On Wed, 23 Jan 2019 14:32:03 +0100
Christian Gudrian <[email protected]> wrote:

> Usually the list of outputArtifacts is known beforehand or can be 
> derived from the inputs. What about scenarios where this is not the 
> case? Is it "allowed" to perform the actual transformation while 
> evaluating the outputArtifacts property, while the prepare step just 
> returns dummy command?

If there is absolutely no other way, then yes. But keep in mind that the 
outputArtifacts script is not parallelized, so everything you do there will 
become a bottleneck. Also, creating actual non-temporary files there might be  
unexpected (think --dry-run mode).
Depending on the concrete use case, it might be the lesser evil to do some work 
twice.


Christian
_______________________________________________
Qbs mailing list
[email protected]
https://lists.qt-project.org/listinfo/qbs

Reply via email to