cleanup: const a few tox pointers for groupchat API

[JFreegman]

This change is 

[codecov]

Codecov Report

Attention: 5 lines in your changes are missing coverage. Please review.

Comparison is base (a1e2458) 9.98% compared to head (576fcfc) 9.99%.
Report is 15 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff            @@
##           master   #2379      +/-   ##
=========================================
+ Coverage    9.98%   9.99%   +0.01%     
=========================================
  Files         141     141              
  Lines       32530   32530              
=========================================
+ Hits         3247    3251       +4     
+ Misses      29283   29279       -4     

Files	Coverage Δ
toxcore/tox.c	7.80% <0.00%> (ø)

... and 5 files with indirect coverage changes

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

[Green-Sky]

Reviewed 2 of 2 files at r1, all commit messages.
Reviewable status: 0 of 1 approvals obtained

[iphydf]

LGTM. There's no API break because there's no release with this API in it, right?

[iphydf]

@robinlinden pointed out, correctly, that we tend to use const in the public API to mean deep-immutable. What do you think? Should all Tox parameters be const everywhere?

[JFreegman]

There's no API break because there's no release with this API in it, right?

Correct (unless you count client releases)

@robinlinden pointed out, correctly, that we tend to use const in the public API to mean deep-immutable. What do you think? Should all Tox parameters be const everywhere?

Is there ever a valid reason for it not to be const?

[iphydf]

If we only make the tox functions const that don't modify things inside the tox state, then it helps client code ensure that if they pass a const Tox* anywhere, then no mutations on internal state will happen.

[iphydf]

@JFreegman I think the documentation aspect of "const means no changes happen to the object" in a broad sense is useful in the public API. Maybe someday I'll write a linter to enforce something like that, but I'm not sure what the syntax and semantics of that should be.

[iphydf]

Closing this for now. Reopen if you want to discuss reasoning for what "const" should express in the public API.