[XviD-devel] Next iA64-patches

Stephan Krause s_kraste at ira.uka.de
Tue Jun 17 22:27:50 CEST 2003


Trying xvid_stat on the mentioned Itanium2 leads to 3 Problems:

1. In xvid_stat.c max_bframes isn't initalized. And it got really big when i 
tested ;)

,-------------

RCS file: /xvid/xvidcore/examples/xvid_stat.c,v
retrieving revision 1.22
diff -r1.22 xvid_stat.c
886a887
>       xparam.max_bframes=0;

'--------------

2. In decoder.c someone tests for 
if((idct == simple_idct_mmx) && (dec->bs_version < 10)) /* rather ugly but 
should work */

Yea, it should work. IF we had MMX. Else simple_idct_mmx isn't defined... A 

,-------------------

RCS file: /xvid/xvidcore/src/decoder.c,v
retrieving revision 1.50
diff -r1.50 decoder.c
1842a1843,1844
> #ifdef ARCH_IS_IA32
> 
1845c1847,1848
< 
---
> #endif
> 
'-------------------

removes at least all *my* problems with this...


3. The usual segfault. Someone trying to add something negative that isnt a 
64bit int to a pointer.

,-------------------

Index: motion/motion_est.c
===================================================================
RCS file: /xvid/xvidcore/src/motion/motion_est.c,v
retrieving revision 1.70
diff -r1.70 motion_est.c
466c466
<       sad = sad32v_c(data->Cur, data->RefP[0] + (x>>1) + 
(y>>1)*(data->iEdgedWidth),
---
>       sad = sad32v_c(data->Cur, data->RefP[0] + (int) ((x>>1) + 
(y>>1)*(data->iEdgedWidth)),

'-------------------

MfG
Stephan

-- 
Sig fault. (core dumped)


More information about the XviD-devel mailing list