Add options for lock statement timeout

[arthurprs]

This change was needed to get pg_repack to get the locks in our reasonably busy servers.

[schmiddy]

Thanks, I think a configurable lock_wait_max is useful. I'm less enthused about adding yet another option for --lock-wait-step, this seems like something we should be able to guesstimate on our own, without contributing to the explosion of command-line options.

What if we just had the lock wait step be 10% of the lock_wait_max, with some minimum like 100 ms.

[arthurprs]

I don't have strong opinions but I'd suggest keeping the step parameter as it's strictly more useful (arguably to a small group but it doesn't matter to the other).

Defaulting wait-step to 10% of max-wait sounds nice a nice default though.

[arthurprs]

I finally had the time to revisit this. Change updated as you suggested.

[andreasscherman]

Thanks for the pull request @arthurprs! This is exactly what I was looking for --- we have a highly contested server and was in need of having a variable timeout for the lock.

@schmiddy would you mind merging this?

[hstefan]

We have a similar patch internally, and would greatly appreciate having it merged so that we can use the upstream package.

[andreasscherbaum]

@arthurprs Can you please rebase this PR? This will trigger the Checks.
Also the documentation is not updated with the new flag.