Barend Gehrels
d25b47f627
[buffer][test] decrease one tolerance which was too high
2015-10-14 12:59:30 +02:00
Barend Gehrels
82b1290849
[buffer][test] increase tolerances for extreme cases to let them pass
...
in clang/clang release/gcc
2015-10-14 12:59:15 +02:00
Barend Gehrels
71143c1b68
[buffer][fix] Fix case where side calculation fails because of length
2015-10-14 12:14:38 +02:00
Adam Wulkiewicz
825815a418
[test][buffer] Add cases causing assertion failure.
2015-09-09 00:50:20 +02:00
Barend Gehrels
49add55205
[buffer] Fix assertion for cases with extreme coordinates.
...
In case of numerical errors, the started ring should be aborted to avoid doing
calculations with them.
2015-07-31 12:40:01 +02:00
Adam Wulkiewicz
2ab10d0e72
[test][buffer] Add test cases causing assertion failure.
2015-07-28 06:44:54 +02:00
Barend Gehrels
1e5c976e19
[buffer][test] update expectations for no-robustness option
2015-07-15 11:58:12 +02:00
Barend Gehrels
91ccb7e6e3
[buffer] re-enable support for no-robustness policy
...
Influence on polygon geometries is limited (countries also little influence)
2015-07-08 20:34:07 +02:00
Barend Gehrels
b1f208d29c
[test] buffer, update expectations and tolerances
...
(after changing rounding/rescaling)
2015-07-01 14:02:30 +02:00
barendgehrels
dd347254d1
[test] remove conditional from spike30 (round) because output is OK
2015-06-27 12:31:28 +02:00
barendgehrels
4025867a85
[test] pending changes/fixes in SVG generation
2015-06-27 12:30:56 +02:00
barendgehrels
c3f1eb9b06
[test] remove two conditionals because these cases now also succeed,
...
either after selecting IP differently, or after rescale/rounding (or both)
2015-06-27 12:12:21 +02:00
barendgehrels
ea940461d9
[test] move SVG creation upwards to avoid invalid SVG if buffered result is empty
2015-06-27 12:04:54 +02:00
barendgehrels
606cb236af
[buffer][test] update expected values after change to rounding in rescaling,
...
and increased the tolerance.
Largest difference was 1.3e-05% (it100 deflate, 9087m2) and results are checked visually
2015-06-27 12:04:10 +02:00
barendgehrels
f77f0652b1
[buffer] update expected values and tolerance for mikado3 case
2015-06-27 11:44:12 +02:00
Adam Wulkiewicz
5fa509783d
[buffer][test] Alter the expected result and tolerance for one MLs case.
...
Increase the tolerance two times. With MSVC and MinGW the results are so
different that the currently used tolerance is too small.
2015-06-25 18:51:33 +02:00
Adam Wulkiewicz
7f91c4a527
[buffer] Update expected test results, syncronize with the change in rescale policy (rounding).
2015-06-25 17:28:38 +02:00
barendgehrels
d40bed9381
[buffer] change testvalues and tolerance in mikado with large distance
...
cases to support for all occuring variants on various platforms (can differ
up to 53 square meter so an average is selected and 30 meter tolerance)
2015-06-24 17:09:02 +02:00
Barend Gehrels
90aaa664c4
[test] add commented section to conveniently create a csv from testresults
2015-06-24 16:29:18 +02:00
Barend Gehrels
688ea21d42
[buffer] fix countries buffer test after change to select IP closest to
...
endpoint or else IP using shortest segment. New values are close to old
values, checked visually and the same for MSVC/gcc/clang
2015-06-24 16:28:51 +02:00
Barend Gehrels
15b14b5d7d
[buffer][test] higher tolerance for mikado cases
2015-06-24 13:18:22 +02:00
Barend Gehrels
9914d7a116
[buffer][test] replace -999 with constant
2015-06-24 13:17:52 +02:00
barendgehrels
181c17be5c
[buffer] fix testcase for county1 which had empty output (but expected area)
2015-06-24 12:18:16 +02:00
barendgehrels
b5756e9e08
[buffer] test: don't use assign_inverse but assign just coordinates to avoid
...
infinite ranges. Check for emptyness as expected
2015-06-24 11:51:06 +02:00
barendgehrels
fbe3cb1c83
[buffer] unit test: don't calculate rescale policy on empty output to
...
avoid empty or invalid boxes and possible assertions
2015-06-24 11:06:23 +02:00
barendgehrels
97c368cdf2
Buffer unit test - small rewrite of access to folder location
2015-06-21 22:54:32 +02:00
barendgehrels
603a2a7c12
Adapt buffer unit tests to change in IP calculation
2015-06-21 22:52:17 +02:00
barendgehrels
452f092e8e
[buffer][test] use a rescale policy calculated over the output, instead
...
of over the input. This shows problems in the newly added test, and also
invalidates one other testcase.
