KDE Traffic #14 For 15�Jun�2001

Editor: Aaron J. Seigo

By Aaron J. Seigo �and� Rob Kaper

Table Of Contents


Welcome to KC KDE! Summer is here and KDE 2.2 is approaching. KDE CVS is in a feature and message freeze until 2.2 is released, which means that only bug fixes and documentation changes will be accepted for the next six weeks, barring any unexpected delays in the release schedule. Six weeks may seem like a long time to not see any new features appear in KDE2 after getting used to the pace of development since the 2.0 release, but this time will be well spent focussing on the stability and quality that KDE has been famous for since the 1.x days.

1. KDE League: questions and answers

4�Jun�2001�-�13�Jun�2001 (36 posts) Archive Link: "[kde-promo] Speaking of the KDE league..."

Summary By Rob Kaper

Topics: Promotion, KDE in the Enterprise

People: Darian Lanx,�Andreas Pour,�Mathias Waack,�David Faure

Darian Lanx asked:

Since I read the interview a moment ago, which I enjoyed, thank you tink, I was asking myself, where is the kde league?

Are there any officials of the KDE league reading this mailing list, if so which? Is there any spokesman or person the various independant promo projects and sites can get in contact with, so the kde league might take actions in favour of them?

Is there any centralized structure for requestion sponsorship or funding? Is there a certain procedure one has to follow, is there a website? are there certain prerequisits one has to fullfill in order to being considered by the League for "help"?

All those questions keep swirling in my head and I honestly wonder...where are they ?

Andres Pour answered all of these questions and explained the lack of visible activities of the League: "There isn't a policy in place at this point for sponsoring projects. We are currently finalizing the budget and once that is done we will move toward addressing issues such as whether a formal procedure makes sense and if so what its outlines will be." . This spawned a discussion about the purpose of the League. Mathias Waack questioned the usefulness of marketing:

KDE is the result of the hard work of hundreds of developers with the help of thousands of users. Most of them had never seen a single cent for her work. And now you are going to spend money to people just to tell other poeple they should use KDE? What kind of result could we expect? One, two or even more percent "market penetration"? Maybe. More feedback and developers? From poeple who didn't heard about KDE until now? Who could this be? Are you really interested in a bunch of bug reports like "my internet is not working with KDE"? Where's the gain for the KDE project?

David Faure explained:

The primary goal is to attract _companies_ to use KDE as a development platform, and to distribute KDE with their stuff (have you seen that HP signed with Ximian to make Ximian Gnome standard all their next machine/unix/whatever it is ?). That's the kind of things we have to work on, getting this kind of deal with big companies, otherwise KDE will go nowhere (compared to its rival). Yes, I agree, it's not users we should target directly with the PR. But directing PR at companies is important.

2. New Kamera Maintainer

5�Jun�2001�-�7�Jun�2001 (11 posts) Archive Link: "Anyone working on kamera?"

Summary By Aaron J. Seigo

Topics: Multimedia

People: Anders Oreb�ck,�Shawn Gorden,�Donghai Liu,�Waldo Bastian

Anders Oreb�ck asked, " I am just curious about if kamera will be fixed for the 2.2 release." Apparently the kioslave front end to gphoto2 was not even compilable due to being unmaintained. Shawn Gorden of theKompany replied, " Even though we did Kamera, we donated it to KDE and I thought it was part of the basic package and being kept up to date with the library changes as part of that move, at least that was the rationale. We haven't had anyone work on it since January though."

At this point Donghai "Alex" Liu said, " I have worked on that piece of code for some time, and it worked for me last night! There are still some problems in the code, i.e. you can't delete files but only get it; you can't get the files size, and kcontrol module still can't compiled and only test it on usb port....but basicly you can list the folder and get the pictures. Since gphoto2 was only beta release, there are some bugs in its libs. I don't think kioslave will work fine without some fixes in the gphoto2." A few posts later Alex asked who the maintainer was so he could send his patches to them. Waldo Bastian replied, " I think the question should be: Do you _want_ to be a KDE developer? :-) Just say the word and we set you up with a CVS account and kamera is all yours!" . Alex answered Waldo's offer saying, " Hmm, I found it's hard to say N word, but before saying Y word I checked the KDE 2.2beta schedule, I'm afraid I can't make it work right before beta1 release. Anyway, _YES_!"

