Update: Logic added to download the last 10 runs to then compare #220
Conversation
The idea is that currently only downloading the last 2 runs means that the script will break if the last 2 runs failed for some reason. Now, we download the last 10 runs, in which we pull out the one that is currently running and a previous run that has a status of success. This will ensure a consistent ability to compare runs.
|
You should have created an issue in this repo to propose and document this change in advance, so then any discussions could have been had and we would know what needed to be reviewed - this should always be done before putting up any pull requests |
There was a problem hiding this comment.
LTS support should be added at some point, but I can see that this was not supported in the previous code either, so it's not a regression.
This code/approach is quite different to how Robot gets the GM installer off GitHub, as we request the list of successful runs (and pre-ordered, I am sure), whereas you're getting all results and then manually filtering to find the successful ones. So again, I suspect there is an optimisation that could be made as a future change here.
Approving because there is a fault that we need this change now in order to fix, but I think we should look to rectify the above two points asap.
(And write up a proper issue first, next time! ;) )
The idea is that currently only downloading the last 2 runs means that the script will break if the last 2 runs failed for some reason.
Now, we download the last 10 runs, in which we pull out the one that is currently running and a previous run that has a status of success. This will ensure a consistent ability to compare runs.