Tuesday, March 30, 2010

ARMs Length

So after reading and tweaking sample openssl code and config here:

http://fedoraproject.org/wiki/Koji/ServerHowTo

I managed to break the code down into 2 separate peices. One for making the CA and one for making the certs for the different users. Now the way my script differs from the original is that it uses the variables passed to it as the default commonName in the ssl.cnf file.

For example, when executing my script to create a cert for a new user...lets say "kojiuser1":

#./certscript kojiuser1

The script as usual asks the various questions about where you are from and the OU name and the province etc, etc, but is then followed by the question of what the commonName should be. This name and/or the OU name should always be different from any other certs already created or you will get a TXT_ error. My script makes pressing enter easier by allowing you to press enter knowing that the commonName will be (by default).. "kojiuser1". Now this may not seem like anything huge for the creation of one user but what if we were creating multiple users (which was the case with me).

#for user in arm{1..25}
>do
>./certscript $user
>done

In addition, a folder called "confs" holds all the conf files used to make the certs, this makes backtracking easier. For now only one line gets modified but it leaves progression for future endeavours. The only time you need to press a key other than enter is when it asks if you want to sign the cert. Of course 95% of the script remains the same as before but this modification makes life one step easier and hence minimizes the length of the cert creation process.

Sunday, March 14, 2010

ARM mock build info for beneath-a-steel-sky-cd

Source RPM: beneath-a-steel-sky-cd-0.0372-4.fc12.src.rpm
Processor: ARM processor (under emulation)
Command used: mock -r fedora-12-arm ~/armPacks/beneath-a-steel-sky-cd-0.0372-4.fc12.src.rpm

Comments: (SUCCESS)

firstbuild:

real 13m43.593s
user 9m27.560s
sys 2m41.840s

secondbuild:

real 13m18.596s
user 9m23.130s
sys 2m15.540s

ARM mock Build info for python-basemap

Source RPM: python-basemap-0.99.2-5.fc12.src.rpm
Processor: ARM processor (under emulation)
Command used: mock -r fedora-12-arm ~/armPacks/python-basemap-0.99.2-5.fc12.src.rpm

Comments:

After approximately 8 minutes the build fails and produces 2 errors saying

1. "Bad build req: No Package Found for geos-devel. Exiting."

2. "Exception(/home/user1/armPacks/python-basemap-0.99.2-5.fc12.src.rpm) Config(fedora-12-arm) 6 minutes 29 seconds"

After investigating the build/error logs located at "/var/lib/mock/fedora-12-arm/" it was realized that geos-devel was an intricate component for python-basemap and could not be excluded.

This issue will be revisited later.

Wednesday, February 17, 2010

Don't Mock Me

After going through a variety of errors and variations, I finally got my wireshark source rpm to work using the fedora-12-x86_64 config file on 'ireland'. Initially I was rebuilding the package with rpmbuild -ba whenever I edited to the spec (which took a very long time), however it was later brought ot my attention that I could simply use rpmbuild -bs to just include the edited specfile into the rpm package. After many, edits my mocked source rpm is up and running

located at http://scotland.proximity.on.ca/slatty/mock/

see you guys in class

Friday, February 5, 2010

Wireshark building 101

After days of failed builds my wireshark package is complete. Apparently, it was always supposed to be trial and error and that there isn't a real approach to follow. My source rpm Can be found at

Http://Scotland.proximity.on.ca/slatty/wireshark-1.2.5-1.fc12.src.rpm

It also passed the rpmlint test on both the spec file and the binary file without any warnings or errors.

Mock and Koji are next on my list...look out for blog tomorrow

Saturday, January 30, 2010

To build or not to build, that is the question

I am not sure if I missed this in class or if it applies to this package only (nled), but when I run the rebuild command my spec file disappears, in the case of wireshark that doesn't happen, however, that could be because wireshark didn't get rebuilt successfully because of missing dependencies. To get the spec file back I have to use the rpm -i on the package.

The second part of this exercise creates a minimal (skeleton) spec file however my current mission is to find out if there is a way to provide an alternative specfile name so that I can compare the minimal vs the main spec file

Wednesday, January 27, 2010

Lab1 continued...1st part done

So it seems I also needed ncurses-devel alongside ncurses..after installing both of those then rebuilding,remaking...bled was operational....the new step will be part b using another source code

So far I'm not being very successful with my choices of packages does anyone have any suggestions?