values need to be bytes when writing back to ldap

[angeloudy]

What does this PR do?

What issues does this PR fix or reference?

fix for #48599

[dwoz]

Should the encoding/decoding be specific to the ldap server? It looks like this PR is basing the encoding/decoding based on the client's system settings. Have you thought about what happens if the client system is using a non utf-8 encoding?

[angeloudy]

@dwoz
That's a good point!
I have no idea how to get the encoding.
What do you suggest?

[dwoz]

@angeloudy I think it's best to use utf-8 here (it looks like that is what LDAPv3 expects by default). We can make it configurable down the road if needed

[cachedout]

@dwoz and @angeloudy There is also __salt_system_encoding__ that should be available.

[cachedout]

Hi @angeloudy. Where are we on this one?

[angeloudy]

@cachedout should salt.utils.data.encode use __salt_system_encoding__ by default?
I did have any 'utf8' hardcoded in my changes.

[terminalmage]

@angeloudy It would try utf8 first, and then try the detected encoding if that raises a UnicodeEncodeError.

[rallytime]

@angeloudy Are you planning on making the requested changes here?

[angeloudy]

@rallytime what requested changes.
I am fine with my current version.

[angeloudy]

@terminalmage @rallytime
if an UnicodeEncodeError exception catching is necessary, shouldn't it be done inside salt.utils.data.encode?

[rallytime]

@terminalmage and @dwoz can you comment/re-review?

[angeloudy]

Hi, any plans to merge this pull request so that I don't have to maintain my own patch?
I don't understand why are you guys worrying about UnicodeEncodeError, the encode method is being used everywhere in saltstack.

[terminalmage]

I don't think anyone is worrying about an encoding error. You still haven't made the suggested changes from a few months ago, which is probably why this hasn't been merged.