Created attachment 699090 [details] dev-python:waitress-2.0.0:20210410-221052.log =dev-python/waitress-2.0.0 fails its test phase, specifically the test DispatcherWithSendTests.test_send. (I wondered if this might just be a timeout, and tried both with a tripled n= value in this test function and with FEATURES=sandbox turned off, and neither helped.)
Created attachment 699093 [details] emerge-info.txt
I have seen exactly the same test failure while attempting to keyword this ebuild for riscv. Cannot reproduce it on any amd64 systems, though.
I'm still/again seeing this with waitress-2.1.1 on one of my two machines, while it passed its test phase on the other. I thought to check dmesg after the most recent attempt to build it, and I see "TCP: request_sock_TCP: Possible SYN flooding on port 58699. Dropping request. Check SNMP counters." (I'm not 100% sure that's correlated with this package, but I can't think what else might cause it ...)
I'm still seeing this with version 2.1.2, and I again noticed the dmesg message about possible SYN flooding.
Created attachment 858145 [details, diff] patch for deselecting the problematic test I also experience this problem with =dev-python/waitress-2.1.2 (on amd64), with the "Possible SYN flooding" in dmesg. If I disable the offending test (see attached patch), all other tests pass.
I can't reproduce here but I see nothing wrong with deselecting it. Thanks!
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=c7f2cb2ea645f19916351a58d341c58ecf6de981 commit c7f2cb2ea645f19916351a58d341c58ecf6de981 Author: Michał Górny <mgorny@gentoo.org> AuthorDate: 2023-03-18 06:56:31 +0000 Commit: Michał Górny <mgorny@gentoo.org> CommitDate: 2023-03-18 06:56:31 +0000 dev-python/waitress: Deselect a problematic test Thanks to Jonathan Lovelace for the report and the suggestion how to deal with it. Closes: https://bugs.gentoo.org/782031 Signed-off-by: Michał Górny <mgorny@gentoo.org> dev-python/waitress/waitress-2.1.2.ebuild | 5 +++++ 1 file changed, 5 insertions(+)
I just updated the upstream issue (https://github.com/Pylons/waitress/issues/405#issuecomment-1567036727): After setting CONFIG_SYN_COOKIES=y in Linux kernel config, the test passes, but still generates "Possible SYN flooding" in dmesg.