Allow endpoint to crash, without compromising future flush #1

Merged
merged 2 commits into from Feb 14, 2012

Conversation

Projects
None yet
2 participants
Contributor

cyberdelia commented Feb 8, 2012

pixel-ping isn't really nice when your endpoint is not accessible on flush, since future flushes are not made anymore. This pull-request makes it a bit more robust.

Owner

jashkenas commented Feb 8, 2012

Good start, but instead of losing data when it can't connect to the endpoint -- how about having it save the data for the next attempt?

Contributor

cyberdelia commented Feb 8, 2012

In my use case, I can just discard data, but definitely yes, we could keep data for next flush (or make a retry).
Can we merge this one as is ? I will try to keep data in another pull-request.

Owner

jashkenas commented Feb 8, 2012

Nope -- if you continue to push to your branch, the commits will be added to this pull request. No need to jump to another.

Contributor

cyberdelia commented Feb 8, 2012

I've added an discard option that allow to not re-send data after failure.

P.S: You can ignore the typo in the commit message

Contributor

cyberdelia commented Feb 14, 2012

Any thoughts on these changes ?

Owner

jashkenas commented Feb 14, 2012

Looks good.

jashkenas added a commit that referenced this pull request Feb 14, 2012

Merge pull request #1 from 20minutes/flush-crash
Allow endpoint to crash, without compromising future flush

@jashkenas jashkenas merged commit 1aa7a6d into documentcloud:master Feb 14, 2012

@xiongchiamiov xiongchiamiov referenced this pull request Oct 17, 2012

Merged

Update gh-pages branch #2

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment