When "git fetch" fails because the remote unexpectedly rewound
its head and fast-forward check triggers, we issued a warning
but kept going anyway. This proposed patch makes the command
exit with non-zero status.
I think this is a sensible change and makes it easier to use
from scripts, but it might have other issues. For example when
you are tracking more than one heads from the remote, and the
first one fast-forwards but the second one doesn't, it updates
the first one and then stops. If we happen to process the
rewound one first, neither is updated because we stop at the
first one. I think this particular discrepancy probably is not
worth worrying about, but there may be other more serious
fallouts we need to fix if we did this.
Comments?
---
diff --git a/git-fetch.sh b/git-fetch.sh
index 0346d4a..6835634 100755
--- a/git-fetch.sh
+++ b/git-fetch.sh
@@ -179,6 +179,7 @@ fast_forward_local () {
;;
*)
echo >&2 " not updating."
+ exit 1
;;
esac
}
-
: 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