Got tipped off about a little hacking challenge over at Rusty Lime. The basic challenge involved an installation of Severn (Red Hat Linux 9.0.93, what later became the original Fedora Core), with a number of services running.
 
After prying around the system, getting a couple of hints from the organiser (see the competition comments page) and running nearly every known bug exploit script against the machine, a cryptic message on a phpbb forum installed on the machine pointed to an unsecure upload script, which allowed... wait for it... the uploading of PHP files. From there it was easy as running an exec inside a PHP script and claiming the prize!
 
Rodney (the organiser) has a great wrap up, outlining the system vulnerabilities and the winning strategy in more depth - check it out!

... because Google indexes everything :)
 
Last year while using Lejos on a uni project I wrote up a quick howto highlighting some common build problems with lejos. Seems it was picked up by someone else who found it useful:
 
Malboro College
 
Its always nice to help, even if its only in a small way!

For those poor souls who can't access their gmail accounts (it just hangs not doing anything) it is most likely to do with you having set it to automatically login. Due to some unknown reason (db recycle? maybe the addition of the IMAP feature everyone is talking about?) these cookies are now invalid.
 
To solve this problem, clear your google.com cookies (Firefox: Tools->Options->Privacy->Show Cookies, for IE see this article) - and you should be right.
 
Happy gmailing!

Every now and then one must just sit back and say "enough!". Today was one of those days, between building a sukkah in the rain and running around between uni and work - I needed a break.
 
The break came in the form of a Perl script that checks which user directories (under a *nix system) have web pages. Basically, it's a very basic script that first scans the defined home directories for users and then checks each user's directory to see if they have modified their default index.html (if it exists).
 
I haven't actually done a lot of Perl development lately so it was a refreshing change - I even (re)discovered the uses of map and grep, which are refreshingly similar to Haskell's list comprehension functions map and filter.
 
You can see a live demo of the script at uni or/and you can download the source.

This is a little old - I wrote it up a while ago and forgot to post it:

I've been running into an annoying problem with Fedora 7 that has stopped me from running httpd as a service. I can spawn it manually (as root) but when I start the service (service httpd start or /etc/init.d/httpd) it fails without an entry in either the messages log (/var/log/messages - SELinux logs its messages here) or the httpd log.
 
In an effort to track down the source of this I put SELinux in permissive mode with:
and then retried the service. Lo and behold - it worked!
 
Unfortunately after 30 minutes of mucking around with permissions I still couldn't convince SELinux that httpd was a trusted service that should be allowed to execute. Hack time.
 
If you are experiencing the same problem, open your /etc/init.d/httpd in your favourite editor and jumpt to the start() function. The offending function call is the call to daemon (part of init.d's functions script) and this is the one that needs to be run with SELinux getting in the way. Add "setenforce 0" and "setenforce 1" around the LANG and RETVAL lines so that you end up with something like:
 

Notes:

1. For non-Fedora systems that may not have "setenforce" you will need to use: "echo 0 >/selinux/enforce" and "echo 1 >/selinux/enforce".
2. You may still have errors with httpd not being able to access files when run as a service. These messages should be available in the log (/var/log/messages) and will tell you which commands (usually a chcon -t public_content_t) need to be issued to fix things

... your name on something (even if they get it wrong ;) ):
 
nsIMsgCustomColumnHandler.idl

Ever since installing the new Linux drivers a number of programs refused to load, the foremost of these being Amarok, complaining about permission problems when accessing libGl.so.1:

Turns out its a problem with SELinux access defaults on Fedora 7 which can be rectified with:

Hope that helps!

Browsing around the source code of different websites may seem to be a boring past time but every now and then you pick up a few nuggets. For example, the index page of UWA's website includes the following comment:
 

Quivis antiburschius Atque irrisores!

 
Which, thanks to this Wikipedia entry, means:
 

Let whoever is against our school

Who laughs at it, perish!

 
It's nice to see some good ol' University pride!
 
Screenshot.

... which is why being in the middle of semester (where did the beginning go? I have a mid-sem next week!) is always the hardest.
 
