When parsing certain pdf's, I'm encountering an IndexOutOfBoundsException in WindowsFontDirFinder.java on line 106:
WindowsFontDirFinder is used by the statement on line 128 in FontMapperImpl.java in my case:
The cause of the error is that the value of windir is "" (empty string), at least during debugging in Windows. However, I should mention that parsing of the same document also fails on ubuntu. While I don't know the exact mechanism of the failure on ubuntu, I assume it is related.
The code checks whether windir is null before doing the substring operation, but not whether the length is in fact >2. I'm not sure why windir ends up being "" here, as the windir env var is C:\WINDOWS.
A working fix for me was to change the windir null check on line 94 to: