I’m developing a testing suite based on Selenium 2. The goal is to test a series of pages under different situations (specified through an arbitrary XML file).
The output is given through:
- Some zipped screenshots
- A raw log file
- An html file
Right now the raw log file and the markdown log file (later used to generate the html) is generated through messages hardcoded in the app.
Now I’m thinking of using some auxiliar data structures to contain every single piece of relevant info, using some kind of convention (codes, enums, etc.) to later process and present. If I’m not wrong, this should make the code easier to mantain and less redundant.
My question is: is there a better approach, or widely accepted practice, for this?
As an example of how the output (just a scratch) is:
<Test> | <name of page> | <browser>:
* Searchbox found and query submitted
* ERROR: timeout while waiting for results
<Test> | <name of page> | <browser>:
* Question found and answer submitted
* Alert box content: < ... >
* OK
Thanks
PS: if anybody fins a more suitable question title, it’ll be more than welcome
Edit
Sorry if my question isn’t as concise as it should, I’m trying my best to make myself clear.
In other words, when my program has to output a potentially big amount of text giving information about the result of the tests executed, how should I do it? hardcode the messages or use some auxiliar data structure (something like a queue/pile) that keeps track of what had happened since the start to the end of the execution (be it error and status codes, for example) and then process it compose a readable document?
Using a presentation-independant data structure is obviously a better choice. It allows