Last holidays Dvorah & I took a trip down South to the Margaret River area and stayed at Bunker Bay. In an effort to make the photo album a little more interactive this time I've pasted the photos on a Google map which can be navigated interactively as you follow our trip from Day 1 to Day 3.
 
Check it out!

West Perth IBM went into overdrive last week to prepare for the upcoming z/OS camp this Friday. The z/OS camp aims to show uni students that mainframes are still alive and kicking - how important they are, where they are used in every day life and how easy it is to make the transition from uni to programming for the mainframes.
 
Each student received a compilmentary 1gb usb drive (not bad for coming in and listening to a day of lectures) and my job was to preload them with content and information about IBM.
 
Since there is sometimes no real point in reinventing the wheel (though sometimes there is!) I took the IBM Australia website and modified it to include the content that we wanted (that's the Hay St. entrance to the office in the background):
 

Unfortunately, once the content was ready I still had to copy it to 40 USB drives. Under Linux the easy (albeit somewhat inefficent) way would be a quick shell script that ran multiple cp commands in the background (eg. cp src dest &). Unfortunately there didn't seem to be an easy way to do this under Windows, which lead me to a quick and dirty bit of C++.
 
Not very user-friendly in it's current incarnation, the file accepts a source (file or directory) and then multiple destinations to which it copies all of the files. The trick is to spawn one thread per destination and then let Windows handle the copying by calling SHFileOperation. Still inefficiant due to multiple reads of the same source per destination (unless Windows has a very good read cacher), but relatively fast.
 
When used in combination with deveject [source included] and a simple batch file this can be used to wipe, relabel, copy and eject multiple usb drives concurrently.
 
Complete Package [ZIP]

Noticing that the new incarnation of Fedora (previously "Fedora Core") - version 7, had been released I decided to give it a try. I hoped that this would be relatively painless as all my personal information resides off Linux's root partition making it easy to just update the system.
 
Think again.
 
When booting the install DVD and selecting any of the boot menu options the boot loader stalls after loading vmlinuz and initrd, with just "Ready.". Not very helpful. A couple of different posts suggested it was a Dual Core problem and that adding "maxcpus=1" to the bootline should help, but to no avail.
 
Turns out the problem is actually with the way the bootloader interacts with the SATA controllers, causing a lockup when SATA controllers are enabled via the BIOS and booting off DVD. This gave rise to two possible solutions:
 

1. Disable all SATA Controllers in BIOS: Since Anaconda (the Fedora installer) will still see be able to see and manipulate the harddisks even if they are disabled via BIOS you can disable them and re-enable them afterwards.


2. Boot off a USB drive: If (1) doesn't work, download diskboot.img from your local Fedora Mirror (you'll find it under releases/7/Fedora/i386/os/images/) and shoot it onto a usb drive by following the instructions here. Note: You'll still need the install DVD - the USB will look for it after boot.
   
   IMPORTANT! When you boot from the USB the installer thinks you always want to boot from the USB - so it will install the bootloader on to the USB stick (though the boot files will still be in /boot). Make sure you change this to the correct harddisk.
   

Happy hunting with Anaconda!

Here's something you don't see everyday... When you search for "footy tips" on google (try it!) a sponsored link for Microsoft's localised Auatralian Live Search is displayed!
 

 
When you follow the link the following website appears, with the following blatant "anti-google search" message:

With a search engine this good, why search for football scores anywhere else?
Millions of Australians have discovered Live.com combines the best of Web Search with Images, News and Maps. Install some of our handy features below.

Once the message dissapears you get the regular live search.
 

 
I'm not sure how many people are going to want to visit Live search and re-search for something they have just searched Google for, but obviously someone at Microsoft (Australia?) seems to think that people will!
 
A couple of things to note...

• The sponsored link comes up via the google.com and the google.com.au search domain and live.com.au is definitely a Microsoft site (here's the output from a dig) though the final redirected link is to gadgets.ninemsn.com.au
• The link you arrive at is: http://gadgets.ninemsn.com.au/default.aspx?q=football%20scores, although when you visit that without being referred from another site (like Google or this one) you get redirected to http://www.live.com/?mkt=en-au
• You can see more full-size screen shots here and here.

