Test routine

[mhantke]

Hi developers!

I guess it would make sense to establish testing routines for keeping stability in the code.

Probably we would have to provide a set of XTC files that are processed. If a new option is introduced you write a function that checks whether you get the expected result from processing a particular file.

What are your thoughts about that?

Cheers,
max.

[antonbarty]

Hi Max

Having a standard set of tests to run in order to check for the introduction of unintended features (er, bugs) would be very useful.
Tom calls this a 'regression test' (see http://en.wikipedia.org/wiki/Regression_testing)

We can set up something similar here on cfelsgi with a few XTC files because we control access rights.
It would be useful to have testing data at SLAC also - I would have to talk to the folks there about where this could be placed.

The question is which data sets?
We'd need something that is not sensitive or political for any experiment, but tests the software.
Something that tests all detectors (pnCCD and cspad), and both single particles and crystals.

How about:
• Lysozyme nanocrystals from CXI (deposited on CXIDB so already open-access)
• CAMP data for single particles (one of the runs from Ilme's data - also already on CXIDB)
• A random dark run of both cspad and pnccd

Cheers

A.

On 19/04/2013, at 11:42 AM, Max Hantke [email protected] wrote:

Hi developers!

I guess it would make sense to establish testing routines for keeping stability in the code.

Probably we would have to provide a set of XTC files that are processed. If a new option is introduced you write a function that checks whether you get the expected result from processing a particular file.

What are your thoughts about that?

Cheers,
max.

—
Reply to this email directly or view it on GitHub.