Added full output test to test-scan.

[nmh] / test / mhshow / test-charset

diff --git a/test/mhshow/test-charset b/test/mhshow/test-charset
index cb16f79893ed89ddbc80c5e8f83eca7f8136e3e7..c3f0687f7025e0e7581280ea168f0655d3ed969e 100755 (executable)
--- a/test/mhshow/test-charset
+++ b/test/mhshow/test-charset
@@ -20,7 +20,17 @@ check_exit '-eq 1' mhshow -
 
 require_locale en_US.UTF-8 en_US.UTF8 en_US.utf-8 en_US.utf8
 
 
 require_locale en_US.UTF-8 en_US.UTF8 en_US.utf-8 en_US.utf8
 

-test "$ICONV_ENABLED" -eq 0  &&  text_size=10  || text_size=11
+iconv_elides_question_marks=0
+if test "$ICONV_ENABLED" -eq 0; then
+    text_size=10
+else
+    text_size=11
+    #### The GNU iconv library normalises charset names by eliding '?', along
+    #### with some other characters.  The iconv libraries used on FreeBSD/NetBSD
+    #### and Mac OS X don't.
+    printf x | iconv -f '?UTF-8' -t UTF-8 >/dev/null 2>&1  &&
+        iconv_elides_question_marks=1
+fi

 
 expected=$MH_TEST_DIR/$$.expected
 actual=$MH_TEST_DIR/$$.actual
 
 expected=$MH_TEST_DIR/$$.expected
 actual=$MH_TEST_DIR/$$.actual


@@ -71,18 +81,30 @@ run_prog mhshow $msgnum > $actual 2>&1
 check "$expected" "$actual" : locale profile component
 
 
 check "$expected" "$actual" : locale profile component
 
 

+#### All of the remaining tests rely on iconv.a
+if test "$ICONV_ENABLED" -eq 0; then
+  finish_test
+  exit $failed
+fi
+
+
+#### RFC 2231 defines the extended value syntax:
+#### extended-initial-value := [charset] "'" [language] "'"
+####                           extended-other-values
+#### used below, where language is blank.
+

 start_test 'Encoded parameter value'
 msgfile=`mhpath new`
 msgnum=`basename $msgfile`
 start_test 'Encoded parameter value'
 msgfile=`mhpath new`
 msgnum=`basename $msgfile`

-#### The encoded charset name is really UTF-8 encoded,
-#### but say it's UTF-16 to exercise the iconv code,
-#### if it's available, in get_param_value().
+#### Use a charset to exercise the iconv code, if it's available,
+#### in get_param_value().  The Content-Type charset is UTF-8, encoded
+#### to exercise iconv.

 cat > $msgfile <<EOF
 From: foo@example.edu
 To: bar@example.edu
 Subject: test
 MIME-Version: 1.0
 cat > $msgfile <<EOF
 From: foo@example.edu
 To: bar@example.edu
 Subject: test
 MIME-Version: 1.0

-Content-Type: text/plain; charset*=UTF-16''%55%54%46%2d%38
+Content-Type: text/plain; charset*=ISO-2022-KR''%1b%24%29%43%55%54%46%2d%38

 Content-Transfer-Encoding: quoted-printable
 Content-Description: bogus quoted-printable data
 Date: Sun, 18 Dec 2005 00:52:39 +0100
 Content-Transfer-Encoding: quoted-printable
 Content-Description: bogus quoted-printable data
 Date: Sun, 18 Dec 2005 00:52:39 +0100


@@ -108,9 +130,11 @@ run_prog mhshow $msgnum > $actual 2>&1
 check "$expected" "$actual" : encoded parameter value
 
 
 check "$expected" "$actual" : encoded parameter value
 
 

+#### mhshow replaces the invalid character in the charset parameter value with
+#### a '?'.  So, a charset name of "?us-ascii" is fed to iconv.  If
+#### $iconv_elides_question_marks, iconv will normalize that to "us-ascii".
+#### Otherwise, we expect iconv to fail.

 start_test 'replacement character in parameter value'
 start_test 'replacement character in parameter value'

-#### The output of this test doesn't show it, but it covers the
-#### noiconv: portion of get_param_value().

 cat > $msgfile <<'EOF'
 Subject: invalid parameter value charset
 MIME-Version: 1.0
 cat > $msgfile <<'EOF'
 Subject: invalid parameter value charset
 MIME-Version: 1.0


@@ -127,7 +151,15 @@ MIME-Version: 1.0
 EOF
 
 run_prog mhshow $msgnum > $actual 2>&1
 EOF
 
 run_prog mhshow $msgnum > $actual 2>&1

-check "$expected" "$actual"
+if [ $iconv_elides_question_marks -eq 1 ]; then
+    check "$expected" "$actual"
+else
+    if grep "Can't convert .*us-ascii to UTF-8" "$actual" >/dev/null; then
+        rm -f "$expected" "$actual"
+    else
+        failed=1
+    fi
+fi

 
 
 finish_test
 
 
 finish_test