Unfortunately I learned the hard way (e.g. haukex/IPC-Run3-Shell#5) that tests should not expect an exact number of warnings, but t/01-mojo.t does. The issue shows up e.g. here: https://travis-ci.org/Corion/future-http/jobs/452344391. My experience has been to either test for the presence of specific warnings that one's own code produces, or to test for the absence of specific warnings that should not happen.
I could provide a patch - which warnings is t/01-mojo.t looking for specifically? Edit: Oh, I see: #2, more specifically mojolicious/mojo#1221. Since it seems the issue was fixed in 555906a7a, maybe it'd be best to remove the "no warnings" test, or make it specific to Use of uninitialized value $chunk in concatenation? Edit 2: Possible patch provided in #7.
haukex
commented
6 years ago
Unfortunately I learned the hard way (e.g. haukex/IPC-Run3-Shell#5) that tests should not expect an exact number of warnings, but
t/01-mojo.tdoes. The issue shows up e.g. here: https://travis-ci.org/Corion/future-http/jobs/452344391. My experience has been to either test for the presence of specific warnings that one's own code produces, or to test for the absence of specific warnings that should not happen.I could provide a patch - which warnings is
t/01-mojo.tlooking for specifically? Edit: Oh, I see: #2, more specifically mojolicious/mojo#1221. Since it seems the issue was fixed in 555906a7a, maybe it'd be best to remove the "no warnings" test, or make it specific toUse of uninitialized value $chunk in concatenation? Edit 2: Possible patch provided in #7.