junitxml doesnt escape everything

[pytestbot]

Originally reported by: Ronny Pfannschmidt (BitBucket: RonnyPfannschmidt, GitHub: RonnyPfannschmidt)

---

as reported by anton on the ml

---

• Bitbucket: https://bitbucket.org/pytest-dev/pytest/issue/126

[pytestbot]

Original comment by Floris Bruynooghe (BitBucket: flub, GitHub: flub):

---

Original report:

Hello all,

I have the following issue with generated junitxml reports:

I'm writing test cases which use scapy functions. So I make asserts on
some packet objects. As I understand scapy uses esc characters for its
own objects representation, and I can see colorful output on the
console. But when test is failed, and its traceback went to xml, I
can't open generated xml due to the error "XML Parsing Error: not
well-formed" In my case the problem with esc charecter 033 (\x1b).
I resolved the issue by adding the following code to append_failure
method in LogXML class:
longrepr = str(report.longrepr).replace('\x1b', '\033')
fail.append(longrepr)

I think that it will be good if you add some cleanup functions to

avoid such situations.

Thank you,

-Anton

[pytestbot]

Original comment by Ronny Pfannschmidt (BitBucket: RonnyPfannschmidt, GitHub: RonnyPfannschmidt):

---

junitxml: use a exclusive match on the legal ranges of xml for binary escaping, fixes issue 126