On Oct 15, 2007, at 2:35 AM, Ralf S. Engelschall wrote:

> On Sun, Oct 14, 2007, Jeff Johnson wrote:
>
>> Grrr, everyone has to have their own private typedefs for ints.
>> [...]
>>     From: Tomasz Mateja <tommat@pimpek.one.pl>
>>     Date: October 14, 2007 3:54:33 PM EDT
>>     To: Jeff Johnson <n3npq@mac.com>
>>     Subject: Re: [Th] rpm@sparc64
>>
>>     Jeff Johnson napisaÅ (a):
>>
>>         Could you try to verify that SizeT is actually "size_t"  
>> and not uint32?
>>         73 de Jeff
>>
>>     I think I've got the solution :)
>>
>>     --- rpm-4.4.9/rpmio/LzmaDecode.h.orig   2007-10-14  
>> 19:23:02.629594398
>>     +0000
>>     +++ rpm-4.4.9/rpmio/LzmaDecode.h        2007-10-14  
>> 19:23:16.628653630
>>     +0000
>>     @@ -35,7 +35,7 @@
>>     Â /* #define _LZMA_LOC_OPT */
>>     Â /* Enable local speed optimizations inside code */
>>
>>     -/* #define _LZMA_SYSTEM_SIZE_T */
>>     +#define _LZMA_SYSTEM_SIZE_T
>>     Â /* Use system's size_t. You can use it to enable 64-bit  
>> sizes supporting*
>>     /
>>
>>     Â #ifndef UInt32
>>
>>     I need to test this but seems to work. I wonder why this is  
>> not on by
>>     default. Thank you Jeff very much, this is great step forward  
>> for PLD@sparc
>> [...]
>
> This _IS_ the default -- at least in RPM 5 ;-) That's exactly the  
> change
> I committed to RPM 5's LzmaDecode.h some time ago. Using size_t is the
> right thing here nowadays.
>

You're good. ;-)

Time to backport the fix to rpm-4_5. todo++, ditto releasing rpm-4.5.

Perhaps time to rip int_32 et al private typedefs in header.h, use  
<stdint.h> instead. Yawn ...

BTW, it's likely time to split rpm-4.5 and rpm-5.0 doxygen doco on  
the web site.
No hurry whatsoever.

73 de Jeff