Testers wanted

The last few days i have been trying to decompose pdftk into a library plus a binary that uses it instead of a monolithic binary so that we can use that lib to implement wish #86787. I have almost finished but i would like that some testers tried the package to see if the autotools thing I've created works (pdftk was using hardcoded makefiles) and if pdftk binary i produce has the same behavior old one had (you have to be a previous pdftk user to know that ;))

The package can be downloaded from http://kgeography.berlios.de/pdftklib.tar.bz2 and to test it you need to do the usual

./autogen.sh
./configure --prefix /usr
make
make install

Keep in mind that --prefix /usr may not be the "right" place in your system and that make install will overwrite your old pdftk so backup it elsewhere

So please if you try it leave a comment saying the system you use and if it worked or not.

KDE hidden powers

Today i discovered one of the lots of the powerful things you can do thanks to kde. Imagine you are writing a script to decompress a tar.bz2 and install it somewhere, imagine the target user may not even have bzip2 installed but you know for sure he has kde installed, what do you do?

kfmclient copy tar:/path/to/the/file.tar.bz2 /path/to/the/new/destination

And voilà, the tar.bz2 is descompressed inside /path/to/the/new/destination :-)

Mad at bash

Something strange is happening with my shell, and it is bad because it makes kde configure fail, let's see if someone can help me and tell me why

x=`gcc -print-prog-name=ld`
y=ld
echo $x
echo $y
if test $x = $y; then
        echo equal
else
        echo not equal
fi

gives me this result

$ ./test.sh
ld
ld
not equal

I'm deseperate :-(

Update: I've fixed it now, it was a hack i did to colorgcc to make it work with unsermake that was making gcc return "strange" things. Thanks to all that ofered help both here and in #kde-devel

More on CIA

You may have noticed that the #kde-commits channel is working again as http://cia.navi.cx/stats/project/kde is listing the commits in the recent messages part. You may have noticed too that for example http://cia.navi.cx/stats/project/kde/kdebase is not getting the kdebase messages, that is due to the script Dirk installed on the KDE subversion server is not the best cia people provide. The problem is that the best one needs regexps to know where each commit really went, so if you want this to work, please have a look at http://cia.navi.cx/clients/svn/ciabot_svn.py and get the correct regexps for each project and subproject. If you make it work do not hesitate to contact me or Dirk so we can try to make it work.

KDE and CIA

No, the intelligence agency is not messing with KDE, this blog is about CIA Open Source Notification System. It provides a web page and some bots like the one present at #kde-commits where you can dig info about commits of KDE and a zillion other different projects.

With the change to SVN the #kde-commits bot stopped working, i contacted the #cia guys and they said we were in our own, either we helped debuging the mail parser or installed one of the cia clients on our SVN server.

He wanted us to debug the mail parser because "I only added KDE mail parser to help CIA bootstrap (read get more exposure)", so i had a look at the mail parser and everything seemed fine, after all the SVN commit mails are 99.99% similar to the CVS ones. So as the CIA server itself was not working because they were rebuilding the DB i thought i'd wait and see if it worked.

Today it was clear the mail parser was not working (#kde-commits bot was still silent) so i entered #cia again and after some messages i saw

<CIA-9> micah * r8234 cia/mail/ (dot_procmailrc filter_kde.py): KDE's filter script no longer works properly. Remove it, since it's easier to just give up the special treatment and have them install a real client script.

So all the CIA and #kde-commits fans you better convince our admins to install the svn client if you still want to see that services work.

For me it has been clearly demostrated that we got used by CIA when they needed some exposure and now that they don't need us anymore nothing stops them to throwing us away, IMHO not a nice move.

