Antranig Basman antranig.basman at
Sat May 14 00:26:11 UTC 2011

On 13/05/2011 12:31, Yura Z wrote:
> Hi Antranig, I've got another patch with a failing test case for you.
> Yura

Thanks JURA... this one is not "clearly" a bug - in that resolution behaviour between renderer decorator 
options and demanded options was never clear. In particular, demands options generally have the status of 
"user options" so the situation is as if the user supplied some direct arguments to some component which was 
then also hit by a demands block... which is of course impossible, since the user would have had to call the 
function directly and not via IoC :P So this behaviour is currently undefined. As you observe, the demands 
block material wins, and the decorator material is discarded.

I responded on -
there is supplied a configuration using "mergeAllOptions" which lets you achieve the effect you were aiming 
for (passing the test) - can you describe the situation and why you need to split the configuration in this 
way? We can perhaps improve this area so you can issue "{options}" rather than "{arguments.1}" but I think 
this situation will always require the use of "mergeAllOptions" since it is inherently ambiguous.



More information about the fluid-work mailing list