On Tue, 07 Jun 2016 17:17:56 +0100, Dominic Hargreaves wrote:
> Test Summary Report
> -------------------
> t/dm-dateformats.t (Wstat: 65280 Tests: 0 Failed: 0)
> Non-zero exit status: 255
> Parse errors: Bad plan. You planned 33 tests but ran 0.
> Files=42, Tests=948, 135 wallclock secs ( 0.47 usr 0.37 sys + 125.12 cusr 8.43 csys = 134.39 CPU)
> Result: FAIL
>
> This appears to be related to this:
>
> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=796944#62
>
> Right now, I can't spot which change in 5.20.2-3+deb8u5 caused the
> problem, but applying the change reference in that bug report, attached,
> fixes the FTBFS.
Confirmed, both the breakage and the fix.
I also don't understand what happens here. My investigations show:
The problem is (where the patch is :)) in
lib/Biber/Input/file/bibtex.pm in the _hack_month() sub.
This is called twice from t/dm-dateformats.t, once with "14" and once
with "january". And what actually breaks is, in the case with
"january", the call to
Unicode::GCString->new($1)->substr(0,3)->as_string
If I add "print $1;" or "my $foo = $1;" before or if I change the
call to ->new("$1") or ->new($1."") [yes I tried silly things], then
all is fine. Or with the patch which uses the new $m variable.
Unicode::GCString is not in core, and in perldebdelta I didn't find
anything obvious, but maybe the above gives someone else a clue.
Cheers,
gregor
--
.''`. Homepage https://info.comodo.priv.at/ - OpenPGP key 0xBB3A68018649AA06
: :' : Debian GNU/Linux user, admin, and developer - https://www.debian.org/
`. `' Member of VIBE!AT & SPI, fellow of the Free Software Foundation Europe
`- NP: David Bowie: Ground Control to Major Tom
Attachment:
signature.asc
Description: Digital Signature