Comment 1:

you might need a dual E5-2670 to see this.

Comment 2:

Another one:
=== RUN TestSetsRemoteAddr-48
--- FAIL: TestSetsRemoteAddr-48 (0.00 seconds)
serve_test.go:469:  Get error: Get /?n=10: stopped after 10 redirects

Comment 3:

=== RUN TestDirectoryIfNotModified-60
--- FAIL: TestDirectoryIfNotModified-60 (0.00 seconds)
fs_test.go:496: Got body "User-agent: go\nDisallow: /something/"; want "I am a fake
index.html file"

Comment 4:

Labels changed: added priority-later, removed priority-triage.

Status changed to Accepted.

Comment 5:

Labels changed: added go1.1.

Comment 6:

Still seeing this stuff. Here's another one:
--- FAIL: TestIdentityResponse-46 (0.00 seconds)
serve_test.go:350: for http://127.0.0.1:39895/?te= expected res.ContentLength of 3; got
-1
serve_test.go:353: for http://127.0.0.1:39895/?te= expected Content-Length header of
"3"; got ""
serve_test.go:357: for http://127.0.0.1:39895/?te= expected len(res.TransferEncoding) of
0; got 1 ([chunked])
serve_test.go:350: for http://127.0.0.1:39895/?te=identity expected res.ContentLength of
3; got -1
serve_test.go:353: for http://127.0.0.1:39895/?te=identity expected Content-Length
header of "3"; got ""
serve_test.go:357: for http://127.0.0.1:39895/?te=identity expected
len(res.TransferEncoding) of 0; got 1 ([chunked])
FAIL
It helps to disable TestStressSurpriseServerCloses (because it takes a while) and then
run:
#!/bin/bash
set -xe
go test -c net/http
while true; do
GOMAXPROCS=$[ 1 + $[ RANDOM % 1024 ]] ./http.test -test.v
done
Also, I've found that TestStressSurpriseServerCloses deadlocks from time to time. It
seems related to some kind of error condition that arises when the system has too many
sockets in the TIME_WAIT state. Running TestStressSurpriseServerCloses in a loop for
about 10 to 20 iterations produces this situation.

Comment 7:

Sorry, I didn't realize this was the same issue when I filed the dup issue #4436.
If you believe this isn't the same issue, feel free to reopen this one.

Status changed to Duplicate.

Merged into issue #4436.

Comment 8:

I think there are two bugs here:
1. TestStressSurpriseServerCloses was hanging, which prevents any of the other failures
from being reproduced
2. whatever is causing the other tests to fail.
I'll retest and report back.

Comment 9:

TestStressSurpriseServerCloses hanging (I've never seen that, but could believe it?) is
probably related to issue #4191 which has a fix but is waiting on related issue #4434.
I think (2) was already fixed by my httptest change
http://code.google.com/p/go/source/detail?r=cd3ad880