Kyle Meyer <[email protected]> wrote: > Thanks, sorry about the mix up.
No worries, will push a regen. > [*] I did run `make check' before sending, and it looks like that also > checks MANIFEST _after_ running the test suite. That didn't help me > catch the MANIFEST sorting issue in this case because the test suite > is failing on my end. > > I believe these failures are recent and was planning on looking into > them today, either sending a patch or just reporting, depending on > whether I could figure out a fix. Anyway, here they are: > > t/psgi_multipart_not.t ....... > ok 1 - use HTTP::Request::Common; > ok 2 - use Plack::Test; > ok 3 - use PublicInbox::WWW; > not ok 4 - /v2test/?q=%22ain't what it used to be%22&x=t > not ok 5 - /v2test/?q=%22ain't what it used to be%22&x=t warns > Failed 2/5 subtests The failure is probably caused by 00d5dff2cce9d2c9 (eml: avoid Encode 2.87..3.12 leak, 2021-10-13) but I can't reproduce it across CentOS 7.x, FreeBSD 11.x, nor Debian 10 & 11. Which versions of Encode and Perl are you using? At least in Debian, libencode-perl is available as a separate package but it's also part of libperl5.xx (possibly w/ a different version); only the latter is required for us, but two packages offering the same thing gets confusing :/ I use: perl -MEncode -E 'say $Encode::VERSION' to determine which gets loaded. -- unsubscribe: one-click, see List-Unsubscribe header archive: https://public-inbox.org/meta/
