I'm getting back into Hadoop coding, trying to get my little patches in while I come to terms with how git works, how the Hadoop code has changed in the year I've been away from it
One thing that has changed, Apache's JIRA server is now unusable from home, at least for submitting patches.
I know my network is having wifi congestion problems in the evenings (too many clashing 802.11g base stations nearby, no clean channel), but this is impossible to use.
For text. I can prepare it in an editor, then submit. If it times out, try again. What doesn't work is patch submission.
When I hit patch a fancy dialog pops pops up, lets me browse to a file, then attach it, then it spins for a period of time doing *something*, but not actually attaching it to the issue. It's only when I hit apply that it tries to attach the issue, and that keeps failing for me, saying "we timed out, we don't know what state the world is in, go back one page and reload it", which I do, to see that the attachment isn't there. So I have to try again, and again
What makes things worse is that I can browse to the issue in another tab. The network is there, jira is there. It's just posting patches that is taking so long it times out.
If the hadoop team weren't so fussy about code review, I'd commit the patches straight to SVN just to avoid JIRA