspects:
> - the bookmarks sidepanel
> - the focus problem
>
> About focus, I really cannot decide what is better to do, so I'm waiting
> for
> your suggestions (and your code ;) ).


What's with the focus?


> COPYRIGHT QUESTION
> I added a string on every rekonq file: the following
>
> * Copyright (C) 2009 rekonq team. Please, see AUTHORS file for details *
>
> This means we need to do also two things:
>
> 1. update Authors file (important)
>        here I wrote 3 columns: name, mail, role
>        You just need to add yours and perform a merge request.


I'm not sure about the role field, we don't have any roles, so it would be a
bit artifitial.

2. update main file (VERY important)
>        here you need to add... I'm pushing one example commented in
> main.cpp file.
>
> We can obviously remove the previous line and push all our names also on
> every
> file. I did this way just for convenience. Decide all together one way and
> just
> do it!!
>

IMO this is bad idea. It's not how things are done. We can have AUTHORS file
but leave the copyrights on per file basis.

I'd like also rise a concern about backporting changes. It doesn't make
sense, because it's easier to port your changes to our fork than vice versa
and you shouldn't port (especially don't change) code that someone write and
tested because you don't know why it was written this way and not another
and the possibility of introducing new bug is big. I've run a diff and I see
there are awful lot of things you didn'd ported and most of this is
important stuff related to bugfixes and missing features. A lots pasts of
code were revritten or heavy refactored and there is no sense in doing it
again.

Also IMO you can't make a release without unittesting (and unit tests I've
implemented are not even adopted in mainline yet), to make it right we need
quite a few new tests to make sure the code can be released to anyone (even
thou this is a preview).


Regards
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.kde.org/pipermail/rekonq/attachments/20090422/5d040fd2/attachment.htm