But I think what we should do it also add an object that safely creates good lock times (and storing previous one internally), and then people just forward this value from that object. Would also allow for future extensions.

In D16821#381670, @bytesofman wrote:

So i this case -- we are using locktime because it's an existing field that is available in script. But we are not really setting a locktime -- we are setting a sort of nonce, that should be a unique ID.

use ===

So i this case -- we are using locktime because it's an existing field that is available in script. But we are not really setting a locktime -- we are setting a sort of nonce, that should be a unique ID.

tabs

fix lints

rebase and fix join

Put this on hold for now, it makes things more complicated. If perf does turn out to be an issue in production we can reconsider doing this.

add MSG_UNDEFINED constant

This needs a rebase after D16812.

Simplify names, dismiss the key instead of nullifying

@bot build-win64 guix-win

Try to fix the windows build

One nit, otherwise looks good

please the compiler warnings
It is not clear to me why on these two the return statement can be omitted but not on the previous lines (93 and 96, 121 and 115)

Version 1.3.0

@bot build-ubsan build-asan build-tsan build-win64

Undo test util linkage

One nit

Otherwise looks good to me!