2164: RB rejects my diff file
- Fixed
- Review Board
| kste*****@gmai***** (Google Code) (Is this you? Claim this profile.) | |
| June 24, 2012 |
What version are you running? 1.6 beta 2 (dev) What's the URL of the page containing the problem? /r/new/ What steps will reproduce the problem? 1. Create a diff with git: git diff --full-index HEAD^ >diff.diff 2. Upload the diff to review board What is the expected output? What do you see instead? Expected: Review board accepts the diff and opens a new review request. Actual: It complains: The file 'ContentViewer/CVDataSource.m' (r6db717bd95e4e128dd2fe2f848b082a7c175e047) could not be found in the repository What operating system are you using? What browser? OSX 10.6.8 Chrome 12.0.742.112 Please provide any additional information below. The file ContentViewer/CVDataSource.m exists in github, and has for months. I have made minor modifications to this file, which are present in the diff.
We don't just need the file in the repository, we need the specific revision you're diffing against. Is HEAD^ pushed to github as well?
-
+ NeedInfo
Update: I tried manually removing the entry for CVDataSource.m from the diff, after which it complained about the very next file in the diff. Maybe it has to do with the number of modified files in the diff? It's a rather large one (190k).
HEAD is pushed to github.
It's strange... I've tried uploading a number of times. About 20% of the time it fails immediately, complaining that the first file in the diff doesn't exist. The rest of the time, it spends a long time thinking (about 20 seconds) before complaining about whatever file occupies the 61st spot in the diff. I can make any file fail simply by putting it in the 61st position.
Hmm. This sounds suspiciously like github is trying to rate limit.
Our github integration just changed a whole lot, and the v3 API has much higher limits, so hopefully this should be fixed.
-
- NeedInfo + Fixed