[mkgmap-dev] Commit: r2353: A few tests for WanMil's style functions.

Wed Oct 31 20:59:36 GMT 2012

> On 31/10/12 15:55, WanMil wrote:
>>>
>>> Version 2353 was commited by steve on 2012-10-31 14:15:43 +0000 (Wed, 31 Oct 2012)
>>>
>>> A few tests for WanMil's style functions.
>>
>> Thanks for adding these tests.
>>
>> Do you think a "negative" test would also be good?
>
> Thanks, I can add that.
>
> I've been looking into making functions a kind of ValueOp so
> they can be compiled rather than looked up at run time, as I
> think you mentioned before.
>
> I can then convert normal tag match highway=primary to
> get_tag(highway)=primary behind the scenes, so that everything
> looks consistent. This also makes it possible to detect and
> give an error if you use length() by itself, without any tag
> matches.

Great! I tried to implement that but got stuck. It's easier for you to 
realize this because you invented the whole style reading and processing.

If possible it would be better to allow using length() on its own in a 
rule. I think sooner or later some style implementors will try rules like:
length() < 100 { set specialdistance=short }
length() > 100 & length() < 1000 { set specialdistance=medium }
length() > 1000 { set specialdistance=long }

Anyhow an error would be a good start!

>
> ..Steve

WanMil