[browser][http] Set HttpResponseMessage.RequestMessage

[campersau]

Fixes #42691

[ghost]

Tagging subscribers to this area: @dotnet/ncl
See info in area-owners.md if you want to be subscribed.

[lewing]

Looks good

[stephentoub]

Are we running our HTTP tests in wasm? I'm wondering if we have a test hole here, or if we do in fact have a test that would catch this but it didn't because we're not running them.

[lewing]

wasm Http tests in CI are waiting on #40786 but we've done recent manual runs that did not catch this so I suspect a test hole.

[stephentoub]

but we've done recent manual runs that did not catch this so I suspect a test hole.

Hmm. I just searched the tests. It's true that there aren't any dedicated to this, and I can fix that. But there are a bunch of tests that just assume this is non-null and would null ref if it was, so something about those manual runs must have skipped all those tests.

[lewing]

Agreed, we're standing up the tests as quickly as possible.

@kjpou1 please look into why the tests aren't catching this.

[lewing]

/backport release/5.0

[lewing]

The issue with the testing here is that browser can't support the loopback server because we don't have sockets and can't listen for requests so all the test that require the loopback server will throw PNSE and have to be skipped.

[stephentoub]

so all the test that require the loopback server will throw PNSE and have to be skipped.

Oookeeyyy... that's the vast majority of the functional tests ;-)

[lewing]

so all the test that require the loopback server will throw PNSE and have to be skipped.

Oookeeyyy... that's the vast majority of the functional tests ;-)

Yeah... I think we're the first platform to hit this but I don't think we will be the last.

[lewing]

Opened #42852 for discussion

[lewing]

Lane failure looks like jaredpar/runfo#47 and this change is browser specific