[14:42] <melodie_> hello
[22:23] <Rarrikins> How do I report a bug in the part of the system that handles #! lines at the top of scripts?
[22:24] <jtaylor> that would be the kernel
[22:24] <jtaylor> but are you sure you found a bug?
[22:24] <jtaylor> thats pretty old and established code
[22:24] <jtaylor> anything odd is more likely a feature :)
[22:24] <Rarrikins> Yes, if you have a script with DOS line endings, it gives a completely wrong error message.
[22:25] <jtaylor> what kind of a message?
[22:25] <Rarrikins> ": No such file or directory"
[22:25] <jtaylor> thats the expected message
[22:26] <jtaylor> if anything goes wrong with the shebang you get that, including when it can'T find a linker etc
[22:26] <jtaylor> its a very hard fix
[22:26] <jtaylor> maybe even impossible
[22:28] <Rarrikins> Couldn't it be something like "Couldn't load script interpreter" or something?
[22:28] <Rarrikins> That's pretty uninformative as is.
[22:28] <jtaylor> no the kernel gives you back a ENOENT#
[22:28] <jtaylor> you would have to change the kernel interface and add a new error code
[22:28] <jtaylor> that can#t really be done
[22:29] <Rarrikins> You could change the error message for ENOENT.
[22:29] <jtaylor> then it would be wrong for all real "file not found" errors
[22:30] <Rarrikins> It could be made suitably generic. I think the specificity is the problem.
[22:32] <jtaylor> see here for some discussion: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=609882
[22:32] <ubot2> Debian bug 609882 in command-not-found "libc6: "No such file or directory" error when attempting to execute LSB executable without lsb-core" [Wishlist,Open]
[22:33] <Rarrikins> OK, thanks
[22:33] <jtaylor> its not exactly the same, but I imagine the reason it is hard to fix is the same
[22:34] <jtaylor> see also here: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=697299#44
[22:34] <ubot2> Debian bug 697299 in command-not-found "misleading error message when ELF interpreter does not exist" [Wishlist,Open]