Closed gkreitz closed 3 weeks ago
Should we mention this in the spec? Or is it obvious that symlinks must point to something and that it mustn't point outside of the package?
It's not obvious that symlinks are allowed, so we may want to specify that. If we do, it certainly doesn't hurt to say that they must point to within the problem package.
It's not obvious that symlinks are allowed
That's a very good point. (And the rest follows).
Well do we want to allow symlinks to point to absolute paths? That seems not so great me
On Fri, May 24, 2024, 19:34 Fredrik Niemelä @.***> wrote:
It's not obvious that symlinks are allowed
That's a very good point. (And the rest follows).
— Reply to this email directly, view it on GitHub https://github.com/Kattis/problemtools/pull/261#issuecomment-2130238033, or unsubscribe https://github.com/notifications/unsubscribe-auth/AB2ZBKUENQQAI5HVFEPNQG3ZD6I6BAVCNFSM6AAAAABIIDKQK6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCMZQGIZTQMBTGM . You are receiving this because you are subscribed to this thread.Message ID: @.***>
Well do we want to allow symlinks to point to absolute paths? That seems not so great me
Excellent point. Absolute symlinks, even if they point to the right place on the machine we're currently running on feel like a very bad idea in a problem package. I added another check for absolute symlinks.
We recently saw issues caused by dangling symlinks being present in a problem package. This PR makes it an error to have any symlinks anywhere in the problem package point to something which does not exist, or point somewhere outside of the problem package.
Example output: