* Reach feature parity with mercator view:
  * Resize tile atlas to reduce draw calls
  * Fill in tiles from other zoom layers

This enables the implementation of TileAtlas::textured_visible_tiles for
different kinds of visible tiles, e.g. for different layers.

and TileLayer -> MercatorTileLayer

This fixes a glutin/winit issue that resulted in 100% CPU usage because
of repeated X11 ConfigureWindow/ConfigureNotify events.

Still a partial solution. Uses region growing starting from the central
tile and checks if at least one of the vertices is visible.

* Split up map_view module into
  * map_view
  * mercator_view
  * orthografic_view

ViewMode::Flat -> Projection::Mercator
ViewMode::Globe -> Projection::Orthografic

Still not solved. Tried to approach by calculating properties of
elliptical arcs, but there are so many edge cases...

* Toggle between ViewMode::Flat and ViewMode::Globe with Ctrl+G
* Approach tile visibility problem at the sphere

This removes the holes of the globe at the poles.

A work in progress...

search::par_search_blocking: Remember node ids for matching ways and
return the positions of these nodes in a second pass over the file.

* Decode/search in parallel
* Start decoding the first Blob as soon as possible
* Update multiple markers at once

Use the parameters -s and --pbf for this.

* Cache the enabled vertex attributes in `Context`
* The vertex attribute parameters (offset, stride, ...) are not cached,
  yet
* Store parameters as `VertexAttribParams` in `Program`

This reduces calls to glUseBuffer.

This would reduce the calls to glUseProgram if there actually was more
than one program.

Each Texture is bound to a unique texture unit (glBindTexture is called
only once). Context keeps track of the current active texture unit and
reduces the number of glActiveTexture calls to a minimum.