Hi chromium-dev@!
We’ll be turning on DCHECKs by default for developer release builds starting next week. We’re doing so to get better coverage of existing DCHECKs in paths not covered by automated testing. This will be done by changing the defaults for dcheck_always_on for non-official builds.
As the DCHECKs-by-default change is only intended for developers and because non-DCHECK paths are important to cover, we’ll change the existing buildbots that do not run with DCHECKs enabled today to use `dcheck_always_on = false`.
If you’re running infrastructure that currently builds without DCHECKs, you may want to consider updating your build configs to retain coverage of both DCHECK and non-DCHECK paths.
If you’re shipping binaries to users it’s still recommended that you build with `is_official_build = true`, otherwise consider using `dcheck_always_on = false` to retain existing behavior.
As we’re increasing exposure to DCHECKs to developers, a failing DCHECK is now a high-priority bug (P1/P0). It is also more likely that your changes get reverted if they introduce a failing DCHECK.
To try this early, build with `dcheck_always_on = false` and `is_debug = false`. If you hit any failing DCHECKs then file bugs. If they are bad enough that you think this should block rollout to fellow colleagues, indicate why and make this bug block crbug.com/1225701.
Thank you for your help!
Peter on behalf of DCHECKs.
Hi chromium-dev@!
We’ll be turning on DCHECKs by default for developer release builds starting next week. We’re doing so to get better coverage of existing DCHECKs in paths not covered by automated testing. This will be done by changing the defaults for dcheck_always_on for non-official builds.
As the DCHECKs-by-default change is only intended for developers and because non-DCHECK paths are important to cover, we’ll change the existing buildbots that do not run with DCHECKs enabled today to use `dcheck_always_on = false`.
If you’re running infrastructure that currently builds without DCHECKs, you may want to consider updating your build configs to retain coverage of both DCHECK and non-DCHECK paths.
If you’re shipping binaries to users it’s still recommended that you build with `is_official_build = true`, otherwise consider using `dcheck_always_on = false` to retain existing behavior.
As we’re increasing exposure to DCHECKs to developers, a failing DCHECK is now a high-priority bug (P1/P0). It is also more likely that your changes get reverted if they introduce a failing DCHECK.
To try this early, build with `dcheck_always_on = false` and `is_debug = false`. If you hit any failing DCHECKs then file bugs. If they are bad enough that you think this should block rollout to fellow colleagues, indicate why and make this bug block crbug.com/1225701.
Thank you for your help!
Peter on behalf of DCHECKs.
--
--
Chromium Developers mailing list: chromi...@chromium.org
View archives, change email options, or unsubscribe:
http://groups.google.com/a/chromium.org/group/chromium-dev
---
You received this message because you are subscribed to the Google Groups "Chromium-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chromium-dev...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAGFX3sEzQc0WA8CNx2-ycs_25JPuJq5WGZ28vLGhOHNqUmYfmA%40mail.gmail.com.
--
--
Chromium Developers mailing list: chromi...@chromium.org
View archives, change email options, or unsubscribe:
http://groups.google.com/a/chromium.org/group/chromium-dev
---
You received this message because you are subscribed to the Google Groups "Chromium-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chromium-dev...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAGFX3sEL0ug%3DHFj7rMWpHw2%3DpdVZWBtyebxcTG9c%2BAL8E7zGMA%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAKFU2SAeX3Bnb6ua9PK_5%2B%3DuxGWfnvoA0X_HSMZ%2Bvi4E%3D67SMA%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAEoffTAvwjPG5%3DHUYqtufBFDim8O2JnSmio4DoNt7jQM9wH6gA%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAOZqsm1OfDyBY%3DMi6HrMm5tHUzzywfqMxTLmEuSdj-LMZaQ9Aw%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAHri2_SnvQivSkqjhM7wm4F2qoCg632NE%2BHhXfJ8i70u%3D-X-QQ%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAJTZ7L%2BeGN0f%2Bxq4p_Ae0dwhan_0dO5tQm%3DGNG1amaB3%3DjoQdw%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAHq1K7Q7Kb923-AFtnC0iYLrS6cH7Fre-6V4kZXg%2B3Q8c3DHyA%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAEoffTAWA1QHTikgUmBX2M-OWZZpqEbuwAMsBP5dchCijsLf8g%40mail.gmail.com.
It can be more subtle than that. In my case it was that CHECK eats its stream under IS_OFFICIAL_BUILD so my death test expectation broke.This happened to an important CL I submitted just before going on 3 weeks holidays. It got reverted by a sheriff because one of its death tests failed on official bots... That CL was supposed to eliminate a class of failures on TSAN bots, instead of that: many more tests were disabled because the fix was reverted (and to this day I have no way to tell how many were disabled under my radar...).Re. each week? No, but the impact goes beyond "minor annoyance" to ultimately evolve into a culture like "avoid death tests". Seems it's cheaper to add that bot than have a way to measure whether we need it or not...
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAJTZ7LJfkW09BeOPFhv3zewjs7KA5dMsZGC-3vJmXEKGrixB%3DQ%40mail.gmail.com.
It can be more subtle than that. In my case it was that CHECK eats its stream under IS_OFFICIAL_BUILD so my death test expectation broke.This happened to an important CL I submitted just before going on 3 weeks holidays. It got reverted by a sheriff because one of its death tests failed on official bots... That CL was supposed to eliminate a class of failures on TSAN bots, instead of that: many more tests were disabled because the fix was reverted (and to this day I have no way to tell how many were disabled under my radar...).Re. each week? No, but the impact goes beyond "minor annoyance" to ultimately evolve into a culture like "avoid death tests". Seems it's cheaper to add that bot than have a way to measure whether we need it or not...
On Thu, Jan 27, 2022 at 1:23 PM Gabriel Charette <g...@chromium.org> wrote:It can be more subtle than that. In my case it was that CHECK eats its stream under IS_OFFICIAL_BUILD so my death test expectation broke.This happened to an important CL I submitted just before going on 3 weeks holidays. It got reverted by a sheriff because one of its death tests failed on official bots... That CL was supposed to eliminate a class of failures on TSAN bots, instead of that: many more tests were disabled because the fix was reverted (and to this day I have no way to tell how many were disabled under my radar...).Re. each week? No, but the impact goes beyond "minor annoyance" to ultimately evolve into a culture like "avoid death tests". Seems it's cheaper to add that bot than have a way to measure whether we need it or not...FWIW, this case is why we have EXPECT_CHECK_DEATH macros which should be used when looking for a CHECK() death.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAJTZ7LJZWWCk0WRWAh%3D3FpZnA5S-%3Djbrjga3ON222DvLc0GbVw%40mail.gmail.com.