The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/fab11cdbc638822aac069e23db9417d028e076a2

commit fab11cdbc638822aac069e23db9417d028e076a2
Author: jdoerrie <jdoerrie@chromium.org>
Date: Thu Apr 13 15:07:24 2017

Revert of When HttpNetworkTransaction encounters QUIC errors, retry the request (patchset #3 id:40001 of https://codereview.chromium.org/2818623002/ )

Reason for revert:
Likely cause of compilation error in Win x64.

BUG= 711316 

Original issue's description:
> When HttpNetworkTransaction encounters QUIC errors, retry the request
> with alt-svc disabled, and if that succeeds then mark QUIC as broken.
>
> Protected by the retry_without_alt_svc_on_quic_errors finch param.
>
> BUG=705033
>
> Review-Url: https://codereview.chromium.org/2818623002
> Cr-Commit-Position: refs/heads/master@{#464391}
> Committed: https://chromium.googlesource.com/chromium/src/+/d0dbccf1f8d23dfa32defdd02c0a3df990729358

TBR=jri@chromium.org,ianswett@chromium.org,rch@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=705033

Review-Url: https://codereview.chromium.org/2811993005
Cr-Commit-Position: refs/heads/master@{#464403}

[modify] https://crrev.com/fab11cdbc638822aac069e23db9417d028e076a2/components/network_session_configurator/network_session_configurator.cc
[modify] https://crrev.com/fab11cdbc638822aac069e23db9417d028e076a2/components/network_session_configurator/network_session_configurator_unittest.cc
[modify] https://crrev.com/fab11cdbc638822aac069e23db9417d028e076a2/net/http/http_network_session.cc
[modify] https://crrev.com/fab11cdbc638822aac069e23db9417d028e076a2/net/http/http_network_session.h
[modify] https://crrev.com/fab11cdbc638822aac069e23db9417d028e076a2/net/http/http_network_transaction.cc
[modify] https://crrev.com/fab11cdbc638822aac069e23db9417d028e076a2/net/http/http_network_transaction.h
[modify] https://crrev.com/fab11cdbc638822aac069e23db9417d028e076a2/net/quic/chromium/quic_network_transaction_unittest.cc

rch@, I reverted your CL as a likely culprit for the compilation error. Do you mind taking another look at this?

Given that my CL landed via the CQ and passed win 64 bots, this is super surprising. In addition, I didn't change anything related to net/http/transport_security_state_static_unittest2.h 

How confident are you that my CL was the culprit?

It is possible that the CQ bots had a pre-existing generated version of gen/net/http/transport_security_state_static_unittest*.h and thus this wasn't caught by them. Given that your CL was the only one touching files in //net/ in the hours before the error and the build passed after reverting your CL I am quite confident that your CL was indeed the culprit.

However, please feel free to reassign this bug if you think another CL is responsible.

If the CQ bots have a pre-existing generated version of gen/net/http/transport_security_state_static_unittest*.h and my CL is actually the culprit (in a win 64 specific fashion) do you have a suggestion for how to repro this? I don't really want to land again and just cross my fingers, but I'm not sure what the alternative is. :(

Looking at the Win 64 bot, my CL was first built by: 

https://build.chromium.org/p/chromium/builders/Win%20x64/builds/10396

This build succeeded.

The bot failed on the next build:

https://build.chromium.org/p/chromium/builders/Win%20x64/builds/10397

My CL was not on the blamelist for that build

The next build after that was:

https://build.chromium.org/p/chromium/builders/Win%20x64/builds/10398

Which was successful again and did NOT include the revert of my CL. The revert was first built by:

https://build.chromium.org/p/chromium/builders/Win%20x64/builds/10399


So if my CL is at fault, it means we have some sort of non-determinism in the build :(

Ok, I re-landed and the build looks stable so I'm going to close this bug.