hi,
these days I was encountering a problem, when I was going through 2 veeeeeeery big hashes to find out which of them have similar keys (not really same keys, just slightly different). I thought, that the easiest way would be to go through both hashes in 2 for-loops and check them out one by one. Unfortunately, this approach took about 2 minutes to go through (2 minutes seem like an eternity if you are staring at a terminal).
My first approach looked like this:
for my $a (keys %b) {
for my $c (keys %d) {
if ($c =~ m/^$a.+$/) {
print "bingo!\n";
}
}
}You see, the keys in the second hash (%d) always look like the keys in the first hash (%b) with the only difference, that they contain the key plus some other stuff. So if $a looks like “abc”, $c looks like “abc123″. Now, if you have 2 hashes with – let’s say – 5000 keys and values, you can easily calculate how long this could take…. (5000 x 5000 = 25.000.000) … as I already said: it takes FOREVER!
Now, after some searching around I found a very nice solution on the Linux and Unix Forums from a guy called “pludi”. He recommended a grep-approach. Here is now my remodelled version:
for my $a (keys %b) {
for my $c (grep { /^$a.+$/ } keys %d) {
print "bingo!\n";
}
}Voila. Instead of taking 2 minutes, the loop now runs through in 15 seconds. 
I love Perl 
because I feel the need to do this:
click
just started playing with my new motorola milestone with android 2.1.
I had no idea how much fun it is to write perl-programs for this nice little device. I use the ASE (android scripting environment) which also supports python, lua, ruby …
You can get it on their website.
I guess I will write some programs soon to see what’s possible. The only thing that sucks on the milestone is, that the dollar sign is only available via touchpad. That’s REALLY annoying. :/
ok, today I finished my work at the fancy little script I mentioned yesterday. You can download it now here. Because wordpress for some reason doesn’t allow me to upload files without a filename-extension (grrr) it now has a txt-ending. please just ignore the extension …..
you call the script like this:
$ ./oebbquery "Innsbruck Hbf" "Wörgl Hbf"
Connections from Innsbruck Hbf to Wörgl Hbf:
Departure Arrival Duration Changes Train(s)
15:13 15:59 0:46 0 rex
15:28 16:24 0:56 0 s
15:54 16:28 0:34 0 oec
Please be sure to use the correct names of the stations. for example: “wörgl” doesn’t work, but “Wörgl Hbf” works. You have to use the names provided by the oebb-website.
the output looks now like this:
have fun!
hi guys,
I just spent 2 hours on a little project here. With this little script written in Perl by using LWP you now have the possibility to view train-connections from A to B in your terminal. You can then use this output to process it with conky or just any application you want. Here is a screenshot:
As you can see, i used it to display it with conky by executing it once every 600 seconds.
As soon as I have cleaned up the code to make it easier to read, you can download it here.
I can’t tell you how often I see scripts and programs which contain a huge amount of echo, print or whatever calls to produce output. It is ok if there are one or two lines to print out, but it becomes inefficient when used explicitly. In shell-scripts and Perl for example you could use so-called Here-Documents. Never heard of them? Well, here some examples:
example: bash
#!/bin/bash
# usage text
usage() {
echo ""
echo "This is the usage of blabla"
echo "Parameters:"
echo " -f ... Parameter 1 blabla"
echo " -g ... Parameter 2 blablubb"
echo ""
}Well, here you can see that there are a lot of echo-calls (often there are MUCH more … :-/). Now, we learn about an exciting new way to shorten this. We use “cat”:
example: bash better
#!/bin/bash
# usage text
usage() {
cat << EOUSAGE
This is the usage of blabla
Parameters:
-f ... Parameter 1 blabla
-g ... Parameter 2 blablubb
EOUSAGE
}So, instead of using tons of calls, just do it with one “cat”. Maybe it seems a bit pedantic of me, but this example makes more sense if you encounter more text (like HTML-output) which is printed out to a document with thousands of “echo”-s.
