On Wed Mar 21 22:24:25 2012, sprout wrote: > On Wed Mar 21 20:20:44 2012, sprout wrote: > > On Wed Mar 21 18:49:00 2012, Hugmeir wrote: > > > > I’ve begun work on the patches following that. I need to ask Brian > > > > Fraser about > > > > < > > > > https://github.com/Hugmeir/gsoc-pad-utf8- > > > safety/commit/4929775f6218457e97e9c11e8b1fcfce20b6316f > > > > >. > > > > Unicode delimiters are not going to go into 5.16, so this patch is > > > > unnecessary for now, right? > > > > > > > > > > I would say throw it out, yes. The one thing it might have fixed > > > (unintendedly, and definitely without tests) is the error message in > > > > > > use open qw( :utf8 :std ); > > > use utf8; > > > use 5.014; > > > > > > eval "q\x{FF01}asdasd\n\x{FF01}say 1"; > > > say $@; > > > > > > But that can be properly addressed later. > > > > Er, that might explain why the tests added by 6961eeb7b3 (toke.c: > > S_missingterm cleanup) are failing for me. :-) > > Yes, that patch for Unicode delimiter groundwork does fix those tests, > so it will go in. Except it causes failures in utf8.t and XS::APItest. So could you (Brian Fraser) try applying ‘toke.c: S_missingterm cleanup’ and ‘toke.c and parser.h: Make multi_(open|close) UVs’ to the latest blead and see why this: print <<"BANG", use utf8; %a = ("\xE1\xA0"=>"sterling"); print 'start'; printf '%x,', ord \$_ foreach keys %a; print "end\n"; BANG when run as: perl whatever | perl produces a screenful of messages instead of a single error? (I find the easiest way to apply unrelated patches like this is to use cherry-pick or, if it is not local, ‘curl https://github.com/Hugmeir/gsoc-pad-utf8-safety/commit/6961eeb7.patch | git am’. The commit IDs are 6961eeb7b3 and 4929775f62.) -- Father Chrysostomos --- via perlbug: queue: perl5 status: open https://rt.perl.org:443/rt3/Ticket/Display.html?id=107008Thread Previous | Thread Next