Geographic buffer line #1045
Geographic buffer line #1045
Conversation
barendgehrels
force-pushed
the
geographic-buffer-line
branch
from
d8c3df8 to
a117766
Compare
| { | ||
| out << id++ << ";3;" << bg::wkt(p) << std::endl; | ||
| } | ||
| } |
There was a problem hiding this comment.
This small .csv part is already there, and looks like:
barendgehrels
force-pushed
the
geographic-buffer-line
branch
from
a117766 to
caa4000
Compare
| typename RangeOut | ||
| > | ||
| inline bool apply(Point const& ip, Point const& vertex, | ||
| Point const& perp1, Point const& perp2, |
There was a problem hiding this comment.
something seems wrong with indentation here?
There was a problem hiding this comment.
similar issues in apply() below
| // If angle is zero, shift angle a tiny bit to avoid spikes. | ||
| calculation_type const eps = angle == 0 ? 1.0e-10 : 0.0; | ||
| auto const dir_rad = direct_t::apply(lon_rad, lat_rad, | ||
| buffer_distance, angle + eps, |
There was a problem hiding this comment.
I think is better to indent below lon_rad or use one tab
| typename direct_t::result_type | ||
| dir_r = direct_t::apply(get_as_radian<0>(point), get_as_radian<1>(point), | ||
| buffer_distance, angle, | ||
| // If angle is zero, shift angle a tiny bit to avoid spikes. |
There was a problem hiding this comment.
is this some fix on buffer point geographic strategy to avoid spikes or is another issue introduced by the new buffer line strategies?
There was a problem hiding this comment.
This was already the case in earlier code. It's now less but not yet 100% disappeared.
We have to come back to it later. If I encounter it, I will create an issue.
| typename DistanceStrategy | ||
| > | ||
| inline result_code apply(Point const& input_p1, Point const& input_p2, | ||
| buffer_side_selector side, |
There was a problem hiding this comment.
should that be passed as const&?
There was a problem hiding this comment.
It is an enumerator. We normally don't do that.
| // For linestrings with flat ends, it's not necessarily the case, there may be points | ||
| // too close, especially on artefacts in heavily curved input with flat ends. | ||
| // Therefore the default expectation can be modified. Inspect the SVG visually before doing this. | ||
| std::size_t too_close = 0, too_far = 0, total = 0; |
There was a problem hiding this comment.
we use one declaration per line style, right?
There was a problem hiding this comment.
Indeed, thanks (and I applied also your other indentation comments)
barendgehrels
force-pushed
the
geographic-buffer-line
branch
3 times, most recently
from
a4a13a4 to
227c35f
Compare
Including tests and testcode
barendgehrels
force-pushed
the
geographic-buffer-line
branch
from
227c35f to
839a4ed
Compare
This is a fairly large PR adding two strategies for geographic buffers:
They are now tested and working for linestrings.
Next phase will be:
end_roundAfter that I will test it with multi linestrings and polygons.
It works pretty well in general, for example:
This picture is created with QGis and
.csv, a next PR will add (optional) possibility to add.csvin testcode. Earlier I often used.svgbut (especially for geographic),.csvor.geojsonis quite convenient.