Some Eclipse Foundation services are deprecated, or will be soon. Please ensure you've read this important communication.

Bug 144763

Summary:

Provide the concept of an object mine inside each test suite

Product:

z_Archived

Reporter:

amehrega

Component:

TPTP

Assignee:

amehrega

Status:

CLOSED FIXED

QA Contact:

Severity:

enhancement

Priority:

CC:

gvaradar, i.am.brs, paulslau, sluiman

Version:

unspecified

Keywords:

plan

Target Milestone:

---

Hardware:

OS:

Windows XP

URL:

http://www.eclipse.org/tptp/groups/Architecture/documents/features/hf_144763.html

Whiteboard:

Deadline:

2006-09-22

Attachments:

Description	Flags
Object Mine Document	none

Description amehrega

2006-05-31 15:58:54 EDT

A major drawback of TPTP’s automated GUI test recorder is its sensitivity to UI changes.  The purpose of this feature is to propose a change that keeps macro modification to a minimal when widget IDs need to be changed.  The recorder currently embeds the widget id of a resolved object inside of a test case’s macro.  The disadvantage to this approach is that if the widget id of an object changes, then the user is required to walkthrough each test case inside of every test suite and change all occurrences of the widget id.

The proposed solution is to make each test suite own an object mine that keeps track of each object detected.  The object mine will keep track of a single id of an object irregardless of the number of times it occurs in the test cases owned by the test suite.  The user will also be able to include external object mines (i.e. object mines owned by other test suites) and contribute any newly detected objects to an external object mine.

See the attached document for more details.

Comment 1 amehrega

2006-05-31 16:00:12 EDT

Created attachment 43161 [details]
Object Mine Document

Comment 2 amehrega

2006-05-31 16:01:39 EDT

I used whatever time I could find during i3 test pass to implement this feature.  The code is implemented and tested and is awaiting approval to be checked in.

Comment 3 Christophe Telep

2006-06-28 02:51:47 EDT

In plan for 4.3

Comment 4 amehrega

2006-07-10 10:48:34 EDT

Fix checked into CVS.  Changes were restricted to a technology preview item.

Comment 5 amehrega

2006-07-14 14:39:59 EDT

Re-opening this feature as per my discussion with Valentina.
The code is checked in but it still needs to go through a formal process.

Comment 6 Valentina Popescu

2006-07-17 13:55:06 EDT

Reviewed and approved by the AG on July 14

Comment 7 Harm Sluiman

2006-07-21 11:35:04 EDT

A common pattern for doing this matching is to have a library of "hints" that list the public properties of an object that should be used to identify it. Some products puttgin weighting on these properties and make the identifaction process more flexible, and often slower.

In addition any identification using strings will be impacted by translation.

It seems that using data pools here would help. The data pool that is used for identification can contain different types of objects, that also support redirection to translated resource bundles. These object could be live with a compare method, and have structured content.

Also having the code before the approved request doesn't lave much room for public discussiob. I suggest a veting of the ideas be done before an impl is chosen and commited.

Comment 8 amehrega

2006-08-23 13:35:06 EDT

This enhancement has been completed and checked into HEAD

Comment 9 amehrega

2006-08-23 15:28:36 EDT

Updating time

Comment 10 Paul Slauenwhite

2006-11-16 09:01:20 EST

ACTION: Please verify/close this defect.

Comment 11 amehrega

2006-11-21 10:17:43 EST

Closing the enhancement