Front page | perl.perl5.porters |
Postings from October 2014
On Mon, Sep 29, 2014 at 2:25 PM, Karl Williamson
<public@khwilliamson.com> wrote:
> I can't figure out the problem here, and would appreciate some help. I will
> comment out the offending line for the time being; it's just a compile-time
> check that we're not overflowing. But these changes have been thoroughly
> smoked, and passed, and I don't understand the Jenkins error message and how
> it applies here.
>
> The error message is
> op_reg_common.h:104:30: error: missing binary operator before token
> "2147483647"
>
> The failing line is
> # if RXf_PMf_COMPILETIME > I32_MAX
I can trigger this if I undef INT32_MAX:
+++ b/handy.h
@@ -222,7 +222,7 @@ typedef U64TYPE U64;
# define UINT64_C(c) PeRl_UINT64_C(c)
# endif
#endif
-
+#undef INT32_MAX
#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
And then build, which gives me a much more helpful error message:
$ make
echo @`sh cflags "optimize='-O2'" perlmini.o` -DPERL_IS_MINIPERL
-DPERL_EXTERNAL_GLOB perlmini.c
@cc -c -DPERL_CORE -fwrapv -fno-strict-aliasing -pipe
-fstack-protector -I/usr/local/include -D_LARGEFILE_SOURCE
-D_FILE_OFFSET_BITS=64 -std=c89 -O2 -Wall -ansi -Werror=$
perl.h:2532:30: error: missing binary operator before token "2147483647"
# define PERL_INT_MAX ((int)INT_MAX)
^
handy.h:260:18: note: in expansion of macro ‘PERL_INT_MAX’
# define I32_MAX PERL_INT_MAX
^
op_reg_common.h:104:30: note: in expansion of macro ‘I32_MAX’
# if RXf_PMf_COMPILETIME > I32_MAX
^
make: *** [perlmini.o] Error 1
This appears to be because I32_MAX is then defined to PERL_INT_MAX,
which uses a cast - (int) - which is compile time not preprocessor
time.
Why INT32_MAX isn't available on that system I don't know, but that
shouldn't matter.
Is there something other than I32_MAX that we can use at compile time
to check size/value here?
-- Matthew Horsfall (alh)
Thread Previous
|
Thread Next