java - Should I chain GUI tests or start the applet anew every time? -
in open source project jchempaint, example, gui tested (using fest framework) collecting dozen individual tests each few java files. applet started once per file, , several independent tests performed in chain.
i know if practice. of course, starting each time cost time. can see problems side effects of previous actions , possible exceptions, i'm no expert. so, practice put several tests 1 applet start?
(i'm looking collection of best practices gui testing can't pose such question, hints welcome nevertheless.)
i'm troubled awkward division between these 2 top-level containers:
org.openscience.jchempaint.application.jchempaint
org.openscience.jchempaint.applet.jchempaintabstractapplet
.
after cursory reading, reluctant critical; re-factoring contents might limit amount of duplicate testing required. in simplified example, common initialization confined initcontainer()
method. comparison, jchempaint
considerably more complex , offers number of applet parameters, correct transfer of should tested.
such re-factoring may on-going. appletests
appear date earlier period of development, while newer jchempaint.src.test
artifacts appear reflect more recent, annotation-based testing architecture.
Comments
Post a Comment