UPDATE: Dirk fixed it, so thanks a lot, next time coolo says "no" i'll speak to you before making it a state afair :-D

Poppler

You may have noticed that recently xpdf source code has had some code auditing and some problems where found in it. And you may have noticed that meant lots of new releases, xpdf, gpdf, kpdf, koffice, cups, evince, latex(IIRC) and some more, why? you may ask. Because xpdf has not a shared lib so all the projects that want to use xpdf code have to include it in their sources. That obviously sucks for many reasons, a shared lib will be obviously better. That is where poppler comes in. Poppler is a fork of xpdf sources that installs itself as a lib other programs can use. Currently only evince is using it, but Brad Hards has some patches to make the pdf kfile_plugin use it and i've sent the poppler mailing list some patches we had in our (kpdf) sources and i hope they will be accepted, so the next time a vulnerability is found in xpdf code you'll have to update less things, a good thing (TM)

For those who fear a fork

I just wanted to inform you that a few moments after the DRM code was added to KPDF we added a configure switch so if you do

./configure --enable-kpdf-drm=no

you still get the old behaviour. IMHO that rules out any possibility of a fork.

Mantainership is a heavy rock

Since i took kpdf mantainership I've been quite proud of the changes Enrico, Tobias and me have made to it. But all things have its bad faces, and kpdf development have had 2 of them.

First of them happened when i introduced the possibility of running external programs on a user click on a link if the pdf specified this, that was put down by kde-cvs list readers as they said it was a security problem even if we added a dialog asking the user if he wanted to execute the command or not. IMHO that can be an interesting feature in some cases like when doing a presentation, but i can understand the security implications

The second has happened today when Enrico and me have added the DRM handling to PDF, that means, if the pdf says you can not print it, we don't let you do it, and some similar things. That has caused some controversy on #kde-devel as people said that this removes freedom from the user side and that there is no need for us to implement it. IMHO there is a need for this, i mean we don't go on opening password protected files without asking for a password, if the author put a password on it it was because he had good reasons, so if the author decided he did not like his pdf to be printed we do the same as with the password protected the file. Maybe you want to print the file, but should not blame kpdf for not being able to print it, blame the author for being an asshole.

Well i suppose being the mantainer of a program means you have to be prepared for such discussions.

Firefox spreads ... non free pdf viewers

If you have downloaded http://www.mozilla.org/press/nytimes-firefox-final.pdf you may have surely noticed that no free pdf viewer can display it, not xpdf(so no kpdf nor gpdf), not gv(so no kghostview nor ggv)

Reading the comments on the news item seems nodoby cares for it ... But i do, why is one of the jewels of the Free Software promoting non Free Software pdf viewers, i'm sure it could have been generated in a way that it looked the same and worked on free viewers.

Update: It seems it works with gs 8.15, the problem is that CUPS only comes with 7.07, any idea if anyone is working on CUPS? Will they ever update their gs version?

When bugs cannot be reproduced...

I hate it, you get a bugzilla mail saying someone have found a crash on the program you mantain, you go to bugzilla and see they have a way of reproduce the crash, open a document and kaboom, kpdf is dead. Then you download the documents, open them and ..... nothings happens :-S

What does that mean? I don't know, but that has already happened twice that week with bug 93352 and bug 93517.

Free Software community rules

This week i received another map for KGeography, that means 5 of the 9 available maps have been made by contributors different than me, which would be rather strange since the program is quite simple and not even finished, but it is not strange, why? Because Free Software community is big and active, so i get maps and translations to more than 10 different languages for my app.

To everyone that makes this possible.

THANKS

Introduction

Ei! My first blog ever in English and about KDE :=)

I will introduce me, my name is Albert Astals Cid, i suppose most of you don't know, but Spanish people have two surnames so Astals is my first surname and Cid my second one. My KDE login name is aacid and i usually hang on #kde-devel as TSDgeos.

My KDE related work has two sides, the first one is as a member of the Catalan (ca) translation team and the second one is as developer. As developer i am currently mantaining kpdf with help from Enrico Ros, also have some programs on kdenonbeta like kgeography, kfly and ktetrinet. Finally as annma usually says in her blog i like to help where i can.

As a news flash i'll say that the kpdf icon contest already has a winner and we're only waiting for the right time to announce it.