39 bytes, text/x-review-board-request
I found a performance optimization in the new pushlog extension. Patch forthcoming.
Created attachment 8538948 [details] MozReview Request: bz://1113458/gps
Attachment #8538948 - Flags: review?(mh+mozilla)
/r/1601 - pushlog: optimize retrieval of pushlog data (bug 1113458) Pull down this commit: hg pull review -r 938d0d2bf978ba2d712bfb441110ae67a2210d51
Attachment #8538948 - Flags: review?(mh+mozilla) → review+
Status: NEW → RESOLVED
Last Resolved: 4 years ago
Resolution: --- → FIXED
And deployed to production. Brief log inspection reveals that pushlog wireproto commands are now using only milliseconds of wall/CPU time, as expected. Thanks for the quick review!
How does pushlog perf now compare to before we switched to the extension? Just wondering if we've helped bug 1079788 at all? Though I guess the next big win will be completing bug 1076826.
I don't expect perf via hgweb (json-pushes) to have changed as a result of this rewrite. However, switching consumers to version=2 and polling via startID and limiting the number of pushes returned should make a significant difference. Also, I wrote https://mozilla-version-control-tools.readthedocs.org/en/latest/hgmo/pushlog.html#writing-agents-that-consume-pushlog-data
Comment on attachment 8538948 [details] MozReview Request: bz://1113458/gps
Created attachment 8618938 [details] MozReview Request: pushlog: optimize retrieval of pushlog data (bug 1113458)
You need to log in before you can comment on or make changes to this bug.