From: Fruechtenicht, T. <TFruechtenicht@PSI.DE> - 2006-06-29 17:57:16
|
>Von: Timothy Wall [mailto:tw...@us...] >Gesendet: Di 27.06.2006 19:29 >An: Fruechtenicht, Torben >Cc: Abbot Users >Betreff: Re: AW: [abbot-users] Using Abbot to perform tests on a = full-blown application, wrong component refs and lookups > >On Jun 27, 2006, at 11:06 AM, Fruechtenicht, Torben wrote: > >>> The fuzzy matching needs some rework in when it is applied, >>> potentially leaving it out entirely unless specifically requested by >>> the user. However, that alone does not necessarily address the >>> problems raised by specific application configurations. The best >>> solution there is to augment abbot's test suite with some >>> representative examples of the hierarchy in question. This has been >>> used in the past to overcome bugs encountered by specific >>> applications which I could never reproduce. >> > >> What do you mean by "representative examples of the hierarchy"? >> > >If my application consists of a single frame with three buttons, then =20 >a representative example would be a single frame with a few buttons. > >If my application consists of many interchangeable frames with the =20 >same deep hierarchy, a representative example would be two or more =20 >such frames with a deep hierarchy. > >Basically I mean a hierarchy similar enough to the original to =20 >produce the same results. > I am still not sure what you mean exactly. How would I augment a test = suite with representative examples of my component hierarchies? Is Abbot = capable of self-learning and how? Torben > >Timothy Wall >http://abbot.sf.net |