Loïc Bartoletti
Loïc Bartoletti
> > @jgrocha this is a very good idea! I made several tests, not yet released, some months ago. QGIS WKT is not readable by some other software. So +1....
OK for this change @jgrocha But you have to fix tests for Oracle, at least, this [file](https://github.com/qgis/QGIS/blob/e688fb104eb3906f69a5051a3b0ce61f480eac43/tests/src/python/test_provider_oracle.py) and maybe other providers.
We're unsure about the intention behind including both 90% and 99% confidence intervals. Was there originally a plan to have an option between 90% and 99%? Your suggestion to use...
Done, sorry. I think I already merged this, months ago...
> As we do not have a `isValid` function, I think `isNull` should return false if there is any NaN value. Or create a distinct `isValid` function. I think, we...
Nice, thanks for this first part.
CI (latest, tests) failed as expected, until https://gitlab.com/sfcgal/SFCGAL/-/merge_requests/516 is not merged
@robe2 @strk as mentioned on IRC/Matrix months ago, I also tried to switch from gnu11 to c11 in [another branch](https://github.com/lbartoletti/postgis/tree/msvc_compat_2). It's OK for *BSD (FreeBSD and macOs tested) but not...
> There are quite a few implementations for this depending on the program. One option to consider could be to provide a simple version (easier to parse for the use...