Check before free

[Smit-tay]

-- inode table array element can be null (ex: element zero intentionally null)

[matt335672]

I can't see the point of this check.

free(NULL) is perfectly valid C, and has been since at least K&R days. The reference from ISO/IEC 9899:1999 (which is the only one I have by my desk) is 7.20.3.2 which quotes for free(ptr):-

If ptr is a null pointer, no action occurs.

See also my comment in #1487

What am I missing? Are you getting a coredump which is pointing you in this area? Are you hitting the issue with 0.9.12 which is addressed by #1487?

Edit : Just read #1502. I think you'll find this is already addressed by #1487.

[Smit-tay]

Indeed.

Although because I am looking at a production server it's not so easy to debug.

The call stack led me to this quite simple function and my (naive) review caught what look(s/ed) like a potential problem.

I was unaware that free(0) is safe.

So, that implies that something that the array holds is no longer available.

For me debugging beyond this level isn't an option, not due to personal ability, but due to environmental problems. I really don't have the resources to investige in a meaningful way

[matt335672]

I might be able to do something with the stack trace you've provided in #1502 to work out whether this is indeed #1487 or something else.

Could you maybe close this PR, re-open #1502, and add some information about the server operating system and version of the XRDP package you've installed? I couldn't see that information.

Is falling back to v0.9.11 something you could do quite easily? I'm not familiar with the remmina client, but if it allows you to map a network drive, this might be a workaround you could use.