WT-13502 Add 'backup_checkpoint' to 'query_timestamp'.

[sueloverso]

Add a configuration setting to be able to query the timestamp that is the timestamp of the checkpoint pinned by an open backup cursor.

@luke-pearson and @keitharnoldsmith here is an implementation of what we discussed. It basically works but I didn't go into situations where timestamps aren't in use or checkpoints sometimes taken without the stable or weird cases like that. This is the basic functionality and plumbing. But I did make a test!

[github-actions]

Thanks for creating a pull request! The below questions and checklist are intended to help with verifying your change is well tested. Response is optional, but if you choose to respond please edit this comment.

What makes this change safe?

A good answer to this question helps the reviewers understand where they should focus their attention, so please consider these questions:

• Is the change risky or not? Why?
• What tests are you adding or changing? Why?
• What existing tests are you relying on?
• What, if anything, are you concerned about that you'd like the reviewer to focus on?
  References:
• Risk level guide
• Testing frameworks

Checklist before requesting a review

• I have performed a self-review of my code.
• I have made corresponding changes to the documentation (if applicable).
• I have added/updated tests that demonstrate my fix is effective or that my feature works correctly.

[tod-johnson-mongodb]

LGTM. I didn't hear the discussion so I didn't mark it approved.

[luke-pearson]

Closing as I don't think this solution is relevant anymore.

[sueloverso]

Reopening this PR because it is now relevant.

[sueloverso]

@keitharnoldsmith @luke-pearson and @tod-johnson-mongodb this PR has been revived and is now part of the solution for the server code. It is now pulled to the current state of develop and ready for review.

[wiredtiger-prbot]

Test coverage is too low, this change probably shouldn't be merged as-is.

Metric (for added/changed code)	Coverage
Line coverage	33% (1/3)
Branch coverage	10% (1/10)

[sueloverso]

Right now the code will return a WT_TS_NONE timestamp if a backup cursor isn't opened. It could return an error instead.

[sueloverso]

@keitharnoldsmith and @tod-johnson-mongodb can you please review this? It is real now and this PR has been reopened.

[luke-pearson]

LGTM, i think one more comment in txn_timestamp.c may be helpful. See https://github.com/wiredtiger/wiredtiger/pull/11011/files#r1801689373