On Thu, Jul 5, 2012 at 4:27 PM, Jed Brown <[email protected]> wrote:

On Thu, Jul 5, 2012 at 2:22 PM, Matthew Knepley <[email protected]> wrote:

...

Are you just running make targets from Python or are you trying to parse

...

all the make tests into a richer (and more compact) Python representation?

The latter.

Right, so you need to interpret a little bit of shell, at least to the extent of expanding some variables and determining which output file to test against. Also, there are some examples that | sort or | grep -v, so you may need to recognize that.

I just do regexp.

I like the semantic information of having multiple tests using the same base options and/or comparing to the same output (rather than normalizing when you specify the tests).

Shell is not the place for that. Years of putting crapy shell into makefiles before having to pull it out should have taught us something. Matt -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener