feat(ffi): enable content scanner by passing an existing instance

[jmartinesp]

Replace setting up the content scanner at the ClientBuilder to being able to set it on the fly when needed.

It's a lot easier for the clients to enable/disable this on the fly than having to create the client with the value ahead of time.

• I've documented the public API changes in the appropriate changelog files (see Writing changelog entries).
• This PR was made with the help of AI.

Signed-off-by:

[codecov]

Codecov Report

❌ Patch coverage is 90.00000% with 3 lines in your changes missing coverage. Please review.
✅ Project coverage is 89.91%. Comparing base (4bd5e81) to head (e323540).
✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
crates/matrix-sdk-contentscanner/src/lib.rs	71.42%	2 Missing ⚠️
crates/matrix-sdk/src/media.rs	92.85%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #6689      +/-   ##
==========================================
+ Coverage   89.89%   89.91%   +0.01%     
==========================================
  Files         396      396              
  Lines      110412   110434      +22     
  Branches   110412   110434      +22     
==========================================
+ Hits        99258    99295      +37     
+ Misses       7370     7360      -10     
+ Partials     3784     3779       -5     

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

[codspeed-hq]

Merging this PR will not alter performance

✅ 50 untouched benchmarks

---

_{Comparing feat/ffi-content-scanner (e323540) with main (4bd5e81)}

[poljar]

A couple of strange things in this PR but ok. Once fixed feel free to merge.

[poljar]

Can't we put this into the if branch, this avoids the need to expose DefaultMediaFetcher publicly.

Also, what's the purpose of the ContentScanner struct? Are you going to scan stuff manually on the app side?

Adding this struct could have been a nice self-contained commit or even PR for that matter, as it's not related to the hot-swapping of a media fetcher.

[poljar]

That's a strange way to do this. Usually we do this type of test with pointers to check that the objects are the same.

Can we at least use the Debug implementation so we don't add unnecessary and possibly confusing Display implementations?