dreamcatcher-tech / artifact

GNU Affero General Public License v3.0
0 stars 1 forks source link

read should not do a write #194

Open inverted-capital opened 3 months ago

inverted-capital commented 3 months ago

This needs to be repeatable, so the same result is returned when run again, and validators can use a range of acceptable results to determine what they will accept.

So it is like a constrained side effect, where we know the results are within a boundary.

Right now, we have to do a write to the branch to do a read, which is very bad.