Conversation
|
When cherry-picking this onto production, I ran into a merge conflict. While easy enough to resolve, it does make me concerned that there have been changes to the files since. How extensively has this been tested on top of the existing production build? |
|
@sparhami I got conflict only in test files. Did you get something else? |
|
@sparhami Can we cherrypick the change to canary first. And only cherrypick to prod once we've verified the fix in inabox RC. |
I just want to make sure there aren't any other potentially conflicting changes. The lack of a merge conflict in the non-test files does not guarantee this. Since the original issue wasn't covered by a test, relying on tests seems to not be sufficient. For example, a hypothetical PR submitted after the PR that is being rolled back, that depends on it could run into issues once we rollback. |
It seems like there might be another cherry-pick request for canary. I'm currently waiting to see if that will happen to combine the two before doing a canary cherry-pick. |
Sounds good. FYI: While the change is in prod already, but it's not affecting ampdoc. The change is still in inabox canary, and we need the canary cherry-pick mainly for inabox RC. |
|
Looks like a test is failing when cherry picking on top of production, in single pass, local tests and sauce lab tests: https://travis-ci.org/ampproject/amphtml/builds/551509527 This seems to fail every time I try it locally. Here is the branch with the cherry pick: Here is the cherry picked commit on top of prod: d0eb22b The changes should be the same as this PR. |
|
That's weird. Let me try running it locally |
|
Hello @sparhami found the reason. unexpected comma when we resolve merge conflict. |
The PR caused side effect on some reporting due to the new "nextTick" behavior in navTiming macro.
/cc @jeffkaufman