This is largely deprecated
This is still pretty raw and under development - expect to have a few teething problems. Basically, we want to create a mysql database to store the results in, and parse a directory full of results into it.
Then we can use cgi scripts to dynamically query the data. Make sure you set up apache to have appropriate aliases, and to allow cgi scripts to be run.
Everything you need is under the tko/ directory (make sure you've checked out the whole source tree, not just the client), both the parser to fill the database, and the frontend.
See the basic instructions under tko/README. You will probably have issues - feel free to email the list (autotest@test.kernel.org), and we'll help you out.
30 second orientation below:
Backend
parse - the main script. "parse <top_level results dir>"
- parse.py - helper library for parsing.
Database
tko - name of the default database in mysql. Stands for "test.kernel.org", not "total knockout"
- .database - two line text file specifiying host and database. Defaults to 'localhost' and 'tko'.
- .priv_login - username and password for the backend (needs read/write!). No defaults. Make sure this IS NOT world readable.
- .unpriv_login - username and password for the frontend (readonly). Defaults to 'nobody' and no password. Make sure this IS world readable.
Frontend
- *.cgi - some fancy cgi scripts. machine_kernel.cgi is the main functional one and machine_benchmark.cgi the main perf one.
- frontend.py - Frontend classes for the database to make it easier to use. All the cgi scripts use this.
- display.py - draw coloured matrix tables, etc (see machine_kernel.cgi for example usage)
- plotgraph.py - python class to draw gnuplot scripts for you (see machine_test_attribute_graph.cgi for example usage)