Stefan Sperling wrote:
> The 1.6.17 release introduced a regression in write-through proxying
> that can break commits to paths which need to be URI-encoded (see
> r1088602 nomination in 1.6.x STATUS).
> This is a conundrum for 1.6.x users who serve write-through proxy
> locations as well as master locations from the same server. Either
> they risk repository corruption bugs fixed in 1.6.18, or they risk
> breaking write-through proxy commits.
> 1.6.17 was released in March. I'd like to ask how people feel about
> issuing another 1.6.x release to ship the currently pending fixes.
> We have a handful nominations for 1.6.x, including a client crash
> which can be triggered by malicious servers and a fix for a server hang.
> I'd volunteer to run the release, and give a week for voting.
> Any objections?
+1. When we fix non-trivial bugs, it's really important to the user community that we actually ship those fixes.
Received on 2012-08-30 18:13:16 CEST
This is an archived mail posted to the Subversion Dev