Hi,
Disclaimer: I've been contributing to SWTBot for several years,
despite I try to be objective, it's very likely that I do not
fully succeed at it; so my answers here shouldn't be the only one
to consider in your decision!
On 11/20/2016 02:10 AM, Evgeny Mandrikov wrote:
My last experience with SWTBot was five years ago,
maybe something changed, but at that time it wasn't easy/stable.
Not much changed for SWTBot. There is now a Test Recorder that
allows to get most of the test code recorded and generated according
to user actions, and there are a few fixes and improvement added
here and there; but it's still a Java API to refer to widgets and
interacts with them in a "human-like" way. See
https://www.eclipsecon.org/europe2016/sites/default/files/slides/Introduction%20to%20Functional%20Testing%20with%20SWTBot%20and%20Maven-Tycho.pdf
for more details. SWTBot per-se has more or less always been stable,
but the fact is that it's not always easy to write tests working in
a stable manner (mostly because stability requires properly dealing
with waits and timeouts).
The nice thing is that it integrates very seamlessly in PDE and
Tycho, as it's just an API.
Another pro for SWTBot is that it has more users than RCPTT at
Eclipse.org I believe (such as EGi and TraceCompass).
Recently heard about RCPTT and quick try was nice.
I can't tell for RCPTT. I've never used it and none of the project
I've worked on has been using it.
I've had good feedback about it by Obeo's people using them on some
Modeling project, but that's about all. Also, last time I heard
about it, it seemed to be on another "grain" than SWTBot, as it
introduced a DSL, a specific runner and was not integrated with
Tycho. But I imagine some things have changed since them.
HTH.
|