Finally! A truly brilliant internet solution that can be installed by anyone, and eliminates the need for expensive infrastructure construction.
This is clearly an ideal replacement for New Zealand’s bad broadband offerings! ;-)
Finally! A truly brilliant internet solution that can be installed by anyone, and eliminates the need for expensive infrastructure construction.
This is clearly an ideal replacement for New Zealand’s bad broadband offerings! ;-)
Ever seen anything like this in your logs?
/home/username/.forward.hostname: Group writable directory: 675 Time(s) /home/username/.forward: Group writable directory: 675 Time(s)
The problem is caused if you have a mailserver setup, it looks in your home directory for a forward file. However, if your home directory permissions has strange privileges (such as being group writable) it complains about it, which gives you these harmless (but annoying) messages.
I came across an interesting issue today – I opened a file in Vim, and yanked (copied) around 140 lines. Then I opened up another file and pasted those lines – however I only ended up with 49 lines!
This caused me a lot of annoyance, until I finally found the fix – in the Vim configuration file, there was the following lines:
set viminfo='20,\"50 " read/write a .viminfo file, don't store more than " 50 lines of registers
The 50 tells Vim not to store any more than 50 lines of data in the copy buffer. Boost that up to whatever you like, and that solves the limitation. :-)
For reference the system-wide configuration file should be in one of the following locations (depending on your distribution):
/etc/vimrc /etc/vim/vimrc
Or you can create a local configuration file at:
~/.vim
Recently a separate git server was setup for kernel.org. (see the news section on kernel.org) Here’s some photos of how it arrived after passing though FedEx. :-O
Amazingly the system booted first time, and is now in production use without any problems. Quite incredible considering the shock it must have suffered…
I’ve written an article giving advice on choosing a suitable Linux distribution for your needs. The article can be found on ReallyLinux.com.
So I had just sat down at my computer, with pizza, ready to relax, and clicked on Evolution expecting to read my email…. but alas, it fails to start. :-(
After running it from the CLI, I got the error:
(evolution-2.6:9203): camel-WARNING **: Failed to initialize NSS
And then evolution just quits. :-(
The Workaround
Undoubtably Ubuntu will release a fix for this ASAP. But for those of you who can’t go 5mins without email, here’s a workaround (in newbie-compatible instructions):
Open a terminal (Applications->Accessories->Terminal).
Type the command: (press enter at the end of it)
LD_LIBRARY_PATH=/usr/lib/firefox evolution
This will start evolution. Keep the terminal window open in the background – if you close it, it will close evolution. Just use this workaround until an update is released to fix this. :-)
WTF is happening?
For the technical readers, here’s what’s happening:
This problem is caused by the upgrade of firefox just recently (upgrade to 1.5.dfsg+1.5.0.10-0ubuntu0.6.06.1), and the problem is that the location of the NSS libraries have moved from /usr/lib to /usr/lib/firefox. Evolution requires the NSS libraries to provide support for SSL and other encryption technologies.
Here’s the bug entry from Ubuntu for this problem.
The command “LD_LIBRARY_PATH=/usr/lib/firefox evolution” works by tells evolution to look in the /usr/lib/firefox directory for the dynamic NSS libraries.
Today’s technical lesson
In case you are interested, LD_LIBRARY_PATH is an environmental variable that tells programs where to look for dynamic libraries. It can be configured in /etc/ld.so.conf. By default, it looks in /lib and /usr/lib. Many distros also specify /usr/local/lib. (Interestingly, I just checked, and it doesn’t appear that Ubuntu has this file created by default).
After making a change in it, you can run ‘ldconfig’ to update the library cache based on the new settings.
Steven J. Vaughan-Nichols (DesktopLinux.com) has written a review of him comparing Vista to Linux, on a step-by-step basis from install up.
You can read his whole review here. However, what you should definitely read, is page 3.
Steven came across big problems with Vista relating to hardware support, such as buggy drivers with missing features, and certain common hardware not being supported. It also sounds like Microsoft has been understating the true system requirements to run vista, and there are some interesting stats in the article.
I fail to see how Microsoft can claim that Vista is ready for people to use, when there is a lack of driver support and missing driver features – it all shows signs of being rushed out of the door – and I wonder what else got rushed – security testing perhaps?
What scares me even more are certain big organisations who are starting to role out Vista installations on a big scale, when the software is so new. It is also unfortunate that Dell has stopped offering WinXP on their desktops, consider that lots of people will still want XP for another year or two at least.
However, I feel the best bit in the article was “When I switched back to Vista, I tried to play Wilco’s Yankee Hotel Foxtrot CD. Whoops! Not a single sound emerged from my speakers. After a little investigation, I found that Vista disables media outputs that don’t incorporate DRM, when you try to play DRM protected media through them.”.
Vista certainly does deliver on the “WOW” experience. The “WOW, this really sucks” experience.
In the weekend I went out and picked up a little Intel Core Duo 1.66Ghz Apple Mac Mini. I got the model with 512MB RAM and 60GB hard disk, and I am using it as my desktop/home server, freeing up my amd64 for developmental work.
Here’s what I love about it:
However, in my adventures with it in the weekend, I found a few things that are a problem:
Other cool stuff:
If you want a box to run Xen on, the Intel Mac Mini’s support the VT-instructions, that allow you to run unmodified operating systems (eg: MS Windows or legacy OSes) on Xen. If you are interested in that, check out the info here.
Overall, I think it’s a great little unit and would recomend it to anyone after a quiet desktop/small server/media center. :-)
In the last few weeks I have been working on the base packages for Jedo Linux. Today I have completed a bootable system and in another week, I will have finished tweaking it, and fixing small issues.
I would say, I am only about 2 weeks behind the (rough!) schedule given in my last status report, which isn’t bad going really. :-)
Once I have the base packages working reliably and problem-free, I will be turning my attentions to the package manager.
Whilst Portage is quite a nice package manager, the emerge program doesn’t suit my needs in the following ways:
I will be writing my own package manager, based partially on Portage. Essentially, it will use the ebuild.sh script from Portage (with a whole bunch of changes) which handles the ebuild files.
I will then write a program called JPM (Jedo Package Manager) to replace emerge. This program will:
So that’s what I’ll be working on for the next month. And hopefully in March, I will be ready to start working on the installer, and in April might be able to release a test ISO of the base system. :-)
I’ve found an interesting problem when I was in the process of building base packages for Jedo Linux. I could compile coreutils okay, but I then found that the command “rm” when used with the -r option, would fail. As an example:
mkdir -p /tmp/junk/foo/bar/mydir rm -rf /tmp/junk rm: cannot remove `/tmp/junk': Function not implemented
After spending a long time looking into this problem, making sure my toolchain was okay, etc, I finally got a hint of the problem from this post, relating to a build problem with BeOS that causes the same/simular problem.
The author goes on to mention that it can be reproduced under Linux, by making /proc inaccessible at build time.
Since I’ve been building in a chroot, /proc wasn’t mounted. After mounting it and building coreutils again – it worked!
So: To anyone out there building coreutils – MAKE SURE /proc IS MOUNTED!! :-)
UPDATE
Interesting…. it seems that the problem occurs anytime if /proc is not mounted! I’m not sure if this is a requirement of coreutils, or some kind of weird bug….