upstream: cannot effectively test posix-rename extension after

changes in feature advertisment.

OpenBSD-Regress-ID: 5e390bf88d379162aaa81b60ed86b34cb0c54d29

diff --git a/regress/sftp-perm.sh b/regress/sftp-perm.sh
index 304ca0ac58770d52628617027047831f70acde63..de96a14da8e835fe22035177eaa783ef2adc3483 100644 (file)
--- a/regress/sftp-perm.sh
+++ b/regress/sftp-perm.sh
@@ -1,4 +1,4 @@
-#      $OpenBSD: sftp-perm.sh,v 1.2 2013/10/17 22:00:18 djm Exp $
+#      $OpenBSD: sftp-perm.sh,v 1.3 2021/03/31 21:59:26 djm Exp $
 #      Placed in the Public Domain.
 
 tid="sftp permissions"
@@ -220,13 +220,15 @@ perm_test \
        "test ! -d ${COPY}.dd" \
        "test -d ${COPY}.dd"
 
-perm_test \
-       "posix-rename" \
-       "realpath,stat,lstat" \
-       "rename $COPY ${COPY}.1" \
-       "touch $COPY" \
-       "test -f ${COPY}.1 -a ! -f $COPY" \
-       "test -f $COPY -a ! -f ${COPY}.1"
+# Can't readily test this because the client falls back to traditional rename.
+# XXX maybe there is a behaviorial difference we can test for?
+#perm_test \
+#      "posix-rename" \
+#      "realpath,stat,lstat" \
+#      "rename $COPY ${COPY}.1" \
+#      "touch $COPY" \
+#      "test -f ${COPY}.1 -a ! -f $COPY" \
+#      "test -f $COPY -a ! -f ${COPY}.1"
 
 perm_test \
        "rename" \