Closed
Description
Originally reported by: Jason R. Coombs (BitBucket: jaraco, GitHub: jaraco)
I'm finding that the junit XML generated by pytest 2.2.4 gets double-encoded output. For example, here's a snippet from one run:
<failure>... Error in parsing
gryphon/models/qsl.py:63: QSLError</failure><system-err>Message 1
{&apos;exception&apos;: [&apos;ParseError: Both inline and named res
ponse lists were used for question lookahead\n&apos;],
&apos;message&apos;: [&apos;ParseError: Both inline and named respo
nse lists were used for question lookahead\n&apos;],
&apos;status&apos;: 500,
...
The stderr stream for that looks something like:
Message 1
{'exception': ['ParseError: Both inline and named response lists were used for question lookahead\n'],
'message': ['ParseError: Both inline and named response lists wer eused for the queston lookahead\n'],
'status': 500,
Notice that the tickmarks are being escaped for XML (to ') and then again the ampersand is getting escaped to '.