Add support for leading BOM when reading import CSV files

Author: stealthycoin

.changes/next-release/bugfix-configure-38726.json

@@ -0,0 +1,5 @@
+{
+  "category": "configure",
+  "type": "enhancement",
+  "description": "Add support for importing CSV credential files that have leading UTF-8 BOM. Fixes `#7721 <https://github.com/aws/aws-cli/issues/7721>`__."
+}

awscli/customizations/configure/importer.py

@@ -12,8 +12,8 @@
 # language governing permissions and limitations under the License.
 import os
 import sys
+import codecs
 
-from awscli.compat import compat_open
 from awscli.customizations.utils import uni_print
 from awscli.customizations.commands import BasicCommand
 from awscli.customizations.configure.writer import ConfigFileWriter
@@ -115,6 +115,9 @@ def __init__(self, strict=True):
         self.strict = strict
 
     def _format_header(self, header):
+        # Remove leading UTF BOM character if present
+        if header.startswith(codecs.BOM_UTF8.decode()):
+            header = header[1:]
         return header.lower().strip()
 
     def _parse_csv_headers(self, header):

tests/unit/customizations/configure/test_importer.py

@@ -109,6 +109,13 @@ def test_csv_parser_simple(self):
         )
         self.assert_parse_matches_expected(contents)
 
+    def test_csv_parser_with_bom(self):
+        contents = (
+            u'\ufeffUser name,Access key ID,Secret access key\n'
+            'PROFILENAME,AKID,SAK\n'
+        )
+        self.assert_parse_matches_expected(contents)
+
     def test_csv_parser_multiple_entries(self):
         contents = (
             'User name,Access key ID,Secret access key\n'
@@ -121,6 +128,18 @@ def test_csv_parser_multiple_entries(self):
         ]
         self.assert_parse_matches_expected(contents)
 
+    def test_csv_parser_multiple_entries_bom(self):
+        contents = (
+            u'\ufeffUser name,Access key ID,Secret access key\n'
+            'PROFILENAME1,AKID1,SAK1\n'
+            'PROFILENAME2,AKID2,SAK2\n'
+        )
+        self.expected_credentials = [
+            ('PROFILENAME1', 'AKID1', 'SAK1'),
+            ('PROFILENAME2', 'AKID2', 'SAK2'),
+        ]
+        self.assert_parse_matches_expected(contents)
+
     def test_csv_parser_multiple_entries_with_bad_entries(self):
         self.parser.strict = False
         contents = (