KDE-CVS-Digest for January 10, 2003

The latest KDE-CVS-Digest is now available. This issue features news about Apple using khtml in Safari and the subsequent merging of Apple's contributions, support for RDP in the KDE Remote Desktop Client (krdc) and numerous bug fixes and new features. Enjoy!

Dot Categories: 

Comments

by Caleb Tennis (not verified)

the new format looks very nice, Derek. Great job as always.

by David Nielsen (not verified)

AWESOME job Derek... I can't thank you enough for these summaries.

by AC (not verified)

Whoops... The page doesn't render properly in Moz 1.0. It gets cut off halfway through the Konqueror bug fixes by a (malformed?) HTML comment.

Renders fine in Konq tho (of course).

by AC (not verified)

works for me with mozilla 1.2.1

by AC (not verified)

I guess I should upgrade. Still, those HTML comments should be escaped with < and >. They are in one place, but not another:

Dirk Mueller committed a change to kdelibs/khtml/html

handle <!--> and <!--->
CCMAIL34302[email protected]
Backported to KDE_3_1_BRANCHhandle (backport)

3.1rc4 Konq cuts the word "and" out in the last line, at least.

by Derek Kite (not verified)

Does it get screwed up after ....

Dirk Mueller committed a change to kdelibs/khtml/ecma

don't crash
Diff

Dirk Mueller committed a change to kdelibs/khtml/html

then goes wonky? If so, this is what is there...

make <select><option> text</select> work....

This is somewhere in the middle of the konq. bug section.

Could be the early mozilla got confused with the escape characters.

Derek
(I fixed the email address. speaking of escape characters....)

by Derek Kite (not verified)

Found it. Fixed it. I didn't escape the second instance of the comment.

Should be ok now.

Ack. Another one. . FIxed.

Derek

by AC (not verified)

Thanks. I love the CVS digest, BTW. :-)

by em (not verified)

Why do they use !(x == y) instead of x != y for example in http://webcvs.kde.org/cgi-bin/cvsweb.cgi/kdelibs/khtml/rendering/render_...
? (And why does the dot prevent me from posting html?)

by Justin (not verified)

Re html posts, the Dot has had them disabled for a couple years now, after I reported an exploit in their forum software (someone had used the hole to change all links on this site to some place unmentionable). This problem has either never been solved, or the admins don't want to risk it again.

by em (not verified)

That is unfortunate, because it leads to posting URLs as text (instead of hyperlinks) which in turn breaks the page layout in a most annoying way.

by Navindra Umanee (not verified)

I don't have any faith in any the HTML parsing code I've seen. It would definitely be nice though, so I'll probably give it a try (this weekend...).

by Roberto Alsina (not verified)

In konqueror, everything is an hyperlink!

Select the URL, then middle-click, IIRC. That should open the URL.

by me (not verified)

Probably to stress that a specific operator (==) is to be used. You can define your own == and != operators for objects, although != will be implicitly mapped to !(==) if it has not been defined. If you have large hierarchies of classes, you tend to lose track of which operator was defined in what class...

by David Faure (not verified)

In my experience != is NOT implicitly mapped to !(operator==).
So for your own classes, you need to define both ... or to use !(a==b), which is probably why khtml does the latter.

by L.Lunak (not verified)

Actually IIRC the reason for using !(a==b) is that libstdc++v2 (shipped with gcc-2.95) wasn't namespace clean and defined a public template operator!=, which took its arguments using references. And since some data members in KHTML classes are bitfields, and one cannot take pointers(i.e. also references) to bitfields, this caused compile errors(the libstdc++ operator was used instead of compiled generating a "normal" one).

by Michael (not verified)

The KDE-VCS-digest mentions that some of the Safari changes are incorporated into Konqueror from KDE_3_1_BRANCH. It is planned to incorporate all the Safari changes to Konqueror before releasing KDE3.1. Would be a good idea, I think, even though this may postpone KDE3.1 a little further.

by Debian User (not verified)

Each changeset will be judged independantly. Some are trivial to oversee and yet have an impact on the performance, those are likely just commited.

Everything that is a harder to oversee change, but makes sense will be commited to HEAD only, probably backported later, when it proved working good enough, for a 3.1.1.

And everything else, that is architectural, will see only HEAD and 3.2

That's the natural way to do it and I guess, the KDE people will only do it better.

Regards, Kay

by Derek Kite (not verified)

>incorporate all the Safari changes to Konqueror before releasing KDE3.1

Some are going in. The easy ones. And the crash fixes. The developers are impressed with the work and using as much as possible as quickly as possible. Someone even told Dirk to go to bed :-)

IIRC the Apple people were working from a 3.04 base. So there is a divergence in the two trees which will take a while to sort out.

I doubt if 3.1 will be delayed any further.

Derek (IMHO)

by anon (not verified)

Is there anyone working on intergrating the Apple changes to khtml?
I assume Apple won't do it,why should they ;)

by AC (not verified)

Have you read the other comments? :)
(yes)

by Debian User (not verified)

To me it came as a surprise to learn that Apple
would not only provide it, but also help merge
it.

After all, the name is Apple.

But if you think of it. While keeping it of course
secret to gain an edge over others (who? would
Sun even try to learn from them?!), after it was
reveiled, it all comes down to cost.

And maintaining a branch costs more than
incorporating the tested KDE releases with
all merges.

So they can focus on the UI. And that's what
Apple believes they can do best.

Seems Ok to me, Kay

by Andy Goossens (not verified)

Heh, it's nice to see your own commit showing up in KDE CVS digest.

One small remark though: my commit originally contained the text "Vladimir is really fast in finding and fixing reported bugs". It looks like those pseudo-tags were lost. Oh, and linebreaks made my commit more understandable :-)

You know, last week I was thinking it would have been great if one of my commits did ever show up in the CVS digest. If only all my dreams could become reality... ;-)

by Derek Kite (not verified)

> originally contained the text "Vladimir...

I guess I took them out too quickly. I'll put them back.

Funny I've never run into this until now. Maybe a plot to keep me on my toes.

Derek

by Dan K. (not verified)

Derek: there's so few comments here, I just want to make sure you know that's because it's so comprehensive and so useful. Wonderful job!

by beergeek (not verified)

Hey Derek, just another post of praise here. Thanks so much for the hard work. I really enjoy the weekly updates!