[TIP] Fwd: Robot Framework 2.0 is available!
stefano.masini at pragma2000.com
Tue Jun 24 23:08:22 PDT 2008
Here's another reason I couldn't find it. It wasn't out yet! :)
By the way, I gave it a try and it looks interesting. The way I see it
is as an imperative language written in tabular form.
Every row is a statement made by a predicate (called "keyword") and a
variable number of parameters.
You can define statements, either within the framework as tables, or
outside, in python code. The latter is the equivalent of a fit
fixture, i.e. the glue between the framework and your System Under
Tests can be arranged into a hierarchical structure the same way as
packages and modules do in python. In fact, the same __init__
convention is used to identify packages, where __init__ can be a file
containing initialization code.
Test files contain the tests, of course, but also constants (since
it's tabular, it makes sense to group large packs of data in a
separate section) and keyword definitions (I see it as function defs).
Test files can be written in html or Tab Separated Value txt. But
providing support for other formats is trivial. I for instance quickly
hacked together support for XLS files using the xlrd library. I can
post the patch is someone is insterested.
As far as reporting is concerned, the runner spits out a nice html
with green/red background and tests grouped according to "packages"
and "modules". Tests can also be assigned tags, and the runner can be
configured to selectively run only tests with specific tags. The
report contains also percentages of success/failure grouped by tags.
Finally, an xml version of the whole execution is available, in case
one needed integration with other tools.
There's only one thing that makes me skeptical, but I guess that might
depend on my inexperience with FIT or other acceptance testing
solutions that work in tabular format: if all was made to resemble so
much an imperative language, why not use the *real* imperative
language? I mean, we're not talking about a syntactically complex
language that's so overloaded to make it impossible to practically
write acceptance tests beforehand (test first approach). We're talking
about Python, that is, in my opinion, clean enough to be used almost
as pseudo code in a planning session, to quickly hack together what's
gonna become the actual test.
But maybe, as I said, I'm just not experienced enough. Maybe I'll find
that hacking together acceptance test tables is even easier. Maybe I
will even be able to do this with the customer! (yeah... sure...) :)
Alright, that was my opinion about Robot Framework.
I'll wait to see other comments if they come.
---------- Forwarded message ----------
From: Pekka Laukkanen <peke at iki.fi>
Date: Tue, Jun 24, 2008 at 8:56 PM
Subject: Robot Framework 2.0 is available!
To: robotframework-users at googlegroups.com,
robotframework-announce at googlegroups.com
Cc: robotframework-devel at googlegroups.com
Robot Framework 2.0 is released. Installation packages can be be
downloaded from the project home page  and also short release notes
 are available. This is the first official open source release of
the framework and thus a huge milestone for the project.
You received this message because you are subscribed to the Google
Groups "robotframework-users" group.
To post to this group, send email to robotframework-users at googlegroups.com
To unsubscribe from this group, send email to
robotframework-users-unsubscribe at googlegroups.com
For more options, visit this group at
More information about the testing-in-python