Saturday, March 13, 2010
Arm Builds
Our attempt is to see which pre-existing SRPMS from FedoraCore 12 will be able to build on ARM, and if there are simple fixes in order to make them build, then to utilize them. At first these builds seemed to take a very long time, then we found out that was due to someone utilizing all the computers resources at the same time. After this changed build times went down. However, the virtualized machines do not seem consistent.
Later on I'm going to post all the build success and build failures and hopefully soon the real ARM machines will be in and we can compare how the builds in a virtualized ARM machine compare to the builds in a real arm machine.
Friday, February 19, 2010
Mock and Koji
Wednesday, February 10, 2010
rpmbuild: First Attempt
http://www.irssi.org/files/irssi-0.8.14.tar.gz. I then extracted
irssi, and looked for files like, "LICENSE, INSTALL, README,
ChangeLog, Makefile" as these are all very important when building the
SPEC file. Before I started building the SPEC file I set up the RPM tree using, "rpmdev-setuptree". This will make
a nice safe area to build RPMs. The rpmbuild directory
was created under ~/rpmbuild. The tarbal needed to be moved to
~/rpmbuild/SOURCE directory. Then the "rpmdev-newspec [packagename.spec]" command is
issued, which will create a bar bones SPEC file that you can fill out
based on the information gathered earlier in the tarball. I found
that reading http://fedoraproject.org/wiki/PackageMaintainers/CreatingPackageHowTo
was fairly helpful to understanding what section means what. Once the
SPEC file is filled out the "rpmbuild -ba packagename.spec" is issued.
Initially I found some errors, and resolved them. after this we issued the rpmlint [packagename]. I had no errors, however I had a warning which was because of trailing spaces.
Tuesday, January 26, 2010
Make -J
Make real 0m55.661s user 0m26.299s sys 0m28.122s | Make real 0m56.408s user 0m26.324s sys 0m28.234s |
Make -j2 real 0m31.994s user 0m25.047s sys 0m24.933s | Make -j2 real 0m31.591s user 0m24.953s sys 0m25.034s |
make -j3 real 0m25.007s user 0m23.618s sys 0m24.088s | Make -j3 real 0m25.661s user 0m23.808s sys 0m23.964s |
make -j4 real 0m39.048s user 0m58.727s sys 0m28.605s | Make -j4 real 0m22.115s user 0m23.160s sys 0m23.179s |
Make -j6 real 0m20.827s user 0m23.418s sys 0m23.376s | Make -j6 real 0m20.331s user 0m23.284s sys 0m23.703s |
Make -j8 real 0m20.179s user 0m23.637s sys 0m23.953s | Make -j8 real 0m20.708s user 0m23.706s sys 0m23.896s |
Make -j
-j [jobs], --jobs[=jobs]
Specifies the number of jobs (commands) to run simultaneously. If
there is more than one -j option, the last one is effective. If
the -j option is given without an argument, make will not limit
the number of jobs that can run simultaneously.
Both machines I used were QUAD cores, and the results were fairly obvious. The -j option increases the amount of jobs and it seems the compilation running just one job at a time is much slower than running multiple jobs. It's seems optimal to run between 6 and 8 jobs on the Q6600 quad core.
Monday, January 18, 2010
SBR600 First Compiles
Under Ubuntu 9.04 in order to finish the compile I had to issue the following commands:
sudo apt-get install build-essentials
sudo apt-get install libncurses5-dev
After doing so the compile ran normally.
#Change the directory to whever NLED was extracted
sudo make
Then enter nled and it's up and working.
This was done on Fedora
The compile I chose to do was asterisk, since I'm interested in asterisk anyways
#Change the Directory to where ever the source code was extracted
./configure
make
make install
Make sure that your running as superuser for these commands.
Friday, January 15, 2010
SBR600 First Blog
http://zenit.senecac.on.ca/wiki/index.php/User:Dgilloch