Ensure that a hunk-header never appears in the middle of an Extraneous range.

A hunk header is always Extraneous.  However Extraneous text before it is
quite separate and is best kept separate.
Start a new Extraneous merger with each hunk header will make it easier
to detect hunk headers and handle them better.

Signed-off-by: NeilBrown <neilb@suse.de>

diff --git a/merge2.c b/merge2.c
index 73b7347299721bed65600ddc186b3cba3f35d702..83be6f16aec9c01c2fc80b9606a76e2099272e39 100644 (file)
--- a/merge2.c
+++ b/merge2.c
@@ -102,6 +102,9 @@ static int isolate_conflicts(struct file af, struct file bf, struct file cf,
         * or any line that has both Changed and Unmatched content.
         * (Unmatched content where nothing is changed is common and not
         *  really a 'wiggle').
+        *
+        * A hunk header is never considered part of a conflict.  It
+        * thereby can serve as a separator between conflicts.
         */
        int i, j, k;
        int cnt = 0;
@@ -276,23 +279,28 @@ struct ci make_merger(struct file af, struct file bf, struct file cf,
                                assert(b < csl1[c1].b);
                                /* some Extraneous text */
                                /* length is min of unmatched on left
-                                * and matched on right
+                                * and matched on right.
+                                * However a hunk-header is not allowed
+                                * in the middle of an extraneous section,
+                                * only at the start.
                                 */
                                rv.merger[i].type = Extraneous;
                                rv.merger[i].al = 0;
-                               rv.merger[i].cl =
-                                       rv.merger[i].bl =
+                               newb = b +
                                        min(csl1[c1].b - b,
                                            csl2[c2].len - (b-csl2[c2].a));
-                               newb = b+rv.merger[i].bl;
                                for (j = b; j < newb; j++) {
                                        if (bf.list[j].start[0] == '\0') {
                                                if (wiggle_found > 1)
                                                        rv.wiggles++;
                                                wiggle_found = 0;
+                                               if (j > b)
+                                                       newb = j;
                                        } else
                                                wiggle_found++;
                                }
+                               rv.merger[i].cl =
+                                       rv.merger[i].bl = newb - b;
                        }
                } else if (match1 && !match2) {
                        /* some changed text
@@ -346,6 +354,7 @@ struct ci make_merger(struct file af, struct file bf, struct file cf,
        }
        rv.merger[i].type = End;
        rv.merger[i].in_conflict = 0;
+       assert(i < l);
        rv.conflicts = isolate_conflicts(af, bf, cf, csl1, csl2, words,
                                         rv.merger, show_wiggles);
        if (wiggle_found)