Dan Baston
Dan Baston
I'd lean towards option 3.
It's not that single reference, it's the entire validity checking process which is 2D-only.
We already have a mechanism for requesting/checking interrupts, so I think a progress callback should handle progress reporting only. I'm not sure about adding new C API signatures with a...
It's just going to take some time to review (1) does this provide a reasonable progress estimate for different datasets / parameters, and (2) is this the signature we want...
So, running the following: ``` gdal vector simplify-coverage ~/data/statscan/lcsd000b21a_e.shp /tmp/simp.gpkg --tolerance 1e-3 --overwrite --progress ``` Hangs at zero for 2:05, and then goes to 100% in 11 seconds. (subdivisions shapefile...
> is this the signature we want to use? Here I'm referring not to the signature of the callback itself, but whether we want functions to accept a progress callback...
> what are you thinking to specifically ? I don't have an alternate solution, but I'm hesitant to hurry and merge this without seeing how the same approach works in...
`GEOSUnaryUnion` would top my list for "single invocation on a huge object."
@rouault OK if I push https://github.com/dbaston/libgeos/commit/def9816fa0f0113f268b13fdffeab3cc4097b54e, to keep the call sites a bit cleaner?
c3d5a9ed4248b2e3c0774224a24740b034514932 reduces the null checks and simplifies the creation of subprogress callbacks a bit.