@rvagg I think you accidentally included some unrelated (http) changes in your commit.

doh! that was from testing #7211, thanks for picking that up @mscdex

@rvagg Thanks for picking up on this! I couldn't find any documentation about it anywhere (e.g. on the TAP website) but I can see that the Jenkins TAP parser understands duration_ms so I guess there's no alternative.

I really should have documented this earlier @gibm, I went though this exact pain very early on because it frustrated me too. I ended up in the Jenkins source code just to verify that it relied on it in this form. I think it's just one of those standards that have appeared and been widely enough adopted to be stuck in a semi-official state.

LGTM

LGTM
(and thanks @rvagg ... my apologies when I reviewed the original change I had completely forgotten that it was a tap thing...)

LGTM

LOOKS GOOD TO ME

@rvagg Looks like there's a PR that will go into a future version of the TAP plugin which requires duration_ms output to be in milliseconds (node's is currently in seconds).

jenkinsci/tap-plugin#6

Not sure how we'd deal with this.

EDIT: It might be worth adding a duration_s option to the Jenkins tap plugin, but you'd still have a breaking change.

oh my, that's terrible, I've added a comment over there

1d4c799 & 05de4d7