refactor: modernize cache.cc

[kvakvs]

• Convert tr_cache to a class
• Review size types and use size_t where appropriate
• Replace C container for blocks with C++
  • Fix tests
• Review and fix narrowing integer conversion warnings where possible
• Add CLion temporary build directory to .clang-format-ignore and .gitignore

[kvakvs]

@ckerr What happened to the readonly array view (Span) merged last October in #1927 and removed by you the following week? Thought I could use it again to replace raw pointers pointing to raw data, and it's gone. For now i am going to use a reference to vector instead.

[ckerr]

@ckerr What happened to the readonly array view (Span) merged last October in #1927 and removed by you the following week? Thought I could use it again to replace raw pointers pointing to raw data, and it's gone.

@kvakvs I removed it because I was going to use https://github.com/martinmoene/span-lite or https://github.com/microsoft/GSL ot https://github.com/gsl-lite/gsl-lite instead, but then I kept looping on which to use (mostly deciding between most-lightweight span-lite or best-maintained GSL) and never did anything.

If this is a blocker for this PR I'll take a look again. There are several places in libtransmission that would be improved by using a span.

[kvakvs]

For the moment it is not a blocker at all.
Worst that can happen, the code taking a pointer and a size remains at that, and no new checks or changes will be added.
But any function taking const any* or any* with a length is a direct candidate for a Span. Allows index out of range checking too, and allows to group size and pointer together forming a slice.

[kvakvs]

A few cache tests fail without actually crashing.
I could use some tips and hints which old trPtrArray operation is rewritten wrong, or at least which function of the new Cache is likely to be failing to make those tests fail.

[ckerr]

@kvakvs I'll try to take a look in the next couple of days. Sorry for the delay; a lot of my time this week has been taken up afk.

Also, FWIW needing help with this code is totally understandable. I kind of hate this code and I'm the one who wrote it 🤷‍♂️

[ckerr]

@kvakvs kvakvs#1