I downloaded and built from all the latest sources and the only warnings
which remain are these (also present in previous release builds):
sljit/sljitNativeX86_32.c(145) : warning C4761: integral size mismatch
in argument; conversion supplied
sljit/sljitNativeX86_common.c(497) : warning C4761: integral size
mismatch in argument; conversion supplied
sljit/sljitNativeX86_common.c(705) : warning C4761: integral size
mismatch in argument; conversion supplied
sljit/sljitNativeX86_common.c(845) : warning C4761: integral size
mismatch in argument; conversion supplied
Regards,
Sheri
On 2/12/2012 12:14 PM, Philip Hazel wrote:
On Sat, 4 Feb 2012, Sheri wrote:
Following are the 13 warnings that appeared during the build:
C:\pcre-8.30\pcre-8.30\pcre_exec.c(6216) : warning C4146: unary minus operator
applied to unsigned type, result still unsigned
I have applied the following patch in an attempt to fix this issue in
pcre_exec.c. I can't test it, because my compiler does not complain
about the use of -sizeof(xxx). (It presumably auto-casts it. It
definitely returns a negative number.)
Index: pcre_exec.c
===================================================================
--- pcre_exec.c (revision 911)
+++ pcre_exec.c (working copy)
@@ -6208,12 +6208,14 @@
const REAL_PCRE *re = (const REAL_PCRE *)argument_re;
/* Check for the special magic call that measures the size of the stack used
-per recursive call of match(). */
+per recursive call of match(). Without the funny casting for sizeof, a Windows
+compiler gave this error: "unary minus operator applied to unsigned type,
+result still unsigned". Hopefully the cast fixes that. */
if (re == NULL&& extra_data == NULL&& subject == NULL&& length == -999&&
start_offset == -999)
#ifdef NO_RECURSE
- return -sizeof(heapframe);
+ return -((int)sizeof(heapframe));
#else
return match(NULL, NULL, NULL, 0, NULL, NULL, 0);
#endif
===================================================================
Please can you run the command "pcretest -m -C" before and after
applying the patch. I expect the output to be different.
The other warnings are all in the JIT code -- Zoltán, did you fix these
with your latest patches?
C:\pcre-8.30\pcre-8.30\pcre_jit_compile.c(6507) : warning C4244: 'function' :
conversion from 'const unsigned short ' to 'unsigned char ', possible loss of
data
C:\pcre-8.30\pcre-8.30\pcre_jit_compile.c(6514) : warning C4244: 'function' :
conversion from 'const unsigned short ' to 'unsigned char ', possible loss of
data
C:\pcre-8.30\pcre-8.30\pcre_jit_compile.c(6507) : warning C4761: integral size
mismatch in argument; conversion supplied
C:\pcre-8.30\pcre-8.30\pcre_jit_compile.c(6514) : warning C4761: integral size
mismatch in argument; conversion supplied
C:\pcre-8.30\pcre-8.30\sljit/sljitNativeX86_32.c(145) : warning C4761:
integral size mismatch in argument; conversion supplied
C:\pcre-8.30\pcre-8.30\sljit/sljitNativeX86_common.c(497) : warning C4761:
integral size mismatch in argument; conversion supplied
C:\pcre-8.30\pcre-8.30\sljit/sljitNativeX86_common.c(705) : warning C4761:
integral size mismatch in argument; conversion supplied
C:\pcre-8.30\pcre-8.30\sljit/sljitNativeX86_common.c(845) : warning C4761:
integral size mismatch in argument; conversion supplied
C:\pcre-8.30\pcre-8.30\pcre_jit_test.c(700) : warning C4090: 'function'
: different 'const' qualifiers
C:\pcre-8.30\pcre-8.30\pcre_jit_test.c(700) : warning C4022: 'pcre_fullinfo' :
pointer mismatch for actual parameter 4
C:\pcre-8.30\pcre-8.30\pcre_jit_test.c(894) : warning C4090: 'function'
: different 'const' qualifiers
C:\pcre-8.30\pcre-8.30\pcre_jit_test.c(894) : warning C4022: 'pcre_config' :
pointer mismatch for actual parameter 2
Philip
--
## List details at https://lists.exim.org/mailman/listinfo/pcre-dev
