[Dev] Perf test cases

[John Anderson]

Heikki Toivonen wrote:

>Since 0.6 is going to be about the calendar I think it makes sense to
>defer non-calendar perf tests until later.
>
>So, I propose that we drop use cases 1 (switching to All view) and 2
>(stamping). By the way, these tests are closest to the acceptable level
>we have defined (in fact, with empty repository we already are
>acceptable for 1, and are acceptable for 2 on Linux).
>
>Also, so as to not kid ourselves about the number of tests, I propose
>that we clarify the remaining tests as follows:
>
>Tests with empty repository:
>
>1. App startup
>  
>
The repository should be built as part of our bulid process so users 
don't have to do it. Building a fresh repository should be something no 
user should ever have to do unless Chandler is not reliable.

>2. New event creation (menu)
>3. New event creation (in-place)
>4. Creating new calendar
>
>Tests with 3000 events:
>
>5. Importing a 3000 event calendar
>6. App startup with 3000 event repository
>7. New event creation (menu) with 3000 event repository
>8. New event creation (in-place) with 3000 event repository
>9. Creating new calendar with 3000 event repository
>
>(Note that the tests are in the order which would make most sense to do
>a full test run.)
>
>  
>
>------------------------------------------------------------------------
>
>_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
>
>Open Source Applications Foundation "Dev" mailing list
>http://lists.osafoundation.org/mailman/listinfo/dev
>  
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osafoundation.org/pipermail/dev/attachments/20050920/f4c38d4d/attachment.htm