Don't depend on github.com/crc-org/vfkit/pkg/rest

[mtrmac]

Warning: Completely untested, and it would have been better authored as a commit just copying the code, + a set of refactors. I’ll happily close this, and leave full credit to, anyone who wants to do this properly.

Via github.com/gin-gonic/gin , this depends on several large encoding / decoding packages, including a JIT compiler. Maintaining <60 lines of code ourselves seems well worth it.

Note that this dependency chain only exists on darwin.

Impact: 927 files changed, 103 insertions(+), 505542 deletions(-)
dropping

github.com/bytedance/sonic
github.com/bytedance/sonic/loader
github.com/cloudwego/base64x
github.com/cloudwego/iasm
github.com/gabriel-vasile/mimetype
github.com/gin-contrib/sse
github.com/gin-gonic/gin
github.com/go-playground/locales
github.com/go-playground/universal-translator
github.com/go-playground/validator/v10
github.com/goccy/go-json
github.com/klauspost/cpuid/v2
github.com/leodido/go-urn
github.com/pelletier/go-toml/v2
github.com/twitchyliquid64/golang-asm
github.com/ugorji/go/codec
golang.org/x/arch

Cc: @Luap99

Does this PR introduce a user-facing change?

None

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: mtrmac

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [mtrmac]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[Luap99]

@baude PTAL would you have any concerns with that?
I have not tried to look up the history of this but the diff makes we very much in favour, the code seems trivial to maintain on our side.

[Luap99]

AFAIK podman always seem to use an http API socket so this could be simplified even further.

(I might need to check why we use a tcp socket at all, unix seems much better suited for local IPC)

[mtrmac]

Yes, I see things like mc.AppleHypervisor.Vfkit.Endpoint = localhostURI + ":" + strconv.Itoa(randPort), but that’s a few layers away from the call site.

I’m all in favor of simplifying further, but I’ll leave that to Podman experts.

[mtrmac]

Note that make-and-check-size reports exactly 0 size delta on Linux.

[mtrmac]

… and another view is that this removes 21.75 % of total lines in vendor.

[baude]

wfm

[mtrmac]

contrib/cirrus/pr-should-include-tests is a fair comment here, or, as a review notes above, maybe we don’t need all that parsing anyway.

@Luap99 I think I’ll leave this as a prototype for now, and I’ll let you finish / replace this as part of the larger size-reducing work.

[Luap99]

@Luap99 I think I’ll leave this as a prototype for now, and I’ll let you finish / replace this as part of the larger size-reducing work.

Sure, works for me.

[mtrmac]

Note to self: #25720 is a superset, this should be closed after that merges.

[mtrmac]

Thanks @Luap99 !