Update comment

• Fix comments
• Fix p2p_inv_download extended test

Back to you for the test failure also

@bot build-native-osx

@bot build-osx

Rebase on latest D11622 (and fix off-by-one conf)

Use HIDDEN category instead of the conflicting hidden_args list

Add comment per feedback

Address feedback:

• Remove now unnecessary init test
• Clarify help text
• Add new arg to hidden args

Fixed an off-by-one issue where configuring -avaproofstakeutxoconfirmations=0 actually meant 1 conf.
This was a hold over from when the parameter was originally named after "utxo depth" but was changed
to confirmations since this is verbage the users are already familiar with.

Fix clang warning

Rebase on D11664

Move UTXO height check after existence check. It doesn't make sense to check
the height unless the UTXO exists. This logic is relevant to making the orphan
pool accept only immature proofs.

The summary could be improved to describe why this is the right way forward. It's not immediately obvious why collecting proofs during IBD is low-value, especially since we bother to track orphans.

Rebase on D11655 to fix linter issue, although it's not a direct dependency.

• Do not use global gArgs when it's unnecessary to do so
• Set -avaproofstakeutxoconfirmations=0 in relevant functional tests

The previous versions were broken and should not have gone up for review.
The fixes to unit tests were minimal so I should not have attempted to blindly
copy that logic from some of the other flags. This is now fixed and test plan
is updated.

Bump min confs back to 2016

Only blocking on the public API change.

• Make comments more accurate
• Made AVALANCHE_MAX_ORPHAN_PROOFS static
• Added missing test cases for replacement and trimming when pool is full

This diff is quite large. Please include a short rationale for why each backported commit was squashed into this one patch. That will help reviewers be more confident that it is complete.

Fix a bug where conflicts would unnecessarily trigger pool trimming.

In D11582#259592, @deadalnix wrote:

Maybe we should just evict proof for which we have no peer after some time instead. We kind of have to do it anyways, and this makes this patch moot as far as I can tell.