[00:42] <wgrant> cjwatson: How extensively did you test your requests timeout port? I've possibly found a regression in the new urlfetch, possibly decoding excessively.
[00:42] <wgrant> Am digging.
[00:46] <cjwatson> Not very extensively.
[00:47] <wgrant> gpg key reactivation crashes, as the urlfetch result is not a str.
[00:47] <wgrant> I assume it must be a unicode.
[00:48] <cjwatson> Huh, I seem to not have ever received the review you posted eight days ago.
[00:48] <cjwatson> Weird.
[00:48] <cjwatson> I think my testing was limited to the test suite.
[00:49] <cjwatson> Possibly when I originally put it together I tried it against turnip too.
[00:50] <wgrant> Which review did I allegedly post eight days ago?
[00:50] <wgrant> Oh.
[00:50] <wgrant> timeout-with-requests never landed.
[00:50] <wgrant> Then what on earth broke this.
[00:51] <cjwatson> Indeed, that review.
[00:51] <cjwatson> Must have somehow deleted it by mistake
[00:52] <wgrant> Works on prod, so I guess a keyserver firewall or something.
[00:52] <wgrant> How odd.
=== xnox_ is now known as xnox
=== blr_ is now known as blr