The name of the C locale w/ UTF-8 support is not always C.UTF-8.

Use a pattern to find it (if present) and use that value instead
of hard-coding C.UTF-8.  This works around a leak sanitizer crash
on certain inputs.

plugins/python/Makefile.in

@@ -36,7 +36,7 @@ cross_compiling = @CROSS_COMPILING@
 # Compiler & tools to use
 CC = @CC@
 LIBTOOL = @LIBTOOL@
-GREP = @GREP@
+EGREP = @EGREP@
 SED = @SED@
 AWK = @AWK@
 
@@ -228,11 +228,9 @@ check-fuzzer:
 
 check: $(TEST_PROGS) check-fuzzer
 	@if test X"$(cross_compiling)" != X"yes"; then \
-	    if locale -a 2>&1 | $(GREP) '^C.UTF-8$$' >/dev/null 2>&1; then \
-		LC_ALL=C.UTF-8; export LC_ALL; \
-	    else \
-		LC_ALL=C; export LC_ALL; \
-	    fi; \
+	    l=`locale -a 2>&1 | $(EGREP) -i '^C.UTF-?8$$'`; \
+	    test -n "$$l" || l="C"; \
+	    LC_ALL="$$l"; export LC_ALL; \
 	    unset LANG || LANG=; \
 	    unset LANGUAGE || LANGUAGE=; \
 	    MALLOC_OPTIONS=S; export MALLOC_OPTIONS; \