On Sun, Jul 25, 2010 at 09:50, Thomas Rast <trast@xxxxxxxxxxxxxxx> wrote: > git-sh-i18n was not picked up by the valgrind wrapper code because > that tests for the executable bit. Since there does not seem to be a > particular reason why it needs to be non-executable, and this is the > path of least special-casing to fix --valgrind testing, just make it > executable. > > Signed-off-by: Thomas Rast <trast@xxxxxxxxxxxxxxx> > --- > 0 files changed, 0 insertions(+), 0 deletions(-) > mode change 100644 => 100755 git-sh-i18n.sh > > diff --git a/git-sh-i18n.sh b/git-sh-i18n.sh > old mode 100644 > new mode 100755 > -- Why is it including git-sh-i18n.sh? It should only be using the generated SCRIPT_LIB git-sh-i18n file. On my system the executable bit of git-sh-i18n.sh has no effect on the generated git-sh-i18n file: $ rm git-sh-i18n; chmod +x git-sh-i18n.sh ; make ; ls -l git-sh-i18n* GEN git-sh-i18n SUBDIR git-gui SUBDIR gitk-git make[1]: Nothing to be done for `all'. SUBDIR perl SUBDIR git_remote_helpers SUBDIR templates -rw-r--r-- 1 avar avar 1965 2010-07-25 10:15 git-sh-i18n -rwxr-xr-x 1 avar avar 1955 2010-07-25 10:08 git-sh-i18n.sh And this still leaves git-sh-setup without an executable bit, why doesn't that fail under valgrind while git-sh-i18n apparently does? Maybe something to do with it being included as <<. "$GIT_EXEC_PATH"/git-sh-i18n>> instead of as <<. git-sh-setup>> ? -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html