You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We want to use Catch2 benchmark for performance regression tests. Therefore we output the data as xml and store it in a database that we can compare the performance over different commits. For comparison reason, we need to store meta data, such like which compiler and which CPU was used and so one.
If we use std::cout, we get a single xml region, which contains all the output and we need to parse the content by our self to split up the different information.
<StdOut>
First output
Second output
second line of the second output
third output
</StdOut>
INFO, WARN and so one is similar. It has the advantage, that each output become it's own xml region, but there is still no difference between the xml region. Each INFO creates the same <Info> xml region and we need to search for an identifier in the content of the xml region to get the right information such like which compiler version.
In the best case, Catch2 provides a mechanism, which produces the following kind of xml output:
There is currently no feature for this, and I am not aware of commonly used benchmarking framework with labels like these -> more often there is a support for user defined numeric counters. However, if someone adds the numeric counters, adding free form text labels should be just some extra boilerplate.
User defined counters are a thing I would like to add at some point, but my own work on Catch2 is on hold at least until October.
We want to use Catch2 benchmark for performance regression tests. Therefore we output the data as xml and store it in a database that we can compare the performance over different commits. For comparison reason, we need to store meta data, such like which compiler and which CPU was used and so one.
If we use
std::cout
, we get a single xml region, which contains all the output and we need to parse the content by our self to split up the different information.INFO
,WARN
and so one is similar. It has the advantage, that each output become it's own xml region, but there is still no difference between the xml region. EachINFO
creates the same<Info>
xml region and we need to search for an identifier in the content of the xml region to get the right information such like which compiler version.In the best case, Catch2 provides a mechanism, which produces the following kind of xml output:
Is there a feature in Catch2 to achieve it and if not, what is the best practice to get something similar?
The text was updated successfully, but these errors were encountered: