That sounds like we should just turn it off, but I'll let the others decide if it's worth it or not.

In general, for shipping configs we have a debug builder on the main waterfall. Why would we skip it for this config?

Because it seems like we haven't needed it?

debug bots for a config is how we get dcheck coverage (since release on main waterfall doesn't have them). Otherwise we're in a situation where a dcheck fires in 64 bit windows builds and CQ is broken while main waterfall is green. To prevent this, all bots on CQ that run tests have debug+release bots on main waterfall running tests. I don't see what's different here, i.e. even if we were lucky in the last year, this seems like a good rule to follow.

I understand the theory you are promoting :).

FWIW, we don't actually have an x64 debug bot running tests on the main waterfall, either. So, in order to prevent the mismatch you're referring to, we'd have to add a bot that runs tests as well. Without that, the compile only bot is buying us very little. My guess is that we've probably seen very few (if any) win x64-specific DCHECKs.

*shrug*. I still would like to see what wfh@ and brucedawson@ think.

I think we should delete the chromium_builder_tests group, or else give it some meaning.

As for what this builder should do, I have no idea. If there is a gap that it is supposed to be filling then it is not clear what that gap is and therefore what this builder should be doing.

@brucedawson - the gap is that we don't have anything building or running tests on x64 debug.

Gotcha. I understand now. A lack of x64 debug coverage on the CQ seems non-ideal. On the other hand, we have other configurations and tests (and targets?) that aren't run on the CQ and those seem to allow bad code through more often than our lack of an x64-debug builder/tester. So, this doesn't bother me particularly since I think there are other places we could more profitably put more CQ resources (although I can't remember what failures have recently made it through the CQ)

@brucedawson: to be clear, this bug isn't about the CQ. This is for main waterfall. If the main waterfall doesn't run these tests, there'll be zero coverage on this config.

Apparently my reading comprehension is poor on this bug. Okay, that makes me more nervous. Even though debug x64 failures are rare, they are probably not unprecedented, and symmetry suggests we should have a compile/test pair if that is practical.

But, although it makes me a bit nervous I don't have strong feelings about it, given that we have debug and x64 coverage (separately).

This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue.

Sorry for the inconvenience if the bug really should have been left as Available.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

bot builds something nowadays