twbs / savage Goto Github PK

Any thought given to creating a new setting that allows members to bypass the whitelist check? Our organization is fairly large. We trust our devs enough to submit Pull Requests with any change, but not enough to give them write access to the repo. 😄

I'm comfortable making a change similar to the following in our instance, but was curious if you'd be interested in a PR back?

Snippet from https://github.com/twbs/savage/blob/master/src/main/scala/com/getbootstrap/savage/server/PullRequestEventHandler.scala#L109

...
case Success(affectedFiles) => {
  log.debug("Files affected by {}: {}", prNum, affectedFiles)
  if (areSafe(affectedFiles) || (trustedCanBypassWhitelistSetting && isTrusted(pr.user))) {
    if (areInteresting(affectedFiles)) {
...

Add an Unknown case.
We're getting errors because the "edited" action isn't being recognized.

https://github.com/twbs/bootstrap/blob/master/CONTRIBUTING.md#pull-request-bots
(Might need to split into subsections)

To my knowledge, our setup/configuration hasn't changed, but we now consistently get "Received Travis request with incorrect hash!"

This in turn causes the status / comment to not be updated / posted on the PR. Looking at Bootstrap, it appears maybe you're seeing the same behavior?

A user wishing to use the @twbs-savage retry syntax must not only be a member of the trusted organization, as is currently specified in the README.md, but they must be a public member.

I suppose in the spirit of this project, that makes sense. Alternatively, you could use isMember as opposed to isPublicMember.

At a minimum, I'm hoping #28 could be accepted to save future users the headache.

This is kind of a weird one and I'm not even sure that title makes sense. We just started seeing a few PRs hang forever and it looks like it's because we're unable to make a new branch to test from? I could be reading the logs wrong, but I was hoping someone had seen this one before.

Here's an example (retry) failure:

18:53:37.581 [on-spray-can-akka.actor.default-dispatcher-6462] WARN  c.g.savage.server.SavageWebService - Received Travis request with incorrect hash!
20:42:11.320 [on-spray-can-akka.actor.default-dispatcher-6495] INFO  c.g.s.server.PullRequestEventHandler - Initiating retry of PullRequestNumber(2395) due to request from trusted GitHubUser(mmcc)
20:42:11.540 [on-spray-can-akka.actor.default-dispatcher-6504] INFO  c.g.s.server.PullRequestEventHandler - Interesting path: /src/js/utils/create-deprecation-proxy.js
20:42:11.541 [on-spray-can-akka.actor.default-dispatcher-6504] INFO  c.g.s.server.PullRequestEventHandler - PR #2395 : Requesting build for safe & interesting PR
20:42:11.734 [on-spray-can-akka.actor.default-dispatcher-6503] INFO  c.g.savage.server.CommitStatusSetter - Successfully created commit status with state Pending and URL https://api.github.com/repos/videojs/video.js/statuses/02523a433250cedc986be517d0b9bc8c7f407ce7 for CommitSha(02523a433250cedc986be517d0b9bc8c7f407ce7)
20:42:12.121 [on-spray-can-akka.actor.default-dispatcher-6504] ERROR c.g.savage.server.PullRequestPusher - Error fetching from misteroneill/video.js:
Exit code: 1
error: 'refs/remotes/scratch/patch/fire-error-empty-source' exists; cannot create 'refs/remotes/scratch/patch'
From https://github.com/misteroneill/video.js
 ! [new branch]      patch      -> scratch/patch  (unable to update local ref)
error: some local refs could not be updated; try running
 'git remote prune https://github.com/misteroneill/video.js.git' to remove any old, conflicting branches

X-Ref: cvrebert/lmvtfy#98

Requires jcabi/jcabi-http#99

https://github.com/blog/1935-see-results-from-all-pull-request-status-checks

See http://sauceio.com/index.php/2014/11/test-automation-link-round-up/

From the logs:

ERROR c.g.savage.server.PullRequestPusher - Error while deleting .git/refs/remotes
java.nio.file.NoSuchFileException: .git/refs/remotes

Would love to get this working for video.js.

See http://doc.akka.io/docs/akka-http/current/scala/http/migration-guide/migration-from-spray.html
Spray has been superseded and is no longer maintained.

TravisPayload.matrix[N].number :: String (e.g. "1.2")

jcabi-github seems to, in general, extract IDs from nested JSON objects and issue redundant API requests to fetch those objects, rather that loading those objects from a nested portion of the already-returned JSON.

Prerequisite: #50, since several Spray and older Akka libraries don't have JARs for Scala 2.12

Problematic scenario:

1. User pushes revision abc
2. Savage pushes abc and starts building it
3. User force pushes revision def
4. Savage force pushes def and starts building it
5. The Travis build in (2) can fail at its fetching step because the revision no longer exists in Savage's repo due to the force push.