if the network service is running in a different process, we can also just crash :D

the main reason i'm not doing that is because it takes down the browser process

I don't think we want a compromised renderer process to be able to crash the network process.

true.

Well, so here's the issue. Every now and then, somebody commits code that incorrectly uses referrers, e.g., sending a https referrer to an http request. And typically, they didn't write tests for this, so a DCHECK() doesn't trigger.

This DumpWithoutCrashing thing has been quite effective in finding those anyways. I mean, we could just strip incorrect referrers, but it feels to odd to have code that just blindly corrects bugs without leaving any trace

I'm completely fine with keeping the DCHECK, if it's useful.  Just wanted confirmation that we should be keeping it.  We could even consider moving the logic into the mojo serialization code for the ResourceRequest, if we wanted, so we'd have a call stack in the consumer process.

Erm...Keeping the DCHECK == keeping the DumpWithoutCrashing.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/10e6cadecfcf0a58ce2a2240e88180d96ec303d9

commit 10e6cadecfcf0a58ce2a2240e88180d96ec303d9
Author: Matt Menke <mmenke@chromium.org>
Date: Wed Jun 20 12:54:20 2018

NetworkService: Make requests with bad referrers fail.

This moves logic for this case from ChromeNetworkDelegate to
NetworkContext's NetworkDelegate, and adds a Mojo parameter to cause
requests with bad referrers to fail (enabled by default).

Also removes the action Net.URLRequest_StartJob_InvalidReferrer rather
than moving it, as it was unowned.

Bug:  852871 
Cq-Include-Trybots: luci.chromium.try:ios-simulator-full-configs;luci.chromium.try:linux_mojo;master.tryserver.chromium.mac:ios-simulator-cronet
Change-Id: I9d6d55157ba28e2b12ca7136b78330100a8673aa
Reviewed-on: https://chromium-review.googlesource.com/1103119
Reviewed-by: Gayane Petrosyan <gayane@chromium.org>
Reviewed-by: David Roger <droger@chromium.org>
Reviewed-by: Tom Sepez <tsepez@chromium.org>
Reviewed-by: Jochen Eisinger <jochen@chromium.org>
Commit-Queue: Matt Menke <mmenke@chromium.org>
Cr-Commit-Position: refs/heads/master@{#568807}
[modify] https://crrev.com/10e6cadecfcf0a58ce2a2240e88180d96ec303d9/chrome/browser/net/chrome_network_delegate.cc
[modify] https://crrev.com/10e6cadecfcf0a58ce2a2240e88180d96ec303d9/chrome/browser/net/network_context_configuration_browsertest.cc
[modify] https://crrev.com/10e6cadecfcf0a58ce2a2240e88180d96ec303d9/ios/chrome/browser/net/ios_chrome_network_delegate.cc
[modify] https://crrev.com/10e6cadecfcf0a58ce2a2240e88180d96ec303d9/net/base/layered_network_delegate.cc
[modify] https://crrev.com/10e6cadecfcf0a58ce2a2240e88180d96ec303d9/net/base/layered_network_delegate.h
[modify] https://crrev.com/10e6cadecfcf0a58ce2a2240e88180d96ec303d9/net/base/layered_network_delegate_unittest.cc
[modify] https://crrev.com/10e6cadecfcf0a58ce2a2240e88180d96ec303d9/services/network/network_context.cc
[modify] https://crrev.com/10e6cadecfcf0a58ce2a2240e88180d96ec303d9/services/network/network_context_unittest.cc
[modify] https://crrev.com/10e6cadecfcf0a58ce2a2240e88180d96ec303d9/services/network/public/mojom/network_context.mojom
[modify] https://crrev.com/10e6cadecfcf0a58ce2a2240e88180d96ec303d9/tools/metrics/actions/actions.xml