testprogs: Ignore escape characters when printing test name

Long story: This was triggered by the addition of the test_trust_ntlm.sh
script in commits 3caca9b and 2de1994. test_trust_ntlm.sh creates a
variable CREDS="$REALM\\$USERNAME%$PASSWORD" that is then used as part
of the test name. subunit.sh uses echo to print the name that is then
picked up by subunithelper.py. test_trust_ntlm.sh also uses /bin/sh as
shell which can be a POSIX compliant shell like dash.

This combination broke 'make test' for any username starting with the
letter c. In this case CREDS contains the escape sequence \c that is
defined to stop producing further output at this point. dash implements
this feature and the echo in subunit.sh as a result skips the output
after \c, including skipping the newline. This means that the data
received by subunithelper.py contains the timestamp from the next line
in the test name, which then breaks the testcase tracking.

Fix this by replacing the echo in subunit.sh with a printf that does not
trigger the special handling of escape characters.

Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>

Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Fri May  5 23:44:16 CEST 2017 on sn-devel-144

1 files changed, 5 insertions, 5 deletions

diff --git a/testprogs/blackbox/subunit.sh b/testprogs/blackbox/subunit.sh
index 5c81ce20a11..aca53f72536 100755
--- a/testprogs/blackbox/subunit.sh
+++ b/testprogs/blackbox/subunit.sh
@@ -28,14 +28,14 @@ timestamp() {
 subunit_start_test () {
   # emit the current protocol start-marker for test $1
   timestamp
-  echo "test: $1"
+  printf 'test: %s\n' "$1"
 }
 
 
 subunit_pass_test () {
   # emit the current protocol test passed marker for test $1
   timestamp
-  echo "success: $1"
+  printf 'success: %s\n' "$1"
 }
 
 # This is just a hack as we have some broken scripts
@@ -48,7 +48,7 @@ subunit_fail_test () {
   # we use stdin because the failure message can be arbitrarily long, and this
   # makes it convenient to write in scripts (using <<END syntax.
   timestamp
-  echo "failure: $1 ["
+  printf 'failure: %s [\n' "$1"
   cat -
   echo "]"
 }
@@ -60,7 +60,7 @@ subunit_error_test () {
   # we use stdin because the failure message can be arbitrarily long, and this
   # makes it convenient to write in scripts (using <<END syntax.
   timestamp
-  echo "error: $1 ["
+  printf 'error: %s [\n' "$1"
   cat -
   echo "]"
 }
@@ -70,7 +70,7 @@ subunit_skip_test () {
   # the error text.
   # we use stdin because the failure message can be arbitrarily long, and this
   # makes it convenient to write in scripts (using <<END syntax.
-  echo "skip: $1 ["
+  printf 'skip: %s [\n' "$1"
   cat -
   echo "]"
 }