refactor: Make Tox_Options own the passed proxy host and savedata.

[iphydf]

This way, client code can immediately free their data after the options setter returns true.

---

This change is 

[nurupo]

can immediately free their data and can pass temporaries to the options setters

This contradicts with:

* The setter tries to copy the string, but if it fails, the value is not
* copied and this member is set to the user-provided pointer. In that
* case, the user must not free the string until the Tox_Options object is
* freed. Client code can check whether allocation succeeded by comparing
* the value of this member to the user-provided pointer.
*/

Looks like the client has to check first if the copying succeeded before they free the data, so not exactly immediately, but could be close enough based on what you meant with that comment. As compared to the previous behavior, you can free it sooner, yes.

Also, if the user clones Tox_Options, freeing it will free the pointers twice.

[robinlinden]

Should this be ck_asserted so there's no potential for a use-after-free in the test?

[iphydf]

Yes, good idea. Done.

[codecov]

Codecov Report

Attention: Patch coverage is 73.46939% with 13 lines in your changes missing coverage. Please review.

Project coverage is 72.18%. Comparing base (aa64916) to head (ebe150f).
Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
toxcore/tox_api.c	73.46%	13 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2819      +/-   ##
==========================================
- Coverage   72.21%   72.18%   -0.04%     
==========================================
  Files         151      151              
  Lines       31122    31176      +54     
==========================================
+ Hits        22476    22505      +29     
- Misses       8646     8671      +25     

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