Thread: plperl fails with perl 5.14.0
It appears from the buildfarm configure logs that member anchovy was updated yesterday from perl 5.12.3 to 5.14.0. It doesn't like plperl.c anymore: ccache cc -flto -march=amdfam10 -fstack-protector -O3 -pipe -Wall -Wmissing-prototypes -Wpointer-arith -Wdeclaration-after-statement-Wendif-labels -Wformat-security -fno-strict-aliasing -fwrapv -g -fpic -I. -I. -I../../../src/include-D_GNU_SOURCE -I/usr/include/libxml2 -I/usr/lib/perl5/core_perl/CORE -c -o plperl.o plperl.c plperl.c: In function 'plperl_trusted_init': plperl.c:929:36: error: lvalue required as left operand of assignment make[3]: *** [plperl.o] Error 1 What it's unhappy about is evidently this: GvCV(sv) = NULL; /* prevent call via GV */ I think we are going to have to find a different way of doing that... regards, tom lane
There was some discussion on this about 6 weeks ago, at which time 5.14 wasn't released, so I didn't want to apply the patch then. I'll look at applying the fix now. cheers andrew On 06/04/2011 12:30 PM, Tom Lane wrote: > It appears from the buildfarm configure logs that member anchovy was > updated yesterday from perl 5.12.3 to 5.14.0. It doesn't like plperl.c > anymore: > > ccache cc -flto -march=amdfam10 -fstack-protector -O3 -pipe -Wall -Wmissing-prototypes -Wpointer-arith -Wdeclaration-after-statement-Wendif-labels -Wformat-security -fno-strict-aliasing -fwrapv -g -fpic -I. -I. -I../../../src/include-D_GNU_SOURCE -I/usr/include/libxml2 -I/usr/lib/perl5/core_perl/CORE -c -o plperl.o plperl.c > plperl.c: In function 'plperl_trusted_init': > plperl.c:929:36: error: lvalue required as left operand of assignment > make[3]: *** [plperl.o] Error 1 > > What it's unhappy about is evidently this: > > GvCV(sv) = NULL; /* prevent call via GV */ > > I think we are going to have to find a different way of doing that... > > regards, tom lane >
On 06/04/2011 01:05 PM, Andrew Dunstan wrote: > > There was some discussion on this about 6 weeks ago, at which time > 5.14 wasn't released, so I didn't want to apply the patch then. > > I'll look at applying the fix now. > > (and apologies for top-replying ;-) ) cheers andrew
Andrew Dunstan <andrew@dunslane.net> writes: > There was some discussion on this about 6 weeks ago, at which time 5.14 > wasn't released, so I didn't want to apply the patch then. Oh, right, I thought the issue seemed familiar: http://archives.postgresql.org/message-id/BANLkTi=qRfqpPfQTO2R5amb_cFsYHdVWwQ@mail.gmail.com It looks like we were waiting to see if they'd put GvCV_set() into ppport.h. If they didn't, Alex's fix looks sane to me. regards, tom lane