trcli sometimes defaults to CP1252 instead of unicode about trcli

[Testinator-X]

According to the installation instructions python 3.10 or newer is recommended. Have you already checked if this also happens when you update your python Version?

[d-rede]

Hello @ch4h2s

Thanks for reporting this issue. Unfortunately, I am not able to replicate it. Could you kindly provide a report file that hasn't worked for you?

Thanks

[bitcoder]

Closing this in the next days unless there's more feedback on it

[ch4h2s]

Note: this may have to do with the fact that python was running on windows (Win 10) and not on linux when this happened.
At any rate, trcli should force utf-8 when communicating with the testrail server and allow an option to read the encoding from the xml file and/or via an option on the command line and/or via environment.

[bitcoder]

As @Testinator-X mentioned, trcli requires at least Python 3.10; tests currently run on 3.10, 3.11, and 3.12.
@ch4h2s , can you please test it using the latest trcli version (1.6.2) and Python 3.10 and let us know if the problem arises? If so, can you please share the JUnit XML report?

[ch4h2s]

Sorry, the old framework is not in use anymore, so cannot repro the issue here myself.
Also that same old framework had been patched to use ASCII to go around this bug well before being deprecated (actually before this bug was even reported), so it would not have been able to reproduce the issue anyway.

If you are sure that trcli now uses UTF-8 for the XML input (or reads the encoding from the xml headers), you can close the issue.

If you need to test it on your machines/frameworks, try to get some non-ascii, non-CP1252 characters inside all of the xml strings and see what happens when run on windows.

[bitcoder]

As neither of us is unable to reproduce it, I'm closing this issue.