KDE Developers don't test their own programs...

Clemens Wacha clemens.wacha at gmx.net
Fri Mar 19 17:45:43 GMT 2004


I just installed KDE 3.2. A lot things got better. I totally like it.
But most of the problems still remain.

Am Mit, den 17.03.2004 schrieb Nathan Toone um 16:30:

> > I.e. KDE Addressbook. Don't even think of trying to use the csv
> > importer. its totally useless to me because I cannot import my adresses.

The Addressbook did change alot. It look great now. It can import more
than one vcard at a time. But this still doesn't help me because I have
lots of names with umlauts (iso8859-1 charset). The addressbook doesnt
respect that. So it remains totally useless for me. Its usable for US
people.. the rest of the world still has to wait.

> Haven't tried it - so I can't comment.  It may be that your csv file is not 
> 100% conforming to specification though.  I don't discount the possibility of 
> bugs, but I do know that for the most part KDE applications try to adhere to 
> previously defined standards (and in my experience, they have done a good job 
> of that - where KDE messes up is when you try to do something with a 
> "nonstandard" file.)

Yes this is true. Anyway my csv file had a job description field that
looks like

"a very
long job description
that lasts over more than one line. Thats why it is encapsulated
in \" \" "

the addressbook doesn't properly import newlines that are "saved" within
"". I don't know of any standard about this but it must be a very stupid
standard if it doesn't allow you to use newlines between "". And don't
tell me that this is difficult to do in C/C++. I have written such
routines on my own more than one time. Hey don't get me wrong. I don't
say that the coder is stupid. He just didn't think of it when he wrote
the code. This happens very often.

> > mime type detection might work for 95% of stupid newbies that never
> > change anything. I tried to add .sid Tunes and Gameboy roms. But its not
> > possible because both of them are recognized as octet-streams... That
> > might be true but they still have different file endings.. well but
> > konqi doesnt care.. hit is hit :-)
> [explanation of difference between MIME and file extensions]

Yes I know all this. I am using computers since almost 15 years now. I
switched to linux 4 wears ago. Before that, I had MS-DOS, windows and
now i am also using Mac OSX at home. I am using Solaris at work. If you
want to detect a file type you can do it the lazy/fast way by looking at
the extension. You can also do it the save way by looking at the content
(which takes more time though).

the Problem I have with kde is not my misunderstanding in filetype
recognition. It seems like KDE does the following:
     1. Check file stats (dir/file/pipe/whatever)
     2. if its a dir: say its a folder and stop
     3. if its a file:
      * check mimetypes
      * if no MIME Hits check extension
      * if no extension hits say: Unknown file

In my case I have two different files with different extensions. KDE
detects both of them as MIME type octet-stream which is not really wrong
but also not very precise. Now kde searches the first entry of known
filetypes that hits the mime type. Actually it should also check the
extension to make sure there is no other known filetype that also hits
the file content. We have 2 problems here:
1. MIME type detection is not good enough
2. Since point 1 can never know all future filetypes it should always
check the extension too.

> - - not by file extension.  Extensions are WAY too easy to change.

So what? The computer is my servant. It has to assist me as good as
possible. Maybe the user WANTs to change the filetype.. so why make it
more difficult than needed?

I would implement it the other way round. first look at the extension.
If there is one assume its correct. If there is none check for MIME
type. Of course this should only be done in konqueror. Application that
open files should always check for MIME correctnes.

[PDF troubles]
I have never heard of KPDF before. Mainly because there is no debian
package for it? KGhostview became much better in KDE 3.2. Rendering is
as slow as before because its done by ghostview. This is mainly a
problem of the way ghostview works. It creates an image of the content
and hands this image over to kghostview. This is bad because PDF consist
of text and hyperlinks and they normally have a structure (table of
contents). Well.. the actual problem is that most pdfs are converted
from ps..  

> > [Bug reports]
> It's a small price to pay for a free operating system and desktop environment.  
> Besides, if you don't log the bugs, you dont' have a right to complain about 
> them.  It's kind of like voting - if you didn't vote, you don't have a right 
> to complain about the current administration, because you didn't *DO* 
> anything about it.  (Now if you voted for someone else, you have all the 
> right in the world to complain about it.)  If you want to lend credibility to 
> your complaints, you have to do everything you can do as well.

You are right. First of all I didn't knew of kdes bug reporting tool.
Its exactly what I was looking for. Secondly there are people that tell
me: Do it better or shut up. I don't like this because I say: Do it
right or leave it. I often forget that my demands are very high :-)
Linux is really great for a free operating system. Although on the
desktop its not yet a real competition to Windows or Mac OS. There is
still to much handwork needed to make it work right. But I am convinced
that one day I am happier with Linux and or KDE that with the other two.

Clemens Wacha           wacha at gmx.ch           ICQ:12620942

proudly running Debian/GNU Linux. See http://www.debian.org

This message is from the kde mailing list.
Account management:  https://mail.kde.org/mailman/listinfo/kde.
Archives: http://lists.kde.org/.
More info: http://www.kde.org/faq.html.

More information about the kde mailing list