Friday, June 3, 2011

Django Libs

This is an old project of mine, a rig to play Mad Libs from a web interface.
Have fun, and if people want to code on it, we should set up some sort of VCS.

Work Log: Fri, June 3

8:30
Arrived 8:20.
Started up web2py, crashed the welcome app.  Turns out I was missing some python-gluon dependencies.  I'll edit my post on how to backport web2py accordingly.
9:00
Small diversion posting my old DjangoLibs project.  Learned how to use dropbox public files, however.  I suspect that will come in handy.
10:00
Got a new python project, will look at while installing Oneiric: getting the datetime in a human-readable format.
11:30
It's not my backport that's broken, it's the Oneiric package.  The debian is broken too.  Installing debian to report the bug.
2:00
Bug appears to be upstream, in the source package.  Ignorable?
2:30
Unclear on how to use CGI.  Frameworks are nice.  Work on web2py app?

Thursday, June 2, 2011

Quickly Backport Web2Py to Natty

I found that downloading the source code and just running the server is most reliable.  But the package currently works.
This package might be broken.  Treat it as you would any other sid package.
wget http.us.debian.org/debian/pool/main/w/web2py/python-gluon_1.96.4-2_all.deb && sudo dpkg -i python-gluon_1.96.4-2_all.deb && wget http.us.debian.org/debian/pool/main/w/web2py/python-web2py_1.96.4-2_all.deb &&  sudo dpkg -i python-web2py_1.96.4-2_all.deb
It's all one line.  Paste it in your terminal for deb based distros.

Investigation on Jabber Servers

Sugar relies on shared roster groups.  Googling this resulted in a page from xmpp.org: http://xmpp.org/extensions/xep-0140.html

At the top was a red line of text:
WARNING: This document has been retracted by the author(s). Implementation of the protocol described herein is not recommended. Developers desiring similar functionality are advised to implement the protocol that supersedes this one (if any).


Well, that's bad.  Incredibly bad.  That tells you that required functionality will only be supplied by old or bizzarely maintained software.  However, there was a mercy on the page: 
the protocol described herein has been retracted in favor of XEP-0144: Roster Item Exchange
That is a draft standard protocol:
http://xmpp.org/extensions/xep-0144.html
I will study
1. How it differs technically from XEP-0140
2. Whether it could be modified to be backwards-compatible with XEP-1040
3. Whether it would be feasible to modify existing Sugar applications to work with XEP-0144 instead of XEP-1040

Work Log: Thurs, June 2

8:30
Arrived.  Working on Openfire pet project till matt gets here.
9:30
Working on openfire, frustrated by lack of dependencies built into RPM.  Back in debianland, a package will tell you when it won't install correctly.  Had to restore from a snapshot; telling RPM to uninstall the package would result in it complaining that it was not installed, and telling RPM to install the package would result in it complaining that the package was already installed.   All the while, the software was on the system, but broken.
10:30
Discarding old projects.
Two new ones: Web2Py work with Stavros, and getting web2py packages working under Lubuntu Natty.
How does Ubuntu release its software?
I'll be doing a backport, of course Debian has instructions on how to do this...
11:30
Web2py installed.  I believe it works.  It starts.
Backport successful.  One must simply manually webget python-gluon and python-web2py and dpkg them.
12:30
Back from lunch at 12:40.
1:30
Made post with step-by-step and immediate instructions for backporting web2py.  It's ready to rock.
2:30
Installed web2py on physical machine.  Also installed Guake.  Both of these things will make my life easier.
Now an activity is not working...
3:30 signed off.

Wednesday, June 1, 2011

Work Log: Wed, June 1

8:30
Arrived 8:20.  Only one fellow student here.  Shall inquire about Fedora vm.  If it isn't already set up, will investigate stripping down a desktop install.
At 8:30, installing from minimal iso.  I'll print out the time sheet.
9:00
Reinstalling fedora 15 with x; the host tools would not install without X -- and then the VM can be stripped down to run on old hardware by inexperts.
9:30
It's running xfce slowly.  I can't drop it down to a console.  The user is admin, the password is default.
9:35 -- yum install jabberd
*presses easy button*
10:00
Got SSH working with host machine.  Put machine in runlevel 3.
Working via SSH from here out, optimally, on ballistophobia.  This will make copy-pasting possible.

If I'm making a vm image, why not make a livecd that runs vms?  I'll look into this.
10:30
Reinstalling, textmode, with better name so as not to break jabber.  Bridged adapter will fix network problems.
11:00
Still encountering texmode problems.  What happened?
11:30
Installing ejabberd on textmode.  Remember kids, ifconfig -a.\
1:15
Reinstalling again.  Ejabber is currently large, fragile software -- I do not see how it is suitable for widespread deployment.


NOTE: DATA FROM HERE TILL 3:00 WAS LOST BECAUSE I DID NOT SAVE

Tuesday, May 31, 2011

Work Log: Tues, May 31

8:30
Arrived.  First task: move some equipment onto a powered cart.
at 8:43, all five laptops were wired to the hub.
Lessons learned: I put things down and needed to spend time looking for them.  If I pick something up, put it in place or back.

8:45
Intent to test sugar on laptops displayed.  Schoolserver?
Matt will be asked many questions.   He will be in at 10:00.
9:45
Researching SchoolServer (XS).  When Matt Gallagher gets here, work may accelerate.
10:45
Fixed Carlos's site; the CSS validation link needs to be changed from check/referrer to check?profile=css3&uri=referer. 
Matt Gallagher arrived.
11:15
I shall be setting up a simpler, more updated, schoolserver than XS.  From scratch.  While documenting.
Physical machine obtained.
12:15
Setting up server - encountering error.  Memtesting, shall try using only one controller.
1:30
Returned from lunch.  Update from 12:15 -- tried using another machine, but it turns out that the disk was incorrect.  Downloading the iso and will test it against known sum.
Will test disks against their sums in future before moving hardware.
1:45
Fedora-15-i386-DVD.iso: OK     I like that line of output.  I'm going to burn the image to disk and start again.

This disk encountered the same error.  Checking the disk's sum.  If it's different from the image's sum, then there may be something wrong with my ROM.  If it's the same, then there is something wrong with a board, or it expects to boot a CD, etc. Checksum failed. Writing slower.

2:30
 Image written.  Checking sum first.

3:15
Future: Set up virtualbox and document procedures.
3:55
Log close, work cease.