Merge pull request #6325 from Silvertree2010/fix/h15445-graph-csv-parser-does-not-strip

Strip leading byte-order mark in graph CSV parser

crates/engine/src/cmd/graph/data.rs

@@ -43,6 +43,7 @@ fn cell(c: &str) -> Cell {
 }
 
 pub(super) fn parse_csv(csv: &str) -> (Vec<String>, Vec<String>, Vec<Vec<Option<f64>>>) {
+    let csv = csv.strip_prefix('\u{feff}').unwrap_or(csv);
     // Records end at line breaks outside quotes; a record is tab-separated if it has a tab outside quotes.
     let rows: Vec<Vec<Cell>> = split_unquoted(csv, '\n')
         .into_iter()

crates/engine/src/cmd/graph/tests.rs

@@ -4205,3 +4205,8 @@ fn character_formats_are_kept_and_bigger_labels_space_out() {
         }
     }
 }
+
+#[test]
+fn csv_leading_byte_order_mark_is_ignored() {
+    assert_eq!(super::parse_csv("\u{feff}1,2\n3,4"), super::parse_csv("1,2\n3,4"));
+}