Mon Nov 19 06:21:18 2012: Request 81172 was acted upon.
Transaction: Correspondence added by peter.smoczyn...@i-new.com
Queue: Inline
Subject: Re: [rt.cpan.org #81172] using Inline with => 'Event' fails to
compile
Broken in: (no value)
Severity: (no value)
Owner: Nobody
Requestors: peter.smoczyn...@i-new.com
Status: open
Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=81172 >
Am 17.11.2012 01:58, schrieb Sisyphus via RT:
> <URL: https://rt.cpan.org/Ticket/Display.html?id=81172 >
>
> On Fri Nov 16 06:51:09 2012, peter.smoczyn...@i-new.com wrote:
>
>> in fact it's the space(!) and not the double quotes!
>>
>> i.e.
>>
>> "-I /opt/perl5.12.2.debug/lib/site_perl/5.12.2/i86pc-solaris/Event"
>>
>> vs.
>>
>> "-I/opt/perl5.12.2.debug/lib/site_perl/5.12.2/i86pc-solaris/Event"
> Well spotted - I had missed that there was a space stuck in there.
>
> So ... the space between the -I and the /opt doesn't cause any problems
> until the whole thing is enclosed in double-quotes ? (Please correct me
> if that's wrong.)
yes, the space between the -I and the /opt does not cause any problems
-- that is correct!
might be reminiscent to the old k&r days..?
> I would have expected that space to be a problem even if there *weren't*
> any enclosing double-quotes.
> But, if that's not the case, then quote_space() probably needs to take
> this into consideration.
>
> I'm thinking quote_space() should be able to do this at the end of its
> processing with
>
> s/"\-I\s+\//"-I\//g
>
> since any spaces between the leading -I and the first / can never be
> meaningful (right ?).
>
> If you want to give that a try, at the end of the quote_space()
> subroutine in C.pm, change:
>
> my $out = join '', @in;
> $_[0] = $out;
>
> to:
>
> my $out = join '', @in;
> out =~ s/"\-I\s+\//"-I\//g;
> $_[0] = $out;
>
> I haven't tested, however, so I'm being a bit premature here ...
I did change the end of quote_space() to
my $out = join '', @in;
$out =~ s/"\-I\s+\//"-I\//g;
$_[0] = $out;
which works fine for me:
[...]
Starting "make" Stage
/opt/perl5.12.2.debug/bin/perl
/opt/perl5.12.2.debug/lib/5.12.2/ExtUtils/xsubpp -typemap
"/opt/perl5.12.2.debug/lib/5.12.2/ExtUtils/typemap" -typemap
"/opt/perl5.12.2.debug/lib/site_perl/5.12.2/i86pc-solaris/Event/typemap"
compile_test_1_t_a7b1.xs > compile_test_1_t_a7b1.xsc && mv
compile_test_1_t_a7b1.xsc compile_test_1_t_a7b1.c
gcc -c -I"/root/peter/inline_fails_to_compile"
"-I/opt/perl5.12.2.debug/lib/site_perl/5.12.2/i86pc-solaris/Event"
-I/usr/local/include -I/usr/local/BerkeleyDB.4.5/include -DDEBUGGING
-fno-strict-aliasing -pipe -fstack-protector -I/usr/local/include
-D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -DPERL_USE_SAFE_PUTENV -g
-DVERSION=\"0.00\" -DXS_VERSION=\"0.00\" -fPIC
"-I/opt/perl5.12.2.debug/lib/5.12.2/i86pc-solaris/CORE"
compile_test_1_t_a7b1.c
Running Mkbootstrap for compile_test_1_t_a7b1 ()
chmod 644 compile_test_1_t_a7b1.bs
rm -f blib/arch/auto/compile_test_1_t_a7b1/compile_test_1_t_a7b1.so
gcc -G -L/usr/local/lib -Wl,-R/usr/local/lib
-L/usr/local/BerkeleyDB.4.5/lib -Wl,-R/usr/local/BerkeleyDB.4.5/lib
-fstack-protector compile_test_1_t_a7b1.o -o
blib/arch/auto/compile_test_1_t_a7b1/compile_test_1_t_a7b1.so
/opt/perl5.12.2.debug/lib/site_perl/5.12.2/i86pc-solaris/auto/Event/Event.so
\
\
chmod 755 blib/arch/auto/compile_test_1_t_a7b1/compile_test_1_t_a7b1.so
cp compile_test_1_t_a7b1.bs
blib/arch/auto/compile_test_1_t_a7b1/compile_test_1_t_a7b1.bs
chmod 644 blib/arch/auto/compile_test_1_t_a7b1/compile_test_1_t_a7b1.bs
Finished "make" Stage
Starting "make install" Stage
Files found in blib/arch: installing files in blib/lib into architecture
dependent library tree
Installing
/root/peter/inline_fails_to_compile/_Inline/lib/auto/compile_test_1_t_a7b1/compile_test_1_t_a7b1.bs
Installing
/root/peter/inline_fails_to_compile/_Inline/lib/auto/compile_test_1_t_a7b1/compile_test_1_t_a7b1.so
Finished "make install" Stage
Starting Cleaning Up Stage
Finished Cleaning Up Stage
Finished Build Compile Stage
compile_test_1.t syntax OK
Sun-A3 root 260/0 (~/peter/inline_fails_to_compile)#
>
> There's also the case that
>
> -I /foo
>
> would be changed to
>
> -I" /foo"
>
> if it was the first element in the string, so that needs to be looked
> at, too.
right
> I'll release 0.51_01 in a couple of days time, and let you know when
> that happens. It will contain, among other things, a fix for this issue.
>
> Thanks for the report, and for the digging.
>
> Cheers,
> Rob
you're welcome!
cheers, peter
