Re: Marking non-automated tests

On Wed, Feb 6, 2013, at 11:36 PM, Kris Krueger wrote:
> We've discussed adding meta-data to the list a few times....each time we
> have decided this was not a good option to pursue.
> That said it's still important to have this data so that it could be
> consumed.
> 
> I'd suggest that we have a text file (format/file type is not a concern
> of mine) that holds this 'meta data' and other metadata (for example test
> is approved).

Well, everything in the master branch should be approved at this stage.
Plus special branches for special cases (like the CR branch).

External metadata is no-good metadata. Having it in the files is much
easier and will be less of a problem. I think the proposal makes sense,
but one thing is that I don't write <html> in my files, but they more
often than not look like this:

  <!doctype html>
  <meta charset=utf-8>
  <title>Awsum test</title>

... Hmm. This is about manual tests you say? Ohwell, yeah, totally fine
with requiring the <html> boilerplate on those.  If not, it could always
look for data-manual on the first few lines.

Anyway, marking up the manual ones, as they are the ones we don't want
people making makes lots of sense.
-- 
  Odin Hørthe Omdal
  odinho@opera.com

Received on Thursday, 7 February 2013 08:59:35 UTC