escape_cdata() leads to double-escaping about exml

[ElectronicRU]

Oh, proptests'd do it, yea. PropEr is one of the thing I miss most about Erlang.

[michalwski]

Hi @ElectronicRU, thanks for reporting the issue. We'll have it fixed as soon as possible, what version did you use previously?
Do you use exml with esl/MongooseIM?

[ElectronicRU]

16 авг. 2016 г., в 10:09, Michał Piotrowski [email protected] написал(а):

Hi @ElectronicRU https://github.com/ElectronicRU, thanks for reporting the issue. We'll have it fixed as soon as possible, what version did you use previously?
Do you use exml with esl/MongooseIM?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub #20 (comment), or mute the thread https://github.com/notifications/unsubscribe-auth/AAW_XR_2F-xm9-YBFHSjYM3l4KXPCgZqks5qgWIOgaJpZM4Jh6Mr.

Not really, no. I just was upgrading from a predating repository for that sweet rebar3 support. ^_^

[ElectronicRU]

Any updates on the matter?

[NelsonVides]

Most of this logic has been so heavily reimplemented since then that I doubt this issue still applies. Closing this very old issue, please let us know if anything here still applies.

[ElectronicRU]

Well, I'm glad @NelsonVides! Perhaps it's a good idea to add some tests against that erroneous behaviour nonetheless.
Unfortunately been a long time since I used the library, so I'm maybe not the best person to do it, sadly.

[NelsonVides]

The cool thing is that escape_cdata is not even exported any longer in the exml interface, and the tests for cdata escaping now have even property-based tests. So this shouldn't be an issue, unless you manually call exml_nif:escape_cdata/1 and then maybe introduce some double-escaping, but, that is not part of the documented API anyway 😄
Thanks for your input, and your prompt answer even after so many years! 🙃