"Maciej W. Rozycki" <ma...@imgtec.com> skribis: > On Fri, 30 Sep 2016, Hans-Peter Nilsson wrote: > >> > The patch below works around the Bash 4.2 bug described at >> > <http://ftp.gnu.org/gnu/bash/bash-4.2-patches/bash42-007>. >> >> > * emulparams/elf32bmipn32-defs.sh: Shift quote of >> > "x$EMULATION_NAME" to the left to work around >> > <http://ftp.gnu.org/gnu/bash/bash-4.2-patches/bash42-007>. >> > >> >> > >> > -case x"$EMULATION_NAME" in >> > +case "x$EMULATION_NAME" in >> > xelf32*n32*) ELFSIZE=32 ;;
[...] > Also for the purpose of such a check what does x"$EMULATION_NAME" expand > to in the buggy case given a specific value of $EMULATION_NAME? Here’s a sample: --8<---------------cut here---------------start------------->8--- $ echo $BASH_VERSION 4.2.0(1)-release $ unset EMULATION_NAME $ echo x$EMULATION_NAME x $ echo x"$EMULATION_NAME" x $ case x"$EMULATION_NAME" in x) echo good;; *) echo bad;; esac bad $ case "x$EMULATION_NAME" in x) echo good;; *) echo bad;; esac good $ echo x"$EMULATION_NAME"y xy $ echo "x$EMULATION_NAME"y xy --8<---------------cut here---------------end--------------->8--- IOW, the problem manifests in ‘case’ statements, but not in ‘echo’. I agree that rejecting such buggy shells at configure time would be safer. There’s been 1 year between Bash 4.2 and Bash 4.3, and 3 weeks between Bash 4.2 and the ‘bash42-007’ fix; not sure how widespread the buggy shell is. HTH, Ludo’.
