Re: posix_spawn[p] less secure on Hurd than on Linux

2020-12-26 Thread Samuel Thibault
Hello, Bruno Haible, le jeu. 24 déc. 2020 02:54:28 +0100, a ecrit: > this misfeature was withdrawn from glibc in > > and > . > > But the test p

Re: posix_spawn[p] less secure on Hurd than on Linux

2020-12-24 Thread Joshua Branson
Thanks for reporting Bruno! Pointing out issues with GNU/Hurd's glibc is super helpful! -- Joshua Branson Sent from Emacs and Gnus https://gnucode.me https://video.hardlimit.com/accounts/joshua_branson/video-channels https://propernaming.org "You can have whatever you want, as long as

posix_spawn[p] less secure on Hurd than on Linux

2020-12-23 Thread Bruno Haible
Hi, In glibc versions < 2.15, posix_spawn and posix_spawnp did execute a script (with execution bits set) even if it did not start with a '#!' marker. Then, /bin/sh was used as an interpreter. You can imagine how easily a text file, which happens to set execution bits set because it was copied fr