3. Music Applications for KDE

6�Jun�2001�-�11�Jun�2001 (28 posts) Archive Link: "Guitar program"

Summary By Aaron J. Seigo

Topics: Multimedia, Audio

People: Anakrewn Meidis,�Guillaume Laurent,�Michael Goffioul

The rate of musicianship among the KDE developers (as with most programmers, it seems) is very high. One would therefore expect to find a large interest in music related programs and indeed there is. Anakrewn Meidis asked, " Has any body written a program for writing tablatures and notes for an instrument? If not is any one writing such a program. I intend to write the program but need to be sure that somebody else hasn't done this allready." Michael Goffioul suggested looking into NoteEdit, which was ported from pure Qt to KDE2. Guillaume Laurent also replied saying, " There is Brahms, which current version is in kde's cvs (in the kmusic component). We also hope to make a 0.1 release of Rosegarden near the end of the summer (/kde/rosegarden.sourceforge.net)."

Michael said he had experience with these programs and compared them to his experience with commercial programs saying, " I never got the same feeling (I mean ease of use) for music writing. Note that my primary interest is not to "hear" the written music, but to print it in a professionnal way. I mean I'm more interested in a powerful notation editor, than in the sequencing engine. When I tested existing tools, I had the impression that the primary interest was the sequencing feature (like in Cubase, which inspired Brahms). However this may have changed now. Finale or Cakewalk are really powerful tools to produce nice scores, and they are easy to use. That's why I would like to have such a tool under KDE. In that context, I found Rosegarden more useful, but the GUI was not very friendly." Guillaume reported that the Rosegarden project (rosegarden.sourceforge.net) was being worked on again with the aim of having something releasable by September.

The thread then turned to future directions in music applications for KDE2 as scores of musicians came out of hiding on the list and started discussing their thoughts on the matter. Ideas for high-end production quality tools were discussed which sparked development in existing and new projects. It will be interesting to see what comes out for KDE2 in the next few months for musicians.

4. KIllustrator and Karbon

7�Jun�2001�-�8�Jun�2001 (13 posts) Archive Link: "Future (KIllustrator, Karbon)"

Summary By Aaron J. Seigo

Topics: KOffice, Merging

People: Lenny Kudling,�Christian Tiberna,�Alex Neundorf

For any given itch there are often several Open Source applications that scratch it. The reasons for this seemingly massive duplication of effort are many, but one of the primary causes is that different developers take different approaches to the same problem. This results in very different programs that allow the user to accomplish similar ends. KOffice is no exception to this rule with KIllustrator and Karbon, two vector drawing applications under active development. Karbon developer Lenny Kudling posted to the koffice-devel list saying:

After finishing my degree i had a bit time again to work on karbon. In these screen-stots

you can see proof-of-concept-implementation (meaning: it works but isn't optimized at all) of 2 (out of hundrets of) issues which i don't like about KIllustrator:

I think now it's the right time (before i start to implement a gui and all these nasty algorithms) to think about this: should we merge karbon and killustrator or should we go seperate ways? The core-question is: "is KIllustrator meant to be a basic vector app or an app with advanced features which i _need_?"

I would be more than happy to put my energy into KIllustrator, but i won't if certain dogmas/implementation are not wished to be changed.

Christian Tiberna replied saying, " This is one *opinionated* way of seeing things. KIllustrator's (and Corel's for that matter) is another. Nobody will be able to say which is better. But I'm sure *both* are useful and *can* be accomodated in the same program. Like, e.g. Corel has a feature that lets the designer to convert a circle (defined by Corel's usual center-and-ray structure) into a path, exactly how you depict it in your screenshot. KIllustrator attempts to offer this too, but right now, turning a circle into a path requires 30 seconds on a rather fast machine, blocks KIllu's GUI for the mean time and it ends into a huge number of nodes plastered on the circle" ... " I believe karbon and killustrator should be merged. killustrator has the user base and the rather large feature selection, karbon has your very nice and new code and the potential to bring to the reunited app a devoted maintainer. Which would be great for KOffice in general."

