Uh oh!
There was an error while loading. Please reload this page.
Relax some asserts in SpanHelpers.T.cs - #1472
Conversation
Uh oh!
There was an error while loading. Please reload this page.
jkotas
commented
Jan 9, 2020
It makes me worried that you are seeing the asserts, but we are not catching them in the CI. @dotnet/runtime-infrastructure Are there any tests runs that run debug/checked flavors of everything? |
trylek
commented
Jan 9, 2020
Both the pre-checkin and the outerloop pipeline only run release bits. AFAIK the only runs that use checked product bits are GC stress and JIT stress runs. |
jkotas
commented
Jan 9, 2020
The dotnet/coreclr repo had coverage for running corefx tests on checked coreclr in the CI. This coverage was proven to be important and caught many bugs early. Is there a workitem to add this back? What it would take to add this back? |
trylek
commented
Jan 9, 2020
I know we discussed that with @BruceForstall when I was working on the PR but looking now I cannot find an explicit item - Bruce, do you know whether there is one or whether I should create a new one? |
safern
commented
Jan 9, 2020
With the single pipeline effort that I'm doing this will be simplified. I might just include running corefx tests on checked coreclr (regular runs, not stress) when coreclr is touched. |
BruceForstall
commented
Jan 9, 2020
I have a work item to add back libraries testing on checked coreclr for coreclr checkins, but I've been waiting for the merged repo and live/live CI changes to stabilize before doing that... and maybe @safern's work will make it easy or do it for me. |
safern
commented
Jan 9, 2020
Yeah, let's sync up after I'm done, I have some ideas. |
jkotas
commented
Jan 10, 2020
@dotnet/runtime-infrastructure A lot of legs are failing with "RestApiException: The response contained an invalid status code 500 Internal Server Error". I am not able to figure out what the problem is - can't find any logs with details. Could you please take a look? |
dagood
commented
Jan 10, 2020
That's happening across a bunch of PRs, including one I'm working on. Filed https://github.com/dotnet/core-eng/issues/8591 to track. |
safern
commented
Jan 10, 2020
So in my PR I'm adding some corefx test runs with a checked runtime and it did hit this issue: I will pull up this change and see if it fixes it. |
safern
commented
Jan 11, 2020
@halter73@jkotas after pulling the changes into a PR only Here is the dump: |
jkotas
commented
Jan 11, 2020
Let's revert the change that added the asserts for now, so that you can get the changes that are fixing the CI coverage in: #1628 |
stephentoub
commented
Jan 15, 2020
Should this be closed now? |
jkotas
commented
Jan 15, 2020
Opened #1767 to track the follow up. |
I saw the following assert failure in System.Memory.Tests
Because of the aggressive inlining, MemoryExtensions.LastIndexOf (a public extension method on Span) doesn't show up in the stack trace, but that's what was calling into the generic SpanHelpers.LastIndexOf method instead of a char-specialized version. I fixed MemoryExtensions.LastIndexOf to call a newly-added char-specialized SpanHelpers.LastIndexOf method.
I also noticed that here are no char-specialized SpanHelpers.LastIndexOfAny methods, so I relaxed the asserts in the generic SpanHelpers.LastIndexOfAny methods.
@jkotas This is a follow up to #1200. Is there an easy way to discover if there are any more assert failures before merging?