Stay posted for some updates as I try and discover more search words that trigger the sponsored link...
 
Search Terms: footy scores, football scores
 

In an effort to help out my local community I put together a hodge-podge of php code to create an online tipping solution! I'll try and get it up and running to some sort of standard by the time this season is over so that it can be ripe for licensing for next year.
 
Check it out here.

... and he writes a tracker to assess it's worthwhileness. Why I waste my time on these things I do not know!

How's this for a patent idea? Traffic lights that have countdown timers on them to tell you how long you have left for each type of light (green / yellow / red)!
 
Imagine how much time could be saved if we knew how long was left for that really long red light? Or how many lives could be saved by people slowing down when they come near to a green light that has 1 second left (err... on second thought people may actually speed up) - so maybe only a red light counter so that people can make the most of their red lights!

Life seems to be shooting ahead at breakneck speed, what with working IBM which started today, our UWA ImagineCup tournament team (we're calling ourselves 277 Hz - let's see if anyone can work out why!) coming together, moving house coming up in just under two weeks and then uni starting in just under four.
 
As you can guess things are pretty hectic, but that's part of the fun. I started work today at IBM on a part time basis, working at Perth's Australian Development Labs (ADL) office. Not sure yet where this will take me - but I'm sure that it will be somewhere interesting!

... is about all I can say!
 
http://google-code-updates.blogspot.com/2006/12/gsoc-romance.html
 
Thank you Leslie, Chris and the rest of the team @ Google for a great Summer of Code (and let us not forget Mozilla too)!

After an amazing wedding followed by an amazing honeymoon we are now settling back into real life (if you call preparing to move into a new home "real life")!
 
I've been away and offline for about a month and it's time to start catching up on things. Unfortunately I won't be able to give all of my projects the attention they deserve - I apologise for letting people down. I promise that once I get properly settled in I'll get back on track!
 
For now, enjoy some pictures from the wedding. The album software ("photominer") is of my creation, I hope to get it up to scratch a bit more and then release the source code for it too.
 
Till next time!

Google's SoC 2006 has drawn to an end and I have submitted my final patches. Now it's time to get working on the documentation which will be up at MDC. Since things aren't totally finished yet you can track the progress of the columns update on bugzilla.
 
More information to come...

Last night (Saturday, 12/8), the Perth community came together to say tehillim and to farewell Avinoam. The number of people from all ages who attended and accompanied Avinoam in a procession from the shul, is testament to the impact he has had on all of us.

The Rabbi of the Perth Hebrew Congregation emphasised that although a wave had taken Avinoam away, his Torah impact on the community will never be washed away and has left a lasting impression.
 
I am grateful that I had the opportunity last shabbos to learn with Avinoam and experience his warm and sincere personality and sharp Torah mind.
 
The thoughts of the Perth Jewish Community are with Avinoam's parents, family and friends.
 

 

Mozilla has always been known for throwing great parties at any excuse - especially the release of their (now) flagship software - Firefox. Firefox 2 is just around the corner and the call has gone out to start organising launch parties. In response to a call on sfx I created a simple party organiser, which has since (with the help of Mozilla volunteers) gotten it's own domain name and hosting, not to mention approval from Mozilla. So head over to http://www.firefoxparty.com and either register or sign up for a party.
 
Party on!

Although the last few weeks have been holidays from uni they have actually proved to be more hectic than usual. For those who have not yet realised, I proposed to my girlfriend, Dvorah, on the 1st of July and I am happy to say that we are engaged! A link to photos and slight info can be found here (external site). Dvorah and I have been together for about a year and half and she is literally the most amazing woman out there!
 
During the holidays I have been working on my SoC project and have completed the first part - a Mail Folder repair extension. This extension repairs mail items that have become concatenated to each other producing interesting (i.e. weird) new mail counts and mail text. The next part of the project is a real "life-simplifier" for extension writers, letting them properly create, populate and sort their own custom columns in Thunderbird.
 
