Three more cases where it's obvious that the code is looking for the
address of the function entry point.

This is to work around the fact that
`SymbolFileNativePDB::FindFunctions` only support
`lldb::eFunctionNameTypeFull` and `lldb::eFunctionNameTypeMethod` now.
Since `main`'s full name is the same as base name (`main`), it's okay to
search with `lldb::eFunctionNameTypeFull` when trying to get the default
file and line. With this, `lldb/test/Shell/Driver/TestSingleQuote.test`
passes on Windows with NativePDB plugin.

Print a warning when the debugger detects a mismatch between the MD5
checksum in the DWARF 5 line table and the file on disk. The warning is
printed only once per file.

Now that more parts of LLDB know about SupportFiles, avoid going through
FileSpec (and losing the Checksum in the process). Instead, use the
SupportFile directly.

This patch updates the source cache dump command to print both the
actual (on-disk) checksum and the expected (line table) checksum. To
achieve that we now read and store the on-disk checksum in the cached
object. The same information will be used in a future path to print a
warning when the checksums differ.

To support detecting MD5 checksum mismatches, deal with SupportFiles
rather than a plain FileSpecs in the SourceManager.

To support detecting MD5 checksum mismatches, store a SupportFile rather
than a plain FileSpec in SourceManager::File.

This is another step towards supporting DWARF5 checksums and inline
source code in LLDB. This is a reland of #85468 but without the
functional change of storing the support file from the line table (yet).

Reverts llvm/llvm-project#85468 because @slackito reports this broke
stepping in one of their tests [1] and this patch was meant to be NFC.

[1]
https://github.com/llvm/llvm-project/commit/d5a277d309e92b1d3e493da6036cffdf815105b1#commitcomment-139991120