(cleanup) pam: Move model code into specific package

[3v1n0]

So we can have it into a non-main package, leaving into main only the code that has to be there for library or binary generations.

This is just a first iteration, let me know if we want to do further cleanups too.

[codecov-commenter]

Codecov Report

Attention: 4 lines in your changes are missing coverage. Please review.

Comparison is base (edc6202) 83.31% compared to head (733a6a3) 83.42%.
Report is 2 commits behind head on main.

Files	Patch %	Lines
pam/pam.go	76.92%	1 Missing and 2 partials ⚠️
pam/adapter/authmodeselection.go	0.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #173      +/-   ##
==========================================
+ Coverage   83.31%   83.42%   +0.10%     
==========================================
  Files          57       57              
  Lines        4562     4568       +6     
==========================================
+ Hits         3801     3811      +10     
+ Misses        590      587       -3     
+ Partials      171      170       -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[denisonbarbosa]

I like the idea, but the name adapter does not feel right/complete. Can we think about a better one?

I'll start thinking about one and will report back as soon as I have some ideas...

[3v1n0]

@denisonbarbosa sure... That was the result of a Friday rambling shouting between me and @didrocks, so just an idea... Used also in #155.

Check the MM logs, but happy to hear if you have better suggestions!

[didrocks]

Some suggestions for you!

[didrocks]

Once you fix the generator for the tests and quality check to pass, you can merge :)

[didrocks]

Not fan about passing them as an unique struct to various functions TBH, especially in the new call where I would build the object from discrete args.

[3v1n0]

Not fan about passing them as an unique struct to various functions TBH, especially in the new call where I would build the object from discrete args.

Well, thing is that in some places I'd need to pass isInteractive, isGdm or whatever... and feel that having just boolean parameters around is not something nice to maintain.

[didrocks]

Yes, but a struct, with is something like an optional struct is worse. At worst, you can say this is something small than a client and embed this struct in the client, but here, initiliazing with a Parameter is really not a pattern which is idiomatic

[3v1n0]

Yes, but a struct, with is something like an optional struct is worse. At worst, you can say this is something small than a client and embed this struct in the client, but here, initiliazing with a Parameter is really not a pattern which is idiomatic

I see, I'll revert this and handle it in the verbose way ;-).

[3v1n0]

Ok, moved also pam_test inside internal package too, and reverted the Parameters change.

[didrocks]

The tests are still failing. Rerequest once the tests pass.

[3v1n0]

The tests are still failing. Rerequest once the tests pass.

Yeah, as always the last "cleanup" change broke things, I hope this last rebuild will be all green :)

[denisonbarbosa]

Yeah... the annoying part is that this package does so much... We might need to take another look into it later, but it's not worth blocking this because of it.

Nice work!

[didrocks]

Looking good now!