Martin Buchholz wrote:
On Sat, Aug 22, 2009 at 20:37, Xueming Shen<xueming.s...@sun.com> wrote:
-------------
31 @@ -39,7 +63,7 @@
32 typedef unsigned int u4;
33 # else
34 # if (ULONG_MAX == 0xffffffffUL)
35 - typedef unsigned long u4;
36 + typedef uLong u4;
37 # else
38 # if (USHRT_MAX == 0xffffffffUL)
39 typedef unsigned short u4;
Using uLong in the above is probably not right,
since comparison against ULONG_MAX means
the corresponding type is unsigned long
(not that it matters)
# if (UINT_MAX == 0xffffffffUL)
typedef unsigned int u4;
# else
# if (ULONG_MAX == 0xffffffffUL)
typedef unsigned long u4;
# else
# if (USHRT_MAX == 0xffffffffUL)
typedef unsigned short u4;
I believe that on all platforms where the JDK will be built,
the first test UINT_MAX == 0xffffffffUL will be true,
so the suggested change will never pass the preprocessor.
I believe it to both be wrong and to have no effect, and increase the
size of local changes - but still OK to commit if you insist!
Oh, I see your point. Are you saying this "local change" is not
necessary since it never gets
preprocessed? OK, I admit I did the replace all in emacs:-) Maybe I
should have touched
those that affected. But given the nature of crc32, I'm pretty sure a
32-bit unsigned integer
is what it should be (and is intended,) as the code purposely defines
the "u4".
An alternative is to keep the crc32.c/h untouched, change the crc32()
declare in zlib.h to
unsigned long and then change the rest of jdk (our CRC.c and the pack
code to prepare
a 64-bit unsigned long on 64-bits), is this the direction you would
prefer to?
Sherman