This sentiment was echoed by the KIllustrator developers in general, including Igor Janssen, Alex Neundorf and Kae-Uwe Sattler. The resulting plan appears to be to merge the work Lenny is doing in Karbon into KIllustrator in the midterm as the changes to KIllustrator internals will be considerable as both projects continue addressing other issues during the feature freeze.

5. --enable-final

7�Jun�2001�-�9�Jun�2001 (6 posts) Archive Link: "What does --enable-final do?"

Summary By Aaron J. Seigo

Topics: Building KDE

People: Gioele Barabucci,�Pavel Troller',�Waldo Bastian,�Pavel Troller

A question that comes up fairly often from people who compile KDE from source is "what does the --enable-final configure option actually do?" Gioele Barabucci was wondering exactly this, causing him to post to the list saying,

am playing a bit around with --enable-final but I cannot see any improvement. From what I have seen --enable-finale does:

and nothing else. So, what is the --enable-final use ?

Pavel Troller answered Gioele's inquiry with:

I think that it can substantially REDUCE the compile time, because all the headers which are used in all the sources, are parsed only once, which is faster with compilers without precompiled headers (current GCC).

> - adds a -ftemplate-depth-99, what is this used for?

Snipped from "info gcc", `-ftemplate-depth-N'

Set the maximum instantiation depth for template classes to N. A limit on the template instantiation depth is needed to detect endless recursions during template class instantiation. ANSI/ISO C++ conforming programs must not rely on a maximum depth greater than 17.
Maybe in such a big code You need to increase the limit.

I'm using it and it makes the compilation about 30% faster (it' my feeling, I didn't make exact measurements).

Waldo Bastian added to this saying, " In general it will also allow the compiler to do better optimisations resulting in slightly faster/smaller code."

6. New Keyboard Shortcut Scheme

7�Jun�2001�-�10�Jun�2001 (5 posts) Archive Link: "List of 4-Modifier Layout"

Summary By Aaron J. Seigo

Topics: Keybindings

People: Ellis Whitehead

Recently a new keyboard shortcut scheme was added to CVS for use by those with a Meta key on their keyboard. It caused a bit of a ruckus at first as it was initially added as the default layout if a Meta key was detected. This was quickly changed back to the previous default set, but the new scheme remains as an option. What makes this new layout rather interesting is how it divides up the tasks between the modifier keys. Ellis Whitehead explained the new scheme in an email saying:

Here is a list of the more significant differences between the 3- and 4- modifier layouts.

Switch To Desktop # Ctrl+F# Meta+#
Show Window List Alt+F5 Meta+0
Pop-up Window Op-Menu: Alt+F3 Alt+Space
K-Menu Alt+F1 Meta+Space
Execute Command Alt+F2 Meta+Return
Window Close Alt+F4 Alt+Escape
Logout Ctrl+Alt+Del Meta+Escape
Close Document* Cltr+W Ctrl+Escape (not yet)
Walk Windows Alt+Tab Alt+Tab
Walk Desktops Ctrl+Tab Meta+Tab
Window Maximize - Meta+Plus
Window Max Vertical - Meta+Bar
Window Max Horiz - Meta+Equal
Window Shade - Meta+Underscore
Window Inconify - Meta+Minus
Minimize all windows - Meta+Ctrl+Minus
Kill Window Alt+Ctrl+Esc Meta+Ctrl+Delete
Show Task Manager Ctrl+Escape Meta+Backspace
Lock Screen Alt+Ctrl+L Meta+ScrollLock

In a later email he also expounded on his reasoning behind this scheme saying, " I've tried to make the key assignments as logical as possible, and spent literally hours drafting up tables of various possible schemes (considering some things which i'll have to wait until the next version to implement)... Of course, every scheme is only more or less arbitrary. ;) Anyhow, i think the some of these settings bring more consistency and are more 'learnable' than the usual. For example, there is a certain correspondence between the window menu and k-menu, and so it makes sense to use the same key with the appropriate modifier. Or the correspondence between quitting a dialog box, quitting a document, quitting a program, and closing the window manager (I like the 'escape' key for this). F#-key assignments often seem highly arbitary (although not in desktop switching). Regarding the desktop switching, i think i would like to change it back to using F#-keys so that the number keys could be used for jumping to a given program in the taskbar..."

