💊 CI failures summary and remediations

As of commit 885d148 (more details on the Dr. CI page):

💚 💚 Looks good so far! There are no failures yet. 💚 💚

@malfet has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

cc. @gdankel on the latest changes to kineto - looks like there're some build errors now

Yep, we need to set USE_KINETO_UPDATED (where?) together with the submodule update. @gdankel can you point to the place?

Checking!

Yep, we need to set USE_KINETO_UPDATED (where?) together with the submodule update. @gdankel can you point to the place?

no, we don't need to set it up (it's not set in oss on purpose) - we use it (btw, similarly to hhvm folks as I was told) to resolve a trivial issue when landing pytorch + bc-breaking kineto changes in both oss and internal repos (because oss version has a submodule at a fixed revision) - we just need to assume that USE_KINETO_UPDATED is defined, remove the usage of USE_KINETO_UPDATED, keep the code sections that are within USE_KINETO_UPDATED and discard the ones that are in the 'else' section

@ilia-cher isn't the process you used in the past to remove the old code and compile flag?

@ilia-cher isn't the process you used in the past to remove the old code and compile flag?

you're right, looks like we just need to remove !USE_KINETO_UPDATED sections

@ilia-cher isn't the process you used in the past to remove the old code and compile flag?

ideally, we just need to remove usages of USE_KINETO_UPDATED by updating the submodule after changes to the kineto were landed, USE_KINETO_UPDATED is supposed to be used only temporarily in between those two events

Yes exactly - it's part of updating submodule.

Like this #59767?

Should we just fork it then?

@malfet has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

@malfet merged this pull request in 827e00c.