Aleksa Sarai

@rata > so I assume the release 1.1 still has the same issue. No, #3599 was never backported to 1.1. > otherwise we release 1.2.0-rc.1 AND 1.2.0 without them I...

I just noticed the JSON field names for the idmap mounts don't match the style of the other fields. This is something we cannot change after we do a runc...

@h-vetinari I understand your point of view, however: > I think there's a clash between the mindset of "we shouldn't release until feature X or refactor Y lands" and time-based...

The code is already ready in #3985. Reverting all of the changes I've done over the past 4 months to fix this code is going to be significantly more work...

An rc1 needs to include the changelog in #4218 since it includes warnings about possible breakages. I can prepare an rc1 update PR with that included though.

I have several other changelog fixes that I've applied to my own branch (wording fixes and your copy of the 1.1.z changes is missing the links and a few other...

> @kolyshkin Never say never, and I would love to remove all the C code from our codebase, but I'm not sure if even on newer kernels and with the...

The binaries built for the release are all static AFAICS, so this isn't a problem with our regular builds: ``` % file release/1.1.8/* | grep ELF release/1.1.8/runc.amd64: ELF 64-bit LSB...

AFAIK there is no standard tooling to get the version information from the way we set versions in the resulting binary. Yes, the `-ldflags "-X ..."` thing is a common...

Yeah, I find it quite strange that `-trimpath` does more than it says on the tin -- it seems to me that `ldflags` is something that really shouldn't be stripped...