Two things that occur to me as possible causes are:
1. You are using different component names for the different actions, and the check is taking place in another tab
2. Other parameters are different (e.g. you are using "relative" as the path type instead of "absolute")
It might be worth posting the relevant part of the test result report - once for the select, and once for the check.
Hm, that is indeed rather strange. One other possible avenue to try is whether it has to do with synchronization in any way. You could try, e.g.
- Selecting Errors/Chronological and then checking the existence (thus assuring that the tree is open - if it is dynamically created, then there could be some timing problems happening here)
- What happens if you just try check existence of the Errors entry?
Andre Renard Messages: 20 Registered: February 2013
Junior Member
I am also facing the same problem. I use the simple match only.
I was trying to match 'Navigation/Executions' in a tree (TEXTPATH in ub_tre_selectNode_byTextpath Test Case)
Worked with: N*/Executions but not with */Executions or N*/Execution*
I am using Jubula 1.3.0.201210161256 on Windows7 with Eclipse RCP 3.7.2 for the AUT.
Pretty ennoying because unreliable.
Your problem sounds like it is a different one - the problem that Yashveer found is specifically in trees with columns, and only for checking. If you put more details in about your problem, then someone might be able to help.
For Yashveer -
I'm afraid the only workaround I can think of would be to try to select the node - you could build a module that:
- Selects the node (if it is there, the test will be successful)
--- The module should contain an empty Test Case as an Event Handler set to action error that continues if an action error occurs.
That would mean that if the node is not there (usually a check failed), then you would get an action error, but the test would continue (i.e. as if a check failed had occurred).
Another more complex route would be to build a module that:
- Selects via textpath the parent of the node you want to check
- Then selects via indexpath (with SEARCH_TYPE relative) the node whose existence you want to check (this presupposes that at least one node is a child of the previously selected node)
- Then performs check text of selected node
Ah, now I understand. Paths are analysed stepwise - that means that if you enter * for the first level, then the first item that matches this will be taken (in this case, General). Executions is then not found, and you get the error. N* works because the first thing that matches this path part is the "Navigation" item.
It can certainly be discussed whether it is the best way of doing things, but at least at the moment it works as designed. You can open an enhancement to discuss the behaviour if you'd like.