Open MauriceHubain opened 3 months ago
This is indeed also wrong and now is wrong for both multi and poly. Thanks for the report. It is not yet fixed by my concept fix for #1293 and #1294 and #1295
It is not fixed by my concept fix for #1226 and #1326
This is a follow up ticket for https://github.com/boostorg/geometry/issues/1288.
Adapting your test code from that ticket with a new data set, we get wrong results for double and long double precision:
Output: Type: float a: 4.26700010347 b: 0.296400005227 diff: 3.97060009825 ints: 0.296400005227 a-b: 3.97060009825 d+i 4.26700010347 validity diff: true ints: true Type: double a: 4.26700000517 b: 0.2964 diff: 0 ints: 0.2964 a-b: 3.97060000517 d+i 0.2964 validity diff: true ints: true Type: long double a: 4.26700000517 b: 0.2964 diff: 0 ints: 0.2964 a-b: 3.97060000517 d+i 0.2964 validity diff: true ints: true
Type: float a: 4.26700010347 b: 0.296400005227 diff: 3.97060009825 ints: 0.296400005227 a-b: 3.97060009825 d+i 4.26700010347 validity diff: true ints: true Type: double a: 4.26700000517 b: 0.2964 diff: 0 ints: 0.2964 a-b: 3.97060000517 d+i 0.2964 validity diff: true ints: true Type: long double a: 4.26700000517 b: 0.2964 diff: 0 ints: 0.2964 a-b: 3.97060000517 d+i 0.2964 validity diff: true ints: true