| Phlip 2006-11-13, 10:05 pm |
| varma wrote:
> Can anybody plz provide some information on this Question, Thank you
>
> Q. While conducting your testing on the data entry page you start to
> get an error message that stops you from being able to progress
> testing.
>
> What kinds of tests would you perform to try and isolate where the
> error is occurring from.
Is this for a quiz?
The goal is to give developers the maximum information, so they can proceed.
Before finding the bug, I would turn on the program's logging feature, and
turn the logging level up so the log rapidly fills with trivial details. The
log should stash every SQL statement the program sends to its data store,
for example.
When I find the bug, I would copy out the log and read it, maybe print it
out and highlight all the SQL statements that look like they contain the
data I entered.
Then I would attempt to reproduce the bug, by entering the same data the
same way, and by taking notes of each step. If the bug is not reproducible,
I might contact a developer to get help reproducing it, or to ask what areas
they feel are high risk.
If the bug is not reproducible, and if I followed good manual testing
procedures, I am exempt from characterizing the bug. And teams should treat
non-reproducibility as very serious, because it typically means the program
corrupts its environment!
If the bug is reproducible, I would then open an SQL database console, and
attempt to reproduce the bug at the SQL level. If I can, the bug's in the
database, and if I can't, the bug's in the client application.
In general, one can "divide and conquer" to look for a bug. The SQL
statements form an interface between two layers, so if I can sample that
interface, then I can attempt to reproduce the bug in each layer separately.
So to continue dividing and conquering, I have to identify and isolate more
interfaces between layers.
I'm curious what the author of this test said to do about it.
--
Phlip
[url]http://www.greencheese.us/Z Land[/url] <-- NOT a blog!!!
|