Subtract output values from offset when inscription spent as fee

[gmart7t2]

Fixes #1841.

Indexes will need rebuilding.

[veryordinally]

@raphjaph this looks good to me - should prioritize for review.

[unisky]

I have found this bug too, it made a big confusion to my current data. hope prioritize for review.

[gmart7t2]

I have found this bug too

Your fix is identical to mine too. That's reassuring to see.

[unisky]

hello guys, when to merge this commit and fix history data? I think this bug is very fatal

[casey]

I think this change is correct. In addition to a test, can you please describe what the bug is, why this fixes it, and the observable change in behavior? There was a lot of discussion in #1841, but I think it also discusses a different issue, so it's a bit confused. If there's already a different issue describing this, you can also link that issue.

[gmart7t2]

I think this change is correct. In addition to a test, can you please describe what the bug is, why this fixes it, and the observable change in behavior? There was a lot of discussion in #1841, but I think it also discusses a different issue, so it's a bit confused. If there's already a different issue describing this, you can also link that issue.

The bug is #1841.

https://github.com/casey/ord/issues/1841#issuecomment-1483950070 explains the bug and why this fixes it.

The observable change in behavior is that ord tracks inscriptions correctly when they are spent as change rather than tracking them incorrectly.

I don't understand the confusion here. Sorry if I'm being slow.

[raphjaph]

I think the bug is very well explained! I just added the changes you made to the test in a new test so it is very clear what the bug was. This is ready for you to have a look @casey!

[casey]

Looks good! For bug fixes, existing tests should not be changed if they aren't broken, new tests should be added, which gives confidence that old and correct behavior, covered by the old tests, isn't being changed.