[Lsb-infrastructure] changes in devchk

[Wichmann, Mats D]

Denis Silakov wrote:
> Wichmann, Mats D wrote:
>> And it seems that devchk takes a segmentation fault
>> shortly after starting the first step of the test -
>> Stew reproduced this manually:
>> 
>> Checking data structures in GL/gl.h
>> Segmentation fault
>> 
>> 
>> No chance to look at what's happening yet.
>> 
> 
> I think I know the reason. It seems that devchk on our side works well
> due to this change which was made after release:
> 
> UPDATE Constant SET Ctype='int' WHERE Cname IN
> ('GL_ATI_blend_equation_separate', 'GL_ARB_multitexture');

okay, I'll try this, thanks!

> In bazaar, these constants are now marked as 'string' and devchk tries
> to check them in incorrect way (i.e. segfault here also indicates that
> devchk found a difference between system headers and the LSB ones:)
> but maybe we need more graceful report...).

ummm, a little more graceful, perhaps...

but how can we test this?