As mentioned in the issue for those places that only need the error code it would've been easier to get it from the headers but I'm unsure whether there are any case where this information is only provided through the body.
So to keep the current behavior I decided to keep parsing it from the body (now based on content-type though)
If you think it would be alright to rely solely on the headers than I'd be happy to change that, it would remove a bit of code and prevent having to parse the body in some cases.
EDIT: Thanks to the all_tests script I discovered the issue of quick-xml being behind the xml feature flag.
I see two options but don't know which one is better:
dont feature gatequick-xml / remove the xml feature
create a non-xml version of the body parsing and gate the current one behind the xml feature (needs documentation)
To quick fix for CI I will remove quick-xml from the xml feature for now. I'm thankful for any further guidance here
closes #1275
As mentioned in the issue for those places that only need the error code it would've been easier to get it from the headers but I'm unsure whether there are any case where this information is only provided through the body. So to keep the current behavior I decided to keep parsing it from the body (now based on
content-typethough)If you think it would be alright to rely solely on the headers than I'd be happy to change that, it would remove a bit of code and prevent having to parse the body in some cases.
EDIT: Thanks to the all_tests script I discovered the issue of
quick-xmlbeing behind thexmlfeature flag. I see two options but don't know which one is better:quick-xml/ remove thexmlfeaturexmlfeature (needs documentation)To quick fix for CI I will remove
quick-xmlfrom thexmlfeature for now. I'm thankful for any further guidance here