-
Notifications
You must be signed in to change notification settings - Fork 130
Comments
Theoretically, we could support the #countries[mapnik/geometry_type=3] {
polygon-fill: #fff;
} |
Y, like @kkaefer this is possible to do. I'm not sure if this syntax is the best as far as css-like-ness, could it be something like #countries[mapnik-geometry-type=3] {
polygon-fill: #fff;
} This would conflict with very uncommon fields called the same, but at the same time doesn't introduce a new syntax for this. |
mapnik-geometry-type sounds good. |
Cool, ping here when it's implemented mapnik-side then we'll start thinking about making this backwards/forwards compatible |
Okay. Makes me realize that Mapnik-reference should likely be responsible as well for declaring these constants. On Oct 9, 2012, at 5:47 AM, Tom MacWright notifications@github.com wrote:
|
Can I help in any way here ? I understand the mapnik side is ready ? |
@artemp, @tmcw, @springmeyer, what about the value type ? Should we aim to use symbolic names rather than (or even better in addition to) the numeric ones ? What about:
|
Forget the above, seems already supported by mapnik |
yeah, keywords I'm still ambivalent about if this really needs to be changed as quoting currently works fine:
|
for Future Rob's reference, @springmeyer says the following works as-is:
|
Personally I'd prefer the (just because I like the concept doesn't mean it's straightforward to implement though 😉) |
Not sure. What about
Is this valid or no? The concept doesn't map cleanly for me - geometry types are attributes of features and should be filtered on. CSS element selectors are for atoms which are contained in other things - a div inside of body and so on.
should be
Just like any other anonymous filter selector. |
Yup, the filter approach definitely works for geometry type. Memory dredge:
Need to dig up Shawn/Josh's notes. |
I think this ticket can be closed, as |
yes quoting, or |
Mapnik 2.1 supports a new magic name to allow sorting on geometry type (mapnik/mapnik#546). A a keyword of
mapnik::geometry_type
is available in the expression engine as if that were an attribute on the data.Testing now I realize that the
::
conflicts with carto attachments so parsing fails.But it can be made to work by quoting:
So, this issue stands to resolve the questions: should we stick with the current name and live with needing quoting? Should it be renamed in mapnik? Should carto start being smart about not interpreting
::
inside selectors as attachments? Other ideas?/cc @artemp and @tmcw.
The text was updated successfully, but these errors were encountered: