xgettext: Assume that Python source files are in UTF-8 by default.
Reported by ilias iliadis <apiuano-bugs@yahoo.gr>
at <https://savannah.gnu.org/bugs/?55231>.
* gettext-tools/src/xg-encoding.h (xgettext_global_source_encoding): Allow a
NULL value.
* gettext-tools/src/xg-encoding.c (xgettext_global_source_encoding): Likewise.
* gettext-tools/src/xgettext.c (main): Initialize
xgettext_global_source_encoding with NULL, not "ASCII". Handle a NULL
xgettext_global_source_encoding.
(extract_from_file): Use ASCII as default for xgettext_global_source_encoding.
* gettext-tools/src/x-javascript.c (extract_javascript): Likewise.
* gettext-tools/src/x-python.c (phase2_getc): Signal an error when encountering
an invalid or incomplete UTF-8 character.
(set_current_file_source_encoding): In the error message, use
xgettext_current_file_source_encoding instead of
xgettext_global_source_encoding.
(extract_python): Use UTF-8 as default for xgettext_global_source_encoding.
* gettext-tools/tests/xgettext-python-3: Verify that if the source file has no
magic coding comment but is UTF-8 encoded, xgettext succeeds.
* NEWS: Mention the change.