During the last couple of days of holidays Dvorah and I flew to Sydney for a short break. I've never been to this enormous city before and I had a great time experienceing all the touristy destinations (like the harbour bridge and the Opera House). More info and pictures to be posted soon...

Now that exams are over it's about time that I got properly into my SoC project. To document my travels with Thunderbird I have started mozilla.nachmore.com to chronicle my journey.
 
Note that the site is far from complete and is more of a patchwork of peices that will slowly fall fully into place.
 
And since there can be no work without play, I have finally managed to compile gnome-games on my system (didn't like some of the libraries I had, so I ended up compiling all of GNOME with garnome.
 
After running into some slight openLDAP problems (that I did not have time to investigate) with "make paranoid-install" a "make install" compiled fully. So now all I have to do is sit back and wait for things to start crashing!
 
So what was the point of gnome-games? gnobots2!!! A game from my past:

So... today exams started - kicking off with Microeconomics. This was one of those exams I was just relieved to have over - I had been studying so intensly for it that I managed to dream up a new term in my sleep and then become fixated on finding it in my book.
 
So, if anyone knows what the Nelson system is in economics (I couldn't find a mention of it on wikipedia, though I didn't try to hard!) - could they please enlighten me?
 
Good Luck to all those students world wide who happen to be doing exams!

I have just received an email accepting me into Google's Summer of Code program!
 
I will be working with Mozilla to develop a set of Javascript API functions for Thunderbird, something that will hopefully simplify the creation of extensions for Thunderbird. I have always wanted to be working with Mozilla on their software - especially after working on Firemonger (which is Mozilla related). Now, not only do I get the chance to work with Mozilla - but I get payed for it too! :)
 
Congratulations to all the other students who were accepted!

Today is ANZAC (Australia and New Zealand Army Corp.) Day - the day in which we commemerate the fallen soldiers of past wars, especially those from WW1, WW2, Korea, S. Vietnam, Iraq and Afganistan.
 
As a reservist myself (although not in ANZAC) I understand the need to remember and never forget. The real soldier is the soldier is the one who yearns the most for peace - war is an ugly reminder of the uglier side of human nature.
 
Let us always remember and never forget those brave men who fought and fell, fighting to protect our way of life, so that today we can live a life of freedom.
 
Lest We Forget.
 
For more information check out:
 
Returned Soldiers League, Western Australia
Wikiepedia Article on ANZAC Day
Australian Government - Department of Veterans' Affairs - ANZAC Day

In a startlingly productive few hours in between studying for mid-terms Glob now has support for comment posting. The support is currently rudimentry (so no trying to hack please) and as you'll see not very pretty (at least, not the view comments section).
 
Over the next week or so I will work on making this more robust and prettier (heaps of new CSS coming your way!)...

In a great show of computer skill a hacker hacked into the system and defaced this page, replacing it with a flash script and an SQL error (which is hard coded into the page).
 
I actually thought it was down right funny until I realised that there is a good chance this was a religion based attack.
 
I hope not.
 
To see the hacked site (cause I still appreciate the art behind these things) go here.

Two weeks have gone by in Australia - and I still can't get over what an amazing place this is. Perth is a beautiful city - full of gardens and parks every few blocks - not to mention Kings Park which is approx. the size of the city that I left in Israel.
 
The skyline is amazing, the river that Perth sits on (Swan river) is blue and beautiful (and don't get me started on the beaches!). Perth is quite hot but not humid at all, so the heat is quite bearable.
 
My only complaint: who was the smart one who decided to drive on the left side of the road?

In a couple of hours my plane takes off heading for Australia! I'm starting out with mixed feelings (especially those of taking my computer offline ;) ) but all in all I must admit to being amazingly excited.
 
Not much time - and I still have to pack up my computer! To work...

As the countdown to my flight ticks on (theres about 30 hours left) I realise that I am finding it hard to disconnect my personal computer ("jerusalem") from the internet and pack it up in a box. For those who haven't noticed the names of my computers come from biblical cities - it was better than calling them "1" and "2". At some stage I am sure that this will become a little confusing, but till then...
 
