John Wiegley wrote:
On Jun 27, 2009, at 1:32 AM, Simon Michael wrote:
I have also been looking at your tests, and plan to use your .test
file format. I think you call these baseline tests, while I want to
call them something like func. tests or shell tests or command-line
tests or.. ? They seem like something useful outside of *ledger
development.
Actually, the "baseline" tests are just a category within the Ledger
tests. I use the same format for the "regress" tests, which I try to
create for each reproducible bug.
The Python scripts used to run the tests are:
test/LedgerHarness.py
test/RegressTests.py
If you find a nice way to abstract it, let me know!
I've made a similar test runner:
http://joyful.com/darcsweb/darcsweb.cgi?r=hledger;a=headblob;f=/tools/shelltest.hs
and a couple of tests, eg:
http://joyful.com/darcsweb/darcsweb.cgi?r=hledger;a=headblob;f=/tests/unbalanced.test
It prepends "hledger" and appends "-f -" when necessary, which is not very general. I'm thinking those should be
overridden if you provide a --exe "progname" option, at least.
I made it accept >>> instead of >>>1 and I think I'll make all parts optional -
anything not specified will not be tested.
