You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm trying to run sf and rgeos revdep checks with forthcoming GEOS 3.9.0, which presents itself as "3.9.0dev". This fails in old_sf_geos():
Error: works with self-intersections: sf (@test-fix_geo_problems.R#94)
Error: works with other problems: sf (@test-fix_geo_problems.R#106)
These fork on GEOS < 3.8.0 for good reason, but do not try to strip trailing string extensions to the declared version number. This only matters if you care about protecting yourselves from the radical changes coming in GEOS 3.9.0. Your code presupposes that the version number is normalised, which it will be when 3.9.0 is released, but then it will be too late to modify code ahead of time. The rgeos tests earlier in the same file all pass, so I think the irritant is just expecting the sf GEOS version string to be normalised.
The text was updated successfully, but these errors were encountered:
@rsbivand this should be fixed now, the non-numeric version components are now converted to numeric before the comparison to v 3.8.0. Thanks for running the revdep checks and catching this!
I'm trying to run sf and rgeos revdep checks with forthcoming GEOS 3.9.0, which presents itself as
"3.9.0dev"
. This fails inold_sf_geos()
:These fork on GEOS < 3.8.0 for good reason, but do not try to strip trailing string extensions to the declared version number. This only matters if you care about protecting yourselves from the radical changes coming in GEOS 3.9.0. Your code presupposes that the version number is normalised, which it will be when 3.9.0 is released, but then it will be too late to modify code ahead of time. The rgeos tests earlier in the same file all pass, so I think the irritant is just expecting the sf GEOS version string to be normalised.
The text was updated successfully, but these errors were encountered: