Release and regressions?


#1

According to this next release expecting to be July 20. I wonder what happens with stable to beta regressions, should they fixed before release, or it is possible to move them to the next release milestone?


#2

P-high issues are more-or-less considered ‘must fix right now’, and I expect the two on this list to be resolved one way or another.

Regressions generally get fixed prior to release, but some number of regressions, often minor, pretty much always slip through. They do not go onto the next-release milestone, but are rather categorized P-high/P-medium etc.

During this cycle we missed a release triage meeting that was scheduled during the Mozilla all-hands, and I personally have been behind on monitoring the release.

This set of regressions is not great.

I will do some triage tomorrow and try to get things in better order. The one assigned to me at least I still intend to fix, but … yeah it’s getting late.

cc @compiler_subteam @alexcrichton regressions!


#3

@brson

Any reason why I-wrong + regression-from-stable-to-beta is not automatically equal to P-high?

This set of regressions is not great.

Or, I thought that you guys have big dashboard with number of regression-from-stable-to-beta and P-high, with online update?


#4

Different bugs and regressions have different impact. P-high is the strongest signal we have on the issue tracker, and with limited resources, we can’t insist on diverting resources to fix every bug.

We don’t have a dashboard like that, but GitHub’s issue tracker can make such queries.


#5

I did big round of triage and posted a summary here: Rust 1.19 release status