You can do this in Perl similarly:
#!/usr/bin/perl
# usage text
sub usage() {
print <<"EOUSAGE";
This is the usage of blabla
Parameters:
-f ... Parameter 1 blabla
-g ... Parameter 2 blablubb
EOUSAGE
}or in Ruby:
#!/usr/bin/ruby
# usage text
def usage
print <<EOUSAGE
This is the usage of blabla
Parameters:
-f ... Parameter 1 blabla
-g ... Parameter 2 blablubb
EOUSAGE
endI guess you can do something like that or similarly in nearly every language. So PLEASE, if you have much text to display in your scripts, use Here-Documents.
Posted by ap0calypse on 28. September 2009 – 13:05
hi,
just wanted to give you a small update about how things are going with the current kernel (2.6.31.1 so far) on the 901 model from Asus.
Graphics:
The intel chipset ( Intel Corporation Mobile 945GME Express Integrated Graphics Controller ) works very well and even KMS works like expected. There were some minor issues that required to build a initial ramdisk in the past, but now that’s not necessary anymore. I don’t know if that’s because of a bugfix or just some lucky configuration of my kernel. 
Wireless:
WiFi works with the rt2860 module from /staging.
Bluetooth:
I managed to use my cell phone as a modem device. It wasn’t very difficult. I’ll write a small article soon how to get this done. wvdial and rfcomm do the dirty work anyway …
I also bought a very nice little IBM thinkpad some days ago which I am currently tweaking to fit my needs. Screenshots will follow. Here is one from my eeePC:
Posted by ap0calypse on 18. September 2009 – 08:58
Filed under general, linux
Ok guys, I figured out how to fix the issue I wrote earlier about. Sometimes KMS didn’t work on boot-up and I didn’t know why. After some research I found the reason. The intel-agp module has to be loaded before the intel driver gets loaded. Sometimes (I don’t even know why) the driver got loaded earlier than the intel-agp and KMS failed. My simple fix is to create an initial ramdisk which loads intel-agp. So far it works quite well … The error didn’t appear again.
Here are the steps for 2.6.31:
# create initrd
mkinitrd -c -k 2.6.31 -m intel-agp
OK: /lib/modules/2.6.31/kernel/drivers/char/agp/agpgart.ko added.
OK: /lib/modules/2.6.31/kernel/drivers/char/agp/intel-agp.ko added.
3224 blocks
# this now created a file named initrd.gz under /boot
Next, you have to include this line to you lilo.conf, under your “root = blabla” line:
initrd = /boot/initrd.gz
and run lilo to update. You have to comment out the “vga = xxx” line too by the way.
I hope this helps some of you. Enjoy!
Posted by ap0calypse on 10. September 2009 – 14:17
Filed under general, linux
today I tried out KMS for the intel-chipset on my eee 901. It results in reduced flickering when starting X and a much higher resolution when booting up. So far, I’m glad it works now. The only thing that sucks is, that there is something wrong with KMS when I reboot via 3-finger-salute. It works perfectly when I type in “reboot”, but it does not come up in KMS again if I reboot via [Ctrl] + [Alt] + [Del] … strange.
Config for kernel 2.6.31:
Device Drivers --->
Graphics Support --->
Direct Rendering Manager (XFree86 4.1.0 and higher DRI support) --->
<*> Intel 830M, 845G, 852GM, 855GM, 856G
[*] Enable modesetting on intel by default
Today I found a very nice ncurses-based calendar tool named calcurse. At first it is a bit hard to get used to it, but once you figured out how it works, you’ll love it.
Friends of console applications will love it anyway . I have always been searching for a reasonable console tool for this task. Now I found one
Here is the link to the website:
calcurse
I’m using it as my primary calendar application to get an overview of my daily tasks and events.
Here a screenshot of me entering my university lessons 
:
entering stuff in calcurse
