[qmtest] The state of GUI

[Vladimir Prus]

On Thursday 07 April 2005 19:51, Mark Mitchell wrote:
> Vladimir Prus wrote:
> > Yea, but if people decide that editing 15 files for 15 versions of "yfc"
> > is not too nice (even with GUI), they should be able to do something.
> > Like using custom file format.
>
> When Stefan and I talked about this yesterday, we decided that the most
> consistent thing would be to add command-line functionality for
> adding/removing/updating/constructing results in a results file.  Then,
> you would just use our friend the shell:
>
>    for x in *.qmr; do
>      qmtest <some command-line here> $x
>    done
>
> For example, one primitive might be "merge".  The idea is that you would
> provide N results files, and the output would be a results files
> containing all of the results in the N files, with results in later
> files overridding earlier ones.  So, you would run the test once, to get
> a failure, or create a results file with a failure in some other way
> (say, via the GUI), and then do:
>
>    for x in *.qmr; do
>      qmtest merge $x fail.qmr
>    done
>
> Would that work for you?

I though about it and think that if I have a file describing expectation for 
all possible toolset, and QMTest requires N expectation files, each for a 
specific toolset, then nothing stops me from writing a script that will 
produce those N expectation files. I'm not even sure that command like 
interface for adding expecation is needed; for me, Python API is file.

- Volodya