Category: Nifty solutions
Yeah, I finally can post to fedora-devel-list! It seems that the moderator of email@example.com accepted me, but I didn't recieve an e-mail notifying me about that. I also know where the problem is now: that weird smtp-server at our university translates all adresses @VTK.UGent.be into @vtk2.ugent.be and the admin of that server doesn't know why. So, I was subscribed with my address @VTK.UGent.be, but the list-system recieved mail from @vtk2.ugent.be who's not subscribed.
You can find my post to the list here. There also was a reply from someone, asking for a diff -u output so I hope they get back on the project again.
For the past weeks, when you entered "fedora stateless" in Google, this blog was the first answer. Now, it seems this has ended and we're back the second answer, after the official fedora stateless website. When you enter "stateless linux", we appear on the second page, but before fedora stateless. Yeah, strange thing, google, but really cool B-)
As you know, I can't post on the fedora-devel-list. So now, I've sent an e-mail to Havoc Pennington, the RedHat developer who firstly announced the Stateless Linux project at the fedora-devel-list. He told me he forwarded my e-mail about the pyGTK dependency to relevant people and he wants some discussion to get more people involved in it. Let's hope they can bring some swing in the project again.
He also told me about the list firstname.lastname@example.org. It seems that you have to be subscribed at that list before you can post on any redhat-list. Strange system and not explained in any bouncing mail or description of other mailinglists, but it'll be so. I've subscribed to that list, confirmed and I'm now waiting for moderator approval again. Let's hope they let me in soon.
I really hope they bring some swing in the project again. At the message that Hans De Goede placed for me at fedora-devel-list, there's still no answer, so no discussion and no progress.
Yeah, the galabal yesterday was fantastic! It was a great time. You can see some pictures here. I was in my bed too late, so I'm not in the mood to study. Well, the ideal time to take a look at stateless-snapshooter.py.
First of all, I downloaded the latest CVS-files as shown here. I think the file sml/snapshooter/stateless-snapshooter.py is the only one I need to patch. The other classes used by this file seem OK for me.
In this file, I removed the line
import gtk (du-uh), the whole definition of the class
progressBar and the defs
snapshot_finished wich also use the progressBar-class. At the bottom of the file, where the execution of the snapshooter is called, I placed following lines in comments:
progress = progressBar ()
shooter.get_transfer ().connect ('update', snapshot_progress, progress)
shooter.connect ('snapshot-finished', snapshot_finished)
I hope it runs fine now. It was my first contact with Python. The only thing I knew about it, is that it's indent-sensitive, so I'm not sure about all this stuff here. If the fedora-devs want, they could always add some text-progressbar to it.
Because of the subnet troubles, I can't test it right now. I'll have to wait till monday, if I've time then. If it still doesn't work, I think about giving it up and just install X/pyGTK/fluxbox to make it work. Yeah, I'm a looser, I know.
About those subnet troubles, I've spoken the person responsible for that. He told me the whole thing is because the network at our university is good protected, but the documentation about it for the people working on it, is quite poor. He doesn't exactly know where to find the problem within a reasonable time. So, he decided to give us shell access to a machine, ftwe.ugent.be, in the same subnet, that is normally reachable from within our university network. So we'll log on onto our student server eduserv, ssh to ftwe and ssh from there to ftwlinux, our machine. So, Steven or Stijn, if you're reading this (you will, I know ;-)), please fix this so I can let the whole world know how fast you are B-)
I wanted to add the file with the exact diffs here, but the blog-engine (of configuration) didn't allow that.