Closed evandjpl closed 1 month ago
Ok looks like there should have been a lower bound on GDAL.jl /GDAL_jll to make sure we get the version where GDT_Int8
exists.
This may be annoying to fix, it needs a version yank but they're hard to get merged. If you really need this I suggest tracking down when Int8 support was added to GDAL.jl, fixing the versions to that and making a PR.
But it might easier to drop 1.8 instead.
Instead of a yank I'd suggest just fixing the compat in the registry, and then also here. That isn't hard to get merged, but needs a manual merge request in the JuliaLang Slack in #pkg-registration. Had to do it recently for GDAL as well: https://github.com/JuliaRegistries/General/pull/105696
https://github.com/JuliaRegistries/General/?tab=readme-ov-file#who-can-approve-an-early-merge https://pkgdocs.julialang.org/v1/registries/#Registry-Compat.toml
Why do you think this is Julia 1.8 related?
It could be 1.8 related if some other dep lower bounds to 1.9 and forces older versions.
(Yes, compat bounds are probably the thing to change, although theres a lot of disagreement, e.g. the registry readme says to yank)
Anyway, main part is finding the GDAL/GDAL_jll version to pin
So, tracking it down I found that GDAL 3.7 is where support for GDT_Int8 was added. https://gdal.org/development/rfc/rfc87_signed_int8.html https://github.com/OSGeo/gdal/issues/4002 And this is the commit that added support for GDAL 3.7 in GDAL.jl:
Same error here.
someone just needs to fix the GDAL versions in the registry for ArchGDAL
https://github.com/JuliaRegistries/General/pull/107486
Also note we are now likely to have conflicts with ImageMagick due to libgeotiff_jll deps.
Thats likely why running update
is not removing this problem - mostly we cant install the latest GDAL. Not having ImageMagick in your environment should help.
Should be fixed now https://github.com/JuliaRegistries/General/pull/107490 is merged.
My CI pipeline hit an error today while trying to precompile Julia libraries. Let me know if you need more info than this:
This is on Julia
v1.8
.