[release/9.0] TypeName fuzzer and bug fixes

[adamsitnik]

This PR combines the PRs that added Fuzzing for both TypeName and AssemblyNameInfo with the most recent bug fixes.

Once it gets merged there won't be known bugs related to TypeName and AssemblyNameInfo parsing in release/9.0

Fuzzing PR:

Backport of #107195 to release/9.0 (it has also contained empty name fix)
Backport of #107257 to release/9.0
Backport of #107206 to release/9.0

Bug fixes:

Backport of #106334 to release/9.0
Backport of #107261 to release/9.0

/cc @jkotas @buyaa-n

Customer Impact

• Customer reported
• Found internally

Two bugs were found by the Fuzzer (sbyte overflow and incorrect handling of empty assembly names), the last one was reported internally by @jkotas.

Regression

• Yes
• No

[If yes, specify when the regression was introduced. Provide the PR or commit if known.]

Testing

All bugs discovered by the Fuzzer has been turned into unit tests (and of course are passing now).

Risk

Low. The bug fixes were relatively simple:

• don't start AssemblyNameInfo parsing when provided name (span) is empty
• don't cast int to sbyte just to pass it to a ctor that accepts an int
• increase max node counter when parsing nested types

[adamsitnik]

@MihuBot fuzz TypeName

[adamsitnik]

@MihuBot fuzz AssemblyNameInfo

[MihaZupan]

I added support to the bot for running against non-main base branches now.
It's been running the fuzzers for a couple of hours now: MihuBot/runtime-utils#647, MihuBot/runtime-utils#648
FYI the continuous fuzzing only runs against main.

[jeffhandley]

@artl93 - Please review this for RC2 inclusion