2013/1/23 Bernardo Dal Seno <[email protected]>: > On 21 January 2013 16:23, Michael Hanselmann <[email protected]> wrote: >> This patch factorizes “_RunXmList” and adds some tests. > > Nice work, thank you. Only, I'm not sure about the fact that the tests > contain parts of the data files, as they check the parsed results and > compare them with hard-coded constants. Wouldn't it be better if the > content of the data files was put in a list, as you did for > testWrongLineFormat()? in this way it would be easier to keep the code > in sync with the data. I don't mind the duplication as much as the > distance between the two things to be kept sync.
testWrongLineFormat uses made-up data, specifically for detecting parsing errors. We already use the concept of having test data in dedicated files in other places, for example for tests regarding parsing /proc/drbd. If Xen were to change the output format, we can simply add a new data file. This is much easier than making sure data in Python code is the same byte-by-byte. Having to keep data and code in sync is unfortunate, but necessary. Michael
