Well, shouldn't this rather count the number of bytes when trimming?
Currently it assumes some average package name length, and might still fail. Also, it might crop although cropping isn't necessary.
To reproduce this easier, some fake packages with very long names might get used.
Well, shouldn't this rather count the number of bytes when trimming?
Currently it assumes some average package name length, and might still fail.
Also, it might crop although cropping isn't necessary.
To reproduce this easier, some fake packages with very long names might get used.