FWIW, I was involved in one of these bugs (bug 1185239). It is
completely unbelievable to me that we seem to not have a single test
that was broken by a change that would introduce a startup crash with an
existing profile.
As Kyle mentioned, it would be nice if someone worked on adding
automated tests for every single one of these issues. Not having landed
the patches in the first place is always better than having to back them
out.
On 2015-07-23 10:40 PM, Naoki Hirata wrote:
Currently our foxfood builds are blocked by two bugs. One gecko/one gaia:
https://wiki.mozilla.org/B2G/QA/Spark#Current_Versions
My question in regards to this is:
1) should we backout of smoke test blockers immediately after finding a
regression window?
2) should we backout of fox food blockers immediately after finding a
regression window?
3) How do we approach this for gecko? How do we approach this for gaia?
The reason why I ask, is that at times a forward fix is a better
solution than a back out. At the same time, this causes our fox fooders
to have to keep waiting. I started to keep track of dogfood breaks so
that I can gain some data in terms of breakages and how long it takes to
fix them : http://mzl.la/1CTpjL4 Sometimes the fixes could take over a
week's time to fix.
The last time we had a good build was 7/9. This is over two weeks ago.
I would like to remind the people that there is a demand for quality and
breaking builds on a constant basis is not a good sign for quality.
Regards,
Naoki
_______________________________________________
dev-gaia mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-gaia
_______________________________________________
dev-b2g mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-b2g
