Meeting notes from this today's Hyades
UI meeting:
Topic:
Discuss a change proposed by Jonathan
to the test navigator code, relating to whether the navigator should show
the name of the test suite, or the name of the file containing the test
suite.
Attendees:
Bianca Xue Jiang
Curtis d'Entremont
David Hodges
Eugene Chan
Jerome Gout
Jonathan Hoover
Time:
Wednesday, Sept. 15, 12:00pm (noon) EST.
Notes:
The rationale behind the test navigator
was not to show the filesystem; we already have a navigator for this. Instead
we would like the test navigator to show a logical representation of the
test artifacts in a meaningful way that is easy for the user to understand.
Currently, the UI is inconsistent in this respect; some things show the
filename while others show the logical name. We want to be consistent throughout
the navigator, so it was proposed to adopt the logical view throughout
the navigator.
There were a few issues raised about
this:
1. Logical names are not necessarily
unique; so how to guarantee that the underlying filename would be unique.
A proposed solution was to annotate the logical name with some unique ID
that would be used to construct the filename.
2. There may be more than one test suite
stored in a file, but there are problems representing this in the UI.
A few other comments that were made:
- Logical names have more flexibility
than filenames, as there are no character restrictions, etc.
- Currently the execution results show
filenames with timestamps which are not readable by the user (they are
32-bit integers). This should no longer be shown in the UI.
- It is difficult to move things around
in the current navigator. For example, there is no move, copy, paste, and
other standard actions like in the eclipse navigators. It is desirable
for the test navigator to provide some basic functionality like this, or
a framework to do it. We don't want different people to implement different
versions of these standard actions where the dialogs might look different,
and inconsistent.
Jerome to open a bugzilla to track the
problem.
Thanks,
Curtis d'Entremont
Problem Determination Tools
IBM Toronto Lab