Relative URLs should be handled first, only as a fallback we should use giturlparse. It won't throw and error, for relative paths so the try/catch was never used. Also the parsed attributes names changed.
Usually we want to iterative over all submodules. In case they are hosted somewhere else, we would still like to access all available properties like submodule path or commit id. I built in a fallback mechanism which could be checked in client code via subproject.project is None.
Relative URLs should be handled first, only as a fallback we should use
giturlparse
. It won't throw and error, for relative paths so thetry
/catch
was never used. Also the parsed attributes names changed.Usually we want to iterative over all submodules. In case they are hosted somewhere else, we would still like to access all available properties like submodule path or commit id. I built in a fallback mechanism which could be checked in client code via
subproject.project is None
.