net-p2p/syncthing: work around ppc64 test failure

For some reason one of the Syncthing tests - TestIssue5063 - times out
on our ppc64 buildhost, with 100% reproducibility. Upstream were not
aware of the problem because apparently they have never run the test
suite on such a system, that said their suggestion to adjust the
time-out threshold appears to do the trick. Make it 20 s i.e. twice what
it originally is, hopefully this will prevent the problem from
resurfacing any time soon.

Closes: https://bugs.gentoo.org/717108
See-also: https://github.com/syncthing/syncthing/issues/6539
Signed-off-by: Marek Szuba <marecki@gentoo.org>

1 files changed, 18 insertions, 0 deletions

diff --git a/net-p2p/syncthing/files/syncthing-1.3.4-TestIssue5063_timeout.patch b/net-p2p/syncthing/files/syncthing-1.3.4-TestIssue5063_timeout.patch
new file mode 100644
index 000000000000..3148dc8a976c
--- /dev/null
+++ b/net-p2p/syncthing/files/syncthing-1.3.4-TestIssue5063_timeout.patch
@@ -0,0 +1,18 @@
+As of early April 2020, running the Syncthing test suite on the Gentoo ppc64
+buildhost fails due to TestIssue5063 timing out. Increase the time-out
+threshold for this test to allow it to pass.
+
+Other arches we currently support (i.e. amd64, arm and x86) shouldn't be
+affected by the change because thay have always passed so far anyway.
+
+--- a/lib/model/model_test.go
++++ b/lib/model/model_test.go
+@@ -914,7 +914,7 @@
+ 	}()
+ 	select {
+ 	case <-finished:
+-	case <-time.After(10 * time.Second):
++	case <-time.After(20 * time.Second):
+ 		pprof.Lookup("goroutine").WriteTo(os.Stdout, 1)
+ 		t.Fatal("Timed out before all devices were added")
+ 	}