tools: bump gyp to eb296f6

[refack]

Bump GYP to eb296f6
Re float all patches since last bump - 1d65b99

Checklist

• make -j4 test (UNIX), or vcbuild test (Windows) passes
• commit message follows commit guidelines

Affected core subsystem(s)

tools,build,gyp

[refack]

CI: https://ci.nodejs.org/job/node-test-pull-request/7432/

[addaleax]

deps: float gyp patch for long filenames 592f89a

It should be safe to drop this now, the patch has been applied upstream

build: enable cctest to use generated objects 6ed488a

This commit seems noticeably different from the original? It might be a hint for solving @gibfahn’s #12448, though…

[refack]

/cc @nodejs/build @addaleax @shigeki @jbergstroem @danbev
Do you know of any older than 1d65b99 patches that might need to be reapplied?

[refack]

build: enable cctest to use generated objects 6ed488a

This commit seems noticeably different from the original? It might be a hint for solving @gibfahn’s #12448, though…

Was a mixed core + GYP commit (see my #11956 (comment)) I picked just the GYP part

[refack]

deps: float gyp patch for long filenames 592f89a

It should be safe to drop this now, the patch has been applied upstream

Dropped

[addaleax]

Was a mixed core + GYP commit (see my #11956 (comment)) I picked just the GYP part

I see that, but including it like this still seems a little misleading… at least the commit message and the PR-URL should probably be different here, to avoid confusion with the original commit.

[refack]

I see that, but including it like this still seems a little misleading… at least the commit message and the PR-URL should probably be different here, to avoid confusion with the original commit.

gotcha

[refack]

@addaleax bb49593 better?

[refack]

As for the PR URL, I think it should stay since it's the actual motivation for the change. I'm just re-applying it in the context of GYP
I guess both should be there.

[shigeki]

I only know the XCode floating patch that is included in this PR. I'm not sure if other else patches are needed.
As for the XCode patch, @bnoordhuis submitted the patch to the upstream in https://codereview.chromium.org/1213123002/ but it's been stalled.

[joaocgreis]

tools/gyp/buildbot/aosp_manifest.xml, tools/gyp/pylib/gyp/generator/android.py and tools/gyp/pylib/gyp/generator/compile_commands_json.py are not present upstream, should be deleted in the first commit.

Isn't 3c46bb9 needed as well?

@refack about the partial commit message, it would be good to use a format similar to 5111e78, with the original message indented. Metadata should be at the end, so that core-validate-commit passes, and you probably want to avoid @-mentioning in the commit message because you'll get a notification every time this is rebased or cherry-picked.

[aqrln]

@refack

As for the PR URL, I think it should stay since it's the actual motivation for the change. I'm just re-applying it in the context of GYP

I believe there's no need in the original metadata since the original commit hash is present in the commit message and it's easy to do git show locally or just click a link on GitHub. Yeah, by the way, that hash is wrong in your last commit.

[refack]

@joaocgreis 3c46bb9 added. As for the partial commit, I'm with @aqrln, same tree, so just a pointer.

[refack]

ReJenk: https://ci.nodejs.org/job/node-test-pull-request/7451/

[aqrln]

@refack still wrong commit hash. 6a09a69 is what you want ;)

[refack]

@refack still wrong commit hash. 6a09a69 is what you want ;)

GitHub is a mysterious place, where does f5a9b16 exist?

[addaleax]

where does f5a9b16 exist?

$ git for-each-ref --contains f5a9b16910878e35eb36f551ed6790c2ed8006f1
2d9967c1c6177c6e9273d6a87bd5ea99571afd74 commit	refs/remotes/upstream/pr/11956

→ #11956, where it came from :)

[refack]

So GitHub creates pseudo-branches for PRs, independent of originating branch... I suspected so after I accidently patched #12231

[aqrln]

@refack as Anna has just noted, in the PR, obviously :) But you don't want to link to the commit from the PR branch, you rather need the commit that has been landed to master.

UPD: oops, I think I'm late for a second with my answer.

[refack]

@refack I think it's been long enough, could you raise a backport PR against v6.x-staging that contains all the relevant commits you mentioned above?

Should I add 444920d (enabling VS2017) as @joaocgreis requested?

[gibfahn]

Should I add 444920d (enabling VS2017) as @joaocgreis requested?

I'm assuming you mean 1c93e8c (the commit you mentioned doesn't have the PR-URL).

Yes why not, that landed a while back, and has been in 8 since 8.0.0. Are there any outstanding bugs/issues with it?

[refack]

3fe980d (typo)
&
780acc2 (tooling edge case)
Both are minor

[gibfahn]

I meant bugs that haven't been fixed (and can't be backported as part of the same PR) 😁 .

Go ahead and open it, if people have issues they can comment on the PR.

[MylesBorins]

ping re: v6.x backport