File name case sensitivity issues in the Windows system about nuitka HOT 8 CLOSED

from nuitka.

https://learn.microsoft.com/en-us/windows/wsl/case-sensitivity

from nuitka.

I removed one use of it the other day, where it checks for .ico suffix, simply because it is code duplication with hasFileExtension but I guess, that one does it too.

from nuitka.

Using os.path.normpath does not allow for case insensitive approaches, so Python doesn't support it. I only tried with 3.10, but that what matters. So it seems we need our own function. The below code illustrates the main use of it. A new function that takes your setting into account is definitely doable, lets see if we find something existing.

def areSamePaths(path1, path2):
    """Decide if two paths the same.

    Args:
        path1: First path
        path2: Second path

    Returns:
        Boolean value indicating if the two paths point to the
        same path.

    Notes:

        Case differences ignored on platforms where that is the
        norm, and with it normalized, and turned absolute paths, and
        even short paths, it then becomes a mere string compare after that.
    """

    path1 = os.path.abspath(os.path.normpath(path1))
    path2 = os.path.abspath(os.path.normpath(path2))

    if os.path.exists(path1) and os.path.exists(path2):
        path1 = getExternalUsePath(path1)
        path2 = getExternalUsePath(path2)

    path1 = os.path.normcase(path1)
    path2 = os.path.normcase(path2)

    return path1 == path2

from nuitka.

I have replaced uses of assignments from os.path.normcase that then can cause these filenames to be wrong, there were a few more things like that, and the code could be done better in all cases by using APIs that compare without modifying.

Generally it cannot be avoided. Comparisons cannot be made real, i.e. it's virtually impossible to detect if a folder has this attribute set in Python. The C API for that is very ugly, and even then, it's very unclear when it must be done or not, e.g. when searching for a module, when it's filename based and wrong in Python or not.

But I believe it might work better for you now.

from nuitka.

Thanks for your report, this is fixed on the factory branch, which is a development version under rapid development. You can try it out by going here: https://nuitka.net/doc/factory.html

Feedback if this is working is very welcome, just please do not share plans of doing it, but rather confirmations or denials of it working.

from nuitka.

These changes are now on develop, let me know if it works for you now.

from nuitka.

This is part of the stable release 2.0 that I just made.

from nuitka.