Checking through Google Analytics again and I came across soemthing interesting. Seems that in one day I got 21 unique visits from "Abu Kabir". Unfortunately I can't tell exactly where, but "Abu Kabir" is famous for being the Israeli Forensic center of Pathology and (nearby) a large holding & interrogation center (by the same name).
 
So, either I have a whole lot of dead guys reading my site or I'm really popular with the prisoners... Great! ;)
 

(click to enlarge)

That's it! nachmore.com's move to an external hosting service is now complete, bringing my old server ("jericho") one step closer to the packing box. It's sad to see the old server go offline after serving up nachmore.com to the world for the last few years.
 
I must admit that the transition to my new host (as-hosting) has been relatively smooth, especially thanks to their online support which helped set up the more intensive services (like the L10N Translator service which is currently being user to translate Firemonger). The transition has also given the site a much needed speed boost - sharing a ~12Kb/s uplink with 3 other computers wasn't doing much for the site's speed.
 
Because of constraints in previous projects all of my current projects (L10N Translator, Glob) all utilise information saving to files. Now that I am on a new host i'll start the slog to get change everything to database (hopefully with backwards compat. for file saving) so that I can speed things up (and organise things better).

According to Google Analytics approximately 45% of my visitors use Internet Explorer. Unfortunately this site didn't look the best to those users, so I decided it was time to do something about it...
 
The site has a new background color - to match IE's default colour for the background of transparent PNGs. This helps the background blend in for IE users, though it isn't perfect - the menu and content bubbles still come up with that horrid background. Maybe one day i'll have time to recreate the PNGs as GIFs (or we can just wait for IE).
 
Also, there is now an "infobar" that appears for IE users prompting them to try out Firefox as an alternative. To use this I had to add
to the start of my page which had the adverse affect of ruining the centering effect of the table and some of the font inheritance (as weird as that seems). The font has been fixed - though the centering is playing up a bit.
 
I have also added icons for "nachmore.com" and "glob", so everything should have an icon now...
 
On a general note, nachmore.com will be moving off my private server and onto a [real] internet host soon, so that it can stay up during my coming move to Australia. Hopefully there will be enough time to make the move and set up all of the services currently running on it (especially the L10N translator which is being used to translate Firemonger) so that there will be no noticeable lags or changes (apart from a great speed boost!).

At last, after days (and sleepless nights ;) ) Firemonger 1.5 is ready to go live. 1.5 sees the first official release of Firemonger using MozCD as a backend, running directly off Firefox. This release comes with huge improvements and is a total code rewrite of the previous versions of Firemonger sporting a new User Interface (thanks Ethan!) and heaps of new features.Some of the new features include:

• "Test Drive" Firefox - the option to try Firefox without installing it
• One click Theme & Extension installation - including multiple installs
• Completely rewritten content

Firemoger 1.5 comes with Firefox 1.5 and Thunderbird 1.0.7 (Tb 1.5 is slated for January 2006, an updated Firemonger will be released when it comes out). It also includes popular Themes & Extensions for both Firefox and Thunderbird, Plugins (Java, Flash, Shockwave) and Utilities (Firetune, BackupFox).
 
Btw, when Firemogner was /. a while ago someone raised legal issues with spreading certain Plugins. All plugins have their original licenses with them on the CD...
 
So rock on over to http://www.firemonger.org/en/downloads.html and download your copy today (and then don't forget to pass it on)!

Moving away from the recent computer related stuff, I took my TOEFL exam today so that I can prove to the different universities that I know my English.
 
The exam itself isn't to hard for anyone who has a working knoweldge of English and if English is your first-language you're not going to have any problems at all.
 
So why am I so happy that it's over? Because it is. And that's another stepping stone passed on my way to even getting my application fully in to be considered...