Both problems (self-intersections) should be solved later
2015-06-13 10:32:00 +02:00
Adam Wulkiewicz
45061a202d
[test][buffer] Add a case for which not valid buffer is generated.
2015-06-11 17:53:14 +02:00
Adam Wulkiewicz
72eb1c7218
[test][buffer] Suppress msvc warning by using the correct type.
2015-06-05 14:58:27 +02:00
barendgehrels
2b1dfaddb6
[buffer] use the new is_empty
2015-06-03 14:50:37 +02:00
barendgehrels
68648da969
[test][buffer] Don't leave envelope uninitialized if input geometry is empty
...
(this is a result of new envelope calculation)
2015-06-03 11:54:48 +02:00
barendgehrels
3b0529d8d9
[test][buffer] Add new test testing boost::geometry::buffer
...
(the detailed tests do test buffer_inserter)
2015-06-03 11:54:03 +02:00
barendgehrels
11d174d60e
[buffer][test] re-comment buffer_countries
2015-05-29 19:59:54 +02:00
barendgehrels
b849991dcd
[buffer][test] renamed testcases
2015-05-29 19:58:42 +02:00
barendgehrels
525613b461
[buffer][test] renamed testfiles
2015-05-29 19:53:11 +02:00
barendgehrels
cdf014e35d
[buffer] propagate error information from strategy to buffer inserter
2015-05-27 10:56:56 +02:00
barendgehrels
ad088fcf02
[buffer] stop in case of numeric errors in side strategy
2015-05-24 11:22:45 +02:00
barendgehrels
62fa2ccc59
[buffer][test] fix commented section (distance_left)
2015-05-23 23:46:14 +02:00
barendgehrels
1eacb2c096
[buffer][test] fix alternate box setting
2015-05-23 23:45:48 +02:00
barendgehrels
5370e2494d
[buffer][math] check for inf when checking on equals
...
This fixes an error in buffer for geometries with extreme coordinate differences
(though the generated buffer is not correct then)
2015-05-23 23:45:24 +02:00
Adam Wulkiewicz
bec216923f
[test][buffer] Fix compilation error - BOOST_MESSAGE unknown.
2015-05-15 13:25:52 +02:00
Barend Gehrels
d881d5811d
[buffer][test] add comment
2015-05-13 13:16:40 +02:00
Barend Gehrels
6158064410
[buffer] performance optimization: set the center point for buffers around
...
(multi)points, calculate min/max radius
and (when checking turn w.r.t. piece) check turn distance to center,
in most cases leave early
Including unit test. For multipoint_b_500k this increases performance with
a factor 6 (340 seconds -> 55 seconds)
Note, this is done for buffers around points but could be done for other
geometry types too (using e.g. centroid)
2015-05-13 13:08:22 +02:00
Barend Gehrels
61d3ae6b78
[buffer] performance improvement for point-pieces:
...
first check if it is in range
Including adaption of unit tests
Performance goes from 24 seconds (with side_of_intersection)
or 14 seconds (without) to 10 seconds (with side_of_intersection)
2015-05-13 10:58:07 +02:00
Barend Gehrels
ff1f2ab05c
[buffer][test] by default turn off labels for within turns
2015-05-05 16:43:23 +02:00
Barend Gehrels
d7a25934c3
[buffer][test] exclude case q2 for new approach.
...
As mailed to Adam: "But that is not related to the new method
- it is only by chance that it now fails and earlier succeeded.
This is a known problem which I still need to solve:
if two IP's are exactly on top of each other,
where a u/u turn is also involved, it can fail
(depending on sort-order, so depending on configuration)."
2015-05-05 16:42:57 +02:00
Barend Gehrels
c5ef4f912d
[buffer][fix] fix errors in multi_point cases
...
by using side_of_intersection in point_piece
2015-05-05 16:01:08 +02:00
Barend Gehrels
01d324288b
[buffer][test] add option to buffer per turn
...
also some other related changes in svg creation
2015-05-05 15:59:31 +02:00
Barend Gehrels
73394e99eb
[buffer][test] use round points for multi_points (not relevant besides
...
creating the rescale policy and svgs)
2015-05-05 15:57:44 +02:00
