On 07/09/2016 06:53 AM, clipka wrote:
> Am 09.07.2016 um 00:46 schrieb William F Pokorny:
>
>> I found this from 2002:
>> http://news.povray.org/povray.beta-test/thread/%3C3c44167a@news.povray.org%3E/
>> which looks like there was intent to remove the option from 3.5.
>
> For the sake of backward compatibility, removing the option is not an
> option (no pun intended).
>
:-) Ah yes, think you are right.

>> If 0 is the only valid option for type at present, but others are
>> planned, should we warn in the parser when it is set to something other
>> than 0? Perhaps also comment the conditional in the code on type to save
>> a few cyles until other options exist ?
>
> A warning might be in order, depending on whether we label the setting
> as obsolete (in which case we should warn) or intended for future
> extensions (in which case we probably shouldn't warn, except maybe if
> the user specifies a value other than the supported value of 0).
>
> As for the conditional, if we go for the "future extension"
> interpretation I suggest replacing it with a switch() and making it the
> default branch. This way we keep the logic there, but make it easy for
> the compiler to optimize it away. (If we do that, we should add a
> warning to the parser if a value other than 0 is specified, as the
> behaviour will change.)
>

OK. I'll work toward the "compatibility and future extension" switch() 
change and warn on non-zero setting.

>> No matter the answer(s) supposing the documentation should be updated to
>> match.
>
> True.
>

I'll first create a branch for the change and submit a pull request for 
review. After refinement/adoption, I'll work up the documentation change 
to match.

Thanks,
Bill P.

Post a reply to this message