So the night started out with me simply wanting to rebuild a new mod_php (5) into my current apache version (which was running a 4.x version). Turns out that might not be so easy. Though Linux has some sort of standardised file location system each distribution seem to do something else, and it can be mighty hard to install apache over a preinstalled copy (one that I installed when I installed Mandrake way back when). Seeing as my version of Apache was a little out of date as is I decided to rebuild apache as I went along. That was the easy part though. Once installed I had to update the httpd config file in init.d with a quick hack to get it to start and stop my server properly (without looking for the other, older version) and I also had to miagrate over the config files (cause the files that Mandrake provide don't seem to like to work with the standard server).
 
Now it was time to build PHP... but I need flex. So I build flex. But no, that needs yacc. So I build yacc. Finally. Go back and build flex. Go back and build PHP. You'd think it might work now. Nup. Missing libxml2 install - go get the source, download and build. But noooo.... what a suprise. My g++ and cpp aren't good enough for libxml2 and since we are already building everything from scratch - lets go download gcc and rebuild it. At least I can still do things on my computer, seeing as we are compiling on the server.
 
After about 2 hours (thats the downside of running the server on an older computer...) of compiling the build fails with

and more stuff like that. ARGHHH!
 
I decide to download the binaries. 135MB??? From a slow server?? 7 hours??? great... might as well try to see what the problem is. Google'ing this error turns up Mainline bootstrap failure in toplev.c which happens to be the problem. Personally, I just vi the toplev.c file and change the name of the offending function. Don't try this at home people! I am probably stuffing up something serious here, but lets hope not :) (I have written it as a to-do to later get and build a newer kernel with proper headers and then rebuild gcc). Meanwhile, the build gurgles on (literally) happily...
 
An hours goes past... If you haven't realised this until now, this blog is being written as all this stuff is compiling... and compiling... and compiling... and... you get the point! :)
 
Well... it's now 4:40 in the morning and gcc is still building itself busily on the system - looks like mod_php will have to wait till next weekend, unless gcc is finished by morning.
 
Have a great week!

I just put the finishing touches to the CSS and Javascript files to get Internet Explorer to display the site without and problems. Now the only problem is that IE doesn't support PNG transparency leaving the site with an "interesting" background in IE (though I am going to set the background of the site to that colour for IE it still isn't the best solution). I think that we are stuck waiting for IE7 which will provide better support.
 
*sigh* standards?

Now that Firemonger 1.5 is nearly ready (it should be released 3-5 days after Firefox 1.5 is released... whenever that is!) we are moving our focus to getting translations for as many of localisations as possible.
 
As it turns out l10n is a real scary thing sometimes - currently we build off DTD files, but every change to the DTDs be it textual or entity (the name of the specific variable) based creates mass mayhem.
 
Currenly we get the translators to go over the files with a diff program to compare between the English files that they have to the newer English files, but thats a lot of work and it's easy to miss things totally. Also, transferring the changes over to the final translation is a nightmare of multiple windows.
 
To solve all of this I am working on a PHP script. Currently the script does most of the things I just wrote about, the big thing left to do is to detect the change in text (whcih requires caching the english translation for each new translation and then each time checking it against the current english translation). An added bonus is that using a PHP script to automate everything I can get it to automatically commit changes to CVS, which should cut down on the multiple e-mails and zip files that float around...

Just thought I'd tell the world that Glob now supports categories - though I haven't quite worked out what icons I'm going to use for each category (so right now only the "alt" text of the category image is going to come up). Currently I'm using a Slashdot style with the icon floating to the right...
 
Have a great week!

I've just revamped the site, so some things (links) will not work yet. The sections have changed, i've removed some content that I didn't like and other things have been removed to sub links rather main sections.
 
Apart from that the layout still needs finalising - there are some small pieces left to put together.
 
Internet Explorer users - I'm sorry to inform you that currently IE is handling the site in a very weird way. IE doesn't yet support PNG transparency and a few other things. I am working on getting it integrated. At least the site works now! (Side Note: I would recommend checking out Firefox)
 
So stick with me for a few days while I get things together and hopefully soon the site will be up and running smoothly without anything missing...

Seems that glob is spurting out HTML and not XHTML, which is what the new layout is built on. So for all of those compatability junkies out there, ill have a pure XHTML version out soon...