Conversation
|
Thanks for your pull request, @MartinNowak! Bugzilla referencesYour PR doesn't reference any Bugzilla issue. If your PR contains non-trivial changes, please reference a Bugzilla issue or create a manual changelog. Testing this PR locallyIf you don't have a local development environment setup, you can use Digger to test this PR: dub fetch digger
dub run digger -- build "master + dmd#7344" |
|
As I recall, back when (Kenji?) tried it it substantially slowed down the compiler. I want to review the best way forward on this first. I think the non-release mode inserts too many checks. |
One way forward is |
Shameless self-plug: https://github.com/dlang/DIPs/blob/master/DIPs/DIP1006.md |
|
I like the idea of finally moving forward with DIP1006! BTW one minor thing: if we start to use a different compiler build (non-release) than the one we ship on the auto-tester, how do we make sure that the version we ship doesn't get accidentally broken? |
|
Guess you can close #6746 then. |
Yes, we could simply do this on a faster CI (e.g. SemaphoreCI is twice as fast as Travis - I checked this recently. Compare SemaphoreCi build, Travis build).
FYI: our CircleCi build already runs without
|
I would say more. On a good day, gdc is built with testsuite and unittests ran in 14 minutes flat on semaphore. Travis barely could do the same in 50 minutes. |
|
What I found implementing #7980, we could hack this already today. |
Heh, wouldn't that be a bug? Or maybe not. It does seem surprising though. |
|
FWIW CircleCI and Azure already run the test suite with debug builds. Is this still required? |
|
Don't think so. Also we should have it in the release, but that's another topic. |
I guess we'll still need to fix a couple of things and might have to disable class invariants to reduce the overhead.