7. KDE2.2beta1 Freeze

7�Jun�2001�-�12�Jun�2001 (29 posts) Archive Link: "Preparing for KDE2.2beta1"

Summary By Aaron J. Seigo

Topics: Release Schedule, KDE 2.2

People: Waldo Bastian,�Torsten Rahn,�Carsten Pfeiffer

Waldo Bastian announced the upcoming beta of 2.2 saying,

This is to remind you that next week we plan to have the first (and possibly only) beta of KDE 2.2. This means no new features as of now. Please report any problems that you encounter in current CVS to one of the mailinglists.

From Monday on, all patches should be posted for review to an appropriate mailinglist. Very simple patches can be included as CVS comment for review.

Some took exception to the idea that this might be the only beta for 2.2, saying that there were numerous grave bugs that needed addressing. Torsten Rahn answered these concerns saying, " We should really try to make this the final beta and take it serious. Otherwise we will definately end up with something that is not beta-quality. We already had this last time when originally it was planned to release a beta and it turned out to be an alpha-release (partly because features weren't in place and partly because people didn't take bugs too seriously). Trying to head for a final beta will certainly give a better quality and make us get faster near the real "final" beta than if we try otherwise."

The discussion revealed numerous show-stopper type bugs that were in the CVS versions of core programs such as Kicker and Konqueror. The consensus was to iron out these large, known bugs first before releasing the beta since, as Carsten Pfeiffer said, " The beta is to let users find new bugs instead of annoying them with known crashes."

8. Enhanced Konqueror Shortcuts

11�Jun�2001 (4 posts) Archive Link: "A more flexible solution for internet keywords"

Summary By Aaron J. Seigo

Topics: Konqueror, Keybindings

People: Andreas Hochsteger

After using Konqueror for a while with its terrific Internet shortcuts feature, one may find themselves automatically typing "gg:" in the location bar of a browser when wanting to do a web search only to be disapointed when it doesn't take you to Google. There was a limitation to this terrific feature, however: only one search term could be defined in the shortcut, making it difficult or impossible to use in conjunction with websites that want more than just a simple search term. Andreas Hochsteger felt the need to rectify this and ended up writing a patch to support arbitrary numbers of query terms. Andreas posted his patch along with an explanation of it, saying:

I've implemented a more flexible solution for the internet keywords which solves some problems with more advanced queries and provides a more flexible usage.

What was my intend to change the current scheme?

Last week I wanted to create a simplified query for telephone numbers with konqeror's internet keywords. But I had troubles to get this working the way it should, because the url to be generated looks like this: http://my.telephone.book/query?name=MyName&city=MyCity As you can see there are there are two different substitutions necessary, which can't be done by the current substitution scheme. It only replaces \1 by the whole (encoded) string of the user query.

I submitted a bug report (whishlist) therefore which can be read here: http://lists.kde.org/?l=kde-bugs-dist&m=99189305712959&w=2

Overview to my implementation:

Since I never really programmed anything with KDE/Qt beyond some simple "Hello World" programms, I wanted to try it myself. It's now finished and working well. I'm writing to this list, because I've got some remaining questions. Perhaps someone can take a look at it, if I've done everything allright.

All changes were done in the method "formatResult". Nothing else had to be modified which speaks a lot for the great design of kde! The parameters query and url of that method have been replaced to reflect the real semantics of that parameters. I've attached a diff for the file kuriikwsfiltereng.cpp and the whole, plain method filterResult (for a quick look).

Andreas went on to explain in detail how the algorithm worked and what his future plans for it were. Discussion on the technical details ensued, such as keeping compatibility with the current shortcuts implementation along with a few bugs that were discovered. Andreas posted several further versions of his patch to address these issues.

Sharon And Joy

Kernel Traffic is grateful to be developed on a computer donated by Professor Greg Benson and Professor Allan Cruse in the Department of Computer Science at the University of San Francisco. This is the same department that invented FlashMob Computing. Kernel Traffic is hosted by the generous folks at kernel.org. All pages on this site are copyright their original authors, and distributed under the terms of the GNU General Public License version 2.0.