Update actions to use libpcre2

[ryfactor]

[ryfactor]

I guess the scope of this has increased to updating the code that uses pcre's functions to however pcre2 does it. I'll take a look this weekend hopefully with a solution, unless someone else beats me to it.

[ryfactor]

Now the vm cannot find libpng. I guess I better update that next.

[ryfactor]

GNU 13.3 now throwing this error:

In file included from /home/runner/work/ivan/ivan/FeLib/Include/config.h:18,
                 from /home/runner/work/ivan/ivan/FeLib/Source/config.cpp:13:
In copy constructor ‘festring::festring(cfestring&)’,
    inlined from ‘festringpile::festringpile(cfestring&)’ at /home/runner/work/ivan/ivan/FeLib/Include/festring.h:145:37,
    inlined from ‘festringpile operator+(cfestring&, type) [with type = const char*]’ at /home/runner/work/ivan/ivan/FeLib/Include/festring.h:159:10,
    inlined from ‘static truth configsystem::NormalStringChangeInterface(stringoption*)’ at /home/runner/work/ivan/ivan/FeLib/Source/config.cpp:[23](https://github.com/Attnam/ivan/actions/runs/13045884947/job/36396279429?pr=661#step:6:24)7:31:
/home/runner/work/ivan/ivan/FeLib/Include/festring.h:172:7: error: writing 8 bytes into a region of size 0 [-Werror=stringop-overflow=]
  172 |       ++REFS(Data);

[ryfactor]

GNU 13.3 now throwing this error: ...

I was unable to reproduce this error in a devcontainer running ubuntu-24.04 with GNU 13.3.0.

GitHub must have shipped a bugged compiler with their ubuntu-latest runner, or I don't know what they've done. Furthermore, devcontainers and runners are not the same beast, making it very inconvenient to reproduce the same errors. So much for solving the "works on my (virtual) machine" issue.