Blog
Page: 0 ... 5 ... 10 ... 15 ... 20 ... 25
i.File
Date: 23/6/2004
I've been hacking away on i.File a bit, fixing things and adding bits and peices. One of the more interesting things I'm working on is a column view, where instead of displaying the files in a list, the files display in columns. This mimics another useful view from Windows Explorer that none of the Linux file managers get right.



On top of that I realised that drawing was slowing down with a lot of text on the page. So I clip it against the current dirty rect and that sped things up a huge amount.
(0) Comments | Add Comment

Mozilla: Love/Hate Relationship
Date: 15/6/2004
My usually whining and praise for Mozilla is currently in 'whining' mode because I can't believe how long they are taking to fix XFT support. Despite the bugzilla report it's not fixed in Mozilla 1.6 xft on XFree86 4.3.0. Some what is it, like years after it was reported.

The work around is simple, manually select the right charset in the menu and all returns to normal. But I don't want to have to do that all the time do I?

Well at least I don't have to worry about adware installing itself. Hahaha if you use IE you're asking for it!
(0) Comments | Add Comment

KDE == 70MB !?!
Date: 10/6/2004
Wow, you really don't know until you look. After reading the nice article I decided to check out the memory usage of the various components I currently have loaded in Redhat9:
  • Scribe: 41mb (yeah ok, I got a big mail folder)
  • Mozilla: 30mb (fine!)
  • X: 21mb (pretty good actually)
  • KDE: 70mb !!!! (being the sum of all the kdeinit tasks and kalarmd)
It seems like all the kdeinit tasks have some common overhead. Does anyone really know what I'm spending my MB's on here? I've also seen it go up to around 140mb at times.

		i.Task Process Listing

Name                 PID   Cpu  Cpu Time Mem     Priority User    Threads Command                                   
-------------------- ----- ---- -------- ------- -------- ------- ------- ----------------------------------------- 
scribe               1478  0.0% 0:24     41.01 M -5       matthew 2       ./scribe                                  
mozilla-bin          2755  0.0% 0:14     30.48 M -5       matthew 2       /usr/local/mozilla/mozilla-bin            
X                    29850 2.7% 5:34     20.23 M -5       root    1       /usr/X11R6/bin/X                          
kdeinit (kicker)     32133 0.0% 0:07     10.82 M -5       matthew 1       kdeinit: kicker                           
kdeinit (kdesktop)   32131 0.0% 0:04     9.40 M  -5       matthew 1       kdeinit: kdesktop                         
kdeinit (knotify)    32125 0.0% 0:00     7.67 M  -5       matthew 1       kdeinit: knotify                          
kdeinit (kded)       32108 0.0% 0:29     7.39 M  -5       matthew 1       kdeinit: kded                             
kdeinit (kwrited)    32136 0.0% 0:00     6.70 M  -5       matthew 1       kdeinit: kwrited                          
kalarmd              32138 0.0% 0:00     6.24 M  -5       matthew 1       kalarmd                                   
kdeinit (ksmserver)  32128 0.0% 0:00     6.23 M  -5       matthew 1       kdeinit: ksmserver                        
kdeinit (klauncher)  32106 0.0% 0:00     5.65 M  -5       matthew 1       kdeinit: klauncher                        
lgiide               3811  0.0% 0:00     5.37 M  -5       matthew 2       /home/lemon/matthew/lgi/Lgi/LgiIde/lgiide 
kdeinit (dcopserver) 32103 0.0% 0:00     5.18 M  -5       matthew 1       kdeinit: dcopserver --nosid               
kdeinit (Running...) 32100 0.0% 0:00     5.16 M  -5       matthew 5       kdeinit: Running...                       
itask                3824  1.0% 0:00     4.91 M  -5       matthew 2       /home/lemon/matthew/lgi/i.Task/itask      
xfs                  3095  0.0% 0:48     4.74 M  -5       root    1       xfs                                       
gconfd-2             1733  0.0% 0:00     2.38 M  -5       matthew 1       /usr/libexec/gconfd-2                     
ntpd                 2863  0.0% 1:36     2.35 M  -5       root    1       ntpd                                      
oafd                 3534  0.0% 0:00     1.64 M  -5       matthew 1       oafd                                      
rcd                  3122  0.0% 1:14     1.45 M  -5       root    1       /usr/sbin/rcd                             
artsd                15025 0.0% 0:20     1.32 M  -5       matthew 1       /usr/bin/artsd                            
smb                  15447 0.0% 0:07     1.01 M  -5       root    1       smb:                  
ssh                  32156 0.0% 0:00     992 K   -5       matthew 1       ssh                                       
[snip]

	Cpu: 3.7%
	Mem: 506.43 M
	Tasks: 27 / 92
(0) Comments | Add Comment

One Less Proprietry Format
Date: 9/6/2004
LgiRes, the resource editor for Lgi applications can now read and write directly to valid XML. This means that the "lr8" format that it has been using for the last year or so is going the way of the dodo. However it was pretty close to XML already, I just needed to make a few changes here and there to be fully compliant.

For those of you that edit the translations by hand it shouldn't make too much difference, but I'll have to update the translation instructions to talk about the differences required to conform to XML.

The upside of course is that with the resource file being XML all sorts of tools can be applied to the data. I'm not sure what benifit that will be yet, but it sounds good in theory right!

Just remember: Proprietry == Bad.
(0) Comments | Add Comment

Mdi Windows On BeOS
Date: 7/6/2004
On linux and windows, child views (HWND etc) can overlap and keep their internal order. This makes implementing Mdi window "Real Easy(tm)". On BeOS however overlapping child windows are not garenteed to paint in any order so you have to implement overlapped windows FROM SCRATCH!

ARRRRRRRRRRRRRRRRGGGGGGGGGGGGGGGGGGGGGGG!!!!!!!!!

Yes, you have to map mouse, keyboard, painting and everything else into virtual windows. Oh boy I'm so not looking forward to getting that right.
Update: I just found a class lib for doing MDI windows on BeOS. Which has given me some ideas on using real BViews. Which should make life a lot easier.

I think by clipping the output of each MDI child and also remapping user input events by hand I can work around the problems with overlapping BViews. Rather than implement a virtual windowing system.
(0) Comments | Add Comment

Why Windows Will Die
Date: 7/6/2004
As a lot of geeks can atest to, keeping a family member or friends boxen running is just par for course. In the old days it involved helping them with hardware purchases, rebuilding after HD's die, installing applications for them and even do some basic computer training.

But in more recent times it's turned into configuring firewalls, getting rid of viruses, setting up secure email and update the virus definitions.

However I'm finding out the hard way that Windows is so insecure that I'm basically going to HAVE to migrate all the computers I have to administer over to Linux simpily to survive. The internet is such a acidic environment for the baby skinned Windows that it doesn't stand a chance. I don't have time to be constantly fixing compromised machines anymore.

There are software solutions for this, primarily software firewalls and av software but it's a moving target. And they don't always work as advertised. And in any case the av software is always out of date with the current batch of viruses.

Where as Linux is secure out of the box and requires a great deal of very skilled knowledge to compromise.

People won't go to Linux because they want to, they'll go to Linux because their tech friends get sick of re-build compromised Windows boxen.
(0) Comments | Add Comment