Hi Max,
Do I have the green light to go for it? :)
Your contributions are very welcome, please do.
I also commented out a bunch of erlang test suite test cases recently which
failed against Java. Not sure if that is related to the TLS thing but could
be, at least each of them involved "ssl":
"java-erl_binary_buffered-ip-ssl",
"java-erl_binary_fastframed-framed-ip-ssl",
"java-erl_binary_framed-ip-ssl",
"java-erl_compact_buffered-ip-ssl",
"java-erl_compact_fastframed-framed-ip-ssl",
"java-erl_compact_framed-ip-ssl",
"java-erl_multi-binary_buffered-ip-ssl",
"java-erl_multi-binary_fastframed-framed-ip-ssl",
"java-erl_multi-binary_framed-ip-ssl",
"java-erl_multic-compact_buffered-ip-ssl",
"java-erl_multic-compact_fastframed-framed-ip-ssl",
"java-erl_multic-compact_framed-ip-ssl",
So that may answer your other question, contribs welcome here as well.
Thank you,
JensG
-----Ursprüngliche Nachricht-----
From: max ulidtko
Sent: Tuesday, September 22, 2020 12:53 PM
To: [email protected]
Subject: Re: Erlang CI again
Hi Jens,
I'm quite sure this is caused by rebar2 → rebar3 upgrade (rebar is the
"package manager" for Erlang, it also drives unit tests). You could guess
from a major version bump, that such an upgrade requires changes to the
project files.
I have a few tweaks to the erl infra collecting dust in a local branch (I
think I explicitly downgraded to rebar2 when making them, to minimize code
churn). BTW, there I also have a workaround for erl↔java TLS1.3 handshake
issue, is that resolved on master already?..
Moving forward, I think I can fix this issue by paying some care to the
erlang infra. Do I have the green light to go for it? :)
Max
On Mon, Sep 21, 2020 at 11:35 PM Jens Geyer <[email protected]> wrote:
Hi all,
https://travis-ci.org/github/apache/thrift/jobs/728933858
Anyone any idea what happens here? Seems as if the Thrift compiler
generates Erlang code, but for some reason the generated files are not
found when running the tests?
Thanks,
JensG
