Implement `copyFromBuffer`, `copyFromArrays`, `copyToBuffer` and `copyToArrays`

[kylebarron]

I'm very inexperienced at C/rust ffi, so happy to get some feedback before adding all four functions and tests.

Closes https://github.com/georust/geos/issues/117

[kylebarron]

Is this crate concerned at all with m values? If they aren't supported anywhere else, should I remove them from these functions?

[kylebarron]

I added a couple more doctests. I think the new_from_buffer function signature should be simplified still, especially because it currently lets you pass in size: 3 but has_z: false and has_m: false. If we assume we never have m values then that makes this a little simpler.

The other thing to note is in get_as_buffer/copyToBuffer. It appears that GEOS always stores coordinates internally as 3 dimensions, so copyToBuffer will fill in NaN values for z coords. Should the API here handle repacking into a 2d buffer? For my case that's unnecessary because I'll be repacking again into arrow memory anyways

[GuillaumeGomez]

I added a couple more doctests. I think the new_from_buffer function signature should be simplified still, especially because it currently lets you pass in size: 3 but has_z: false and has_m: false. If we assume we never have m values then that makes this a little simpler.

size is independent from the m as far as I can see: https://github.com/libgeos/geos/blob/0a55b3af552470754893485b51407cc0219dbaae/capi/geos_ts_c.cpp#L2435-L2443

[kylebarron]

size is independent from the m as far as I can see: libgeos/geos@0a55b3a/capi/geos_ts_c.cpp#L2435-L2443

That's so interesting; I should've known to look at the source! So passing in a buffer of XY is not a straight memcpy

[kylebarron]

I added a couple more doctests. I think the new_from_buffer function signature should be simplified still, especially because it currently lets you pass in size: 3 but has_z: false and has_m: false. If we assume we never have m values then that makes this a little simpler.

size is independent from the m as far as I can see: libgeos/geos@0a55b3a/capi/geos_ts_c.cpp#L2435-L2443

Oops I meant to write dims, not size. size is the number of coordinates in the array; dims is the number of dimensions; we should be able to infer dims from has_z and has_m

[GuillaumeGomez]

Time for the last part of the PR: fixing CI errors. :)

Don't worry about errors from valgrind though. Also, don't forget to rebase on master.

[kylebarron]

I updated with better docstrings, a bug fixed in as_buffer, better handling of m values.

[GuillaumeGomez]

Clippy is still unhappy.

[kylebarron]

I put a feature flag on the type def. Now I'm getting a pass locally with cargo clippy -- -D warnings on Clippy 0.1.67

[GuillaumeGomez]

Clippy is still failing for the same reason.

[kylebarron]

Clippy is still failing for the same reason.

It is? The Clippy output is different in my last two commits. The first time it had a warning related to my changes; now all the clippy warnings are unrelated to my changes. I can change the lifetime syntax there if you'd like

[GuillaumeGomez]

Oh sorry, didn't pay attention. I opened https://github.com/georust/geos/pull/122 to fix it. In the meantime, please squash your commits.

[kylebarron]

I guess this is why on my own repos I usually let github squash my commits for me, because I'm inept at git 😂

[kylebarron]

I guess there's no way for me to re-open this PR, even if I push a new commit to the branch now; sorry about that! I'll have to open a new PR

[GuillaumeGomez]

I wrote a tuto a long time ago about it here. But the other easy way with small PRs is to simply remove all changes and then do them again and commit cleanly. Sometimes, the "not so great way" works well enough.

[GuillaumeGomez]

Nah, reopening in progress. :p

[kylebarron]

My issue was I followed the second suggestion in https://stackoverflow.com/a/25357146 and ran

git reset --soft $(git merge-base master HEAD)

but then ran git push -f, forgetting to commit again 😆

Nah, reopening in progress. :p

thanks! didn't know you could do that

[GuillaumeGomez]

thanks! didn't know you could do that

Me neither until we ended up in this situation. ;)

[GuillaumeGomez]

Thanks!

[kylebarron]

thanks for your help!