> On 23 Mar 2023, at 18:15, Maxim Dounin <mdounin@mdounin.ru> wrote:
>
> Hello!
>
> On Wed, Mar 22, 2023 at 01:46:33PM +0400, Sergey Kandaurov wrote:
>
>>> On 18 Mar 2023, at 18:14, Maxim Dounin <mdounin@mdounin.ru> wrote:
>>>
>>> # HG changeset patch
>>> # User Maxim Dounin <mdounin@mdounin.ru>
>>> # Date 1679107816 -10800
>>> # Sat Mar 18 05:50:16 2023 +0300
>>> # Node ID 97b09b6633f69747c0d6ef13c76739bdd6b7f3bb
>>> # Parent 125fb8461d88a81a62ccb40d0e205a01ecc759f5
>>> Tests: separate SSL session reuse tests in mail.
>>>
>>> Instead of being mixed with generic SSL tests, session reuse variants
>>> are now tested in a separate file.
>>>
>>> [..]
>>> sub get_ssl_socket {
>>> - my ($port, $ses, $alpn) = @_;
>>> + my ($port, $ses) = @_;
>>>
>>> my $s = IO::Socket::INET->new('127.0.0.1:' . port($port));
>>> my $ssl = Net::SSLeay::new($ctx) or die("Failed to create SSL $!");
>>> Net::SSLeay::set_session($ssl, $ses) if defined $ses;
>>> - Net::SSLeay::set_alpn_protos($ssl, $alpn) if defined $alpn;
>>> Net::SSLeay::set_fd($ssl, fileno($s));
>>> Net::SSLeay::connect($ssl) == 1 or return;
>>
>> Since this won't fail anymore on ALPN checks,
>> you can convert this back to die.
>
> Strictly speaking, it shouldn't die regardless of whether
> connect() can legitimately fail during tests or not. If connect()
> fails (for example, due to a bug in some of the session reuse
> variants), it still should run other tests.
>

In general I agree with you.

> Fixing this (and other issues, such as missing timeout and SIGPIPE
> handling) is out of the scope of this patch series though.

Yes, let's postpone this.

--
Sergey Kandaurov
_______________________________________________
nginx-devel mailing list
nginx-devel@nginx.org
https://mailman.nginx.org/mailman/listinfo/nginx-devel