git/list[1] front-page[2] threads[3] people[4] search[5] about
 

[PATCH 4/8] xdl_change_compact(): do one final shift or the other, not both

From
Michael Haggerty <mhagger@alum.mit.edu>
Date
Aug 3, 2016, 22:00 UTC
Message-ID
<39a135da93834fd72ee923d95d0cebfe525dfe7a.1470259583.git.mhagger@alum.mit.edu>
In-Reply-To
<cover.1470259583.git.mhagger@alum.mit.edu>

There is no need to shift the group to match a diff in the other file if we're just going to override that shift based on the compaction heuristic. Note that this changes the behavior if the matching shift would have shifted the group higher than the last blank line: the old code would have ignored the compaction heuristic in that case, whereas the new code always gives precedence to the compaction heuristic when it is turned on.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
---
 xdiff/xdiffi.c | 44 ++++++++++++++++++++++----------------------
 1 file changed, 22 insertions(+), 22 deletions(-)
diff --git a/xdiff/xdiffi.c b/xdiff/xdiffi.c
index 0f235bc..c67cfe3 100644
--- a/xdiff/xdiffi.c
+++ b/xdiff/xdiffi.c
@@ -547,11 +547,28 @@ int xdl_change_compact(xdfile_t *xdf, xdfile_t *xdfo, long flags) {
 			}
 		} while (groupsize != end - start);
 
-		/*
-		 * Try to move back the possibly merged group of changes, to match
-		 * the recorded position in the other file.
-		 */
-		if (end_matching_other != -1) {
+		if ((flags & XDF_COMPACTION_HEURISTIC) && blank_lines) {
+			/*
+			 * Compaction heuristic: if a group can be moved back and
+			 * forth, then if possible shift the group to make its
+			 * bottom line a blank line.
+			 *
+			 * As we already shifted the group forward as far as
+			 * possible in the earlier loop, we only need to handle
+			 * backward shifts, not forward ones.
+			 */
+			while (start > 0 &&
+			       !is_blank_line(recs, end - 1, flags) &&
+			       recs_match(recs, start - 1, end - 1, flags)) {
+				rchg[--start] = 1;
+				rchg[--end] = 0;
+			}
+		} else if (end_matching_other != -1) {
+			/*
+			 * Move the possibly merged group of changes back to line
+			 * up with the last group of changes from the other file
+			 * that it can align with.
+			 */
 			while (end_matching_other < end) {
 				rchg[--start] = 1;
 				rchg[--end] = 0;
@@ -561,23 +578,6 @@ int xdl_change_compact(xdfile_t *xdf, xdfile_t *xdfo, long flags) {
 					io--;
 			}
 		}
-
-		/*
-		 * If a group can be moved back and forth, see if there is a
-		 * blank line in the moving space. If there is a blank line,
-		 * make sure the last blank line is the end of the group.
-		 *
-		 * As we already shifted the group forward as far as possible
-		 * in the earlier loop, we need to shift it back only if at all.
-		 */
-		if ((flags & XDF_COMPACTION_HEURISTIC) && blank_lines) {
-			while (start > 0 &&
-			       !is_blank_line(recs, end - 1, flags) &&
-			       recs_match(recs, start - 1, end - 1, flags)) {
-				rchg[--start] = 1;
-				rchg[--end] = 0;
-			}
-		}
 	}
 
 	return 0;
-- 
2.8.1
Previous: Michael HaggertyNext: Michael Haggerty
Message 21 of 57 in “Better heuristics make prettier diffs”
  1. 0/8 Better heuristics make prettier diffsMichael Haggerty, Aug 3, 2016
  2. 2/8 xdl_change_compact(): clarify codeMichael Haggerty, Aug 3, 2016
  3. Stefan BellerAug 3, 2016
  4. Michael HaggertyAug 3, 2016
  5. Stefan BellerAug 3, 2016
  6. Jeff KingAug 4, 2016
  7. Michael HaggertyAug 10, 2016
  8. Stefan BellerAug 10, 2016
  9. 5/8 xdl_change_compact(): fix compaction heuristic to adjust ioMichael Haggerty, Aug 3, 2016
  10. Jeff KingAug 4, 2016
  11. Michael HaggertyAug 10, 2016
  12. Michael HaggertyAug 10, 2016
  13. Jeff KingAug 11, 2016
  14. Junio C HamanoAug 4, 2016
  15. Michael HaggertyAug 10, 2016
  16. 7/8 is_blank_line: take a single xrecord_t as argumentMichael Haggerty, Aug 3, 2016
  17. Junio C HamanoAug 4, 2016
  18. 6/8 xdl_change_compact(): keep track of the earliest endMichael Haggerty, Aug 3, 2016
  19. Junio C HamanoAug 4, 2016
  20. Michael HaggertyAug 10, 2016
  21. 4/8 xdl_change_compact(): do one final shift or the other, not bothMichael Haggerty, Aug 3, 2016
  22. 1/8 xdl_change_compact(): rename some local variables for clarityMichael Haggerty, Aug 3, 2016
  23. Jeff KingAug 4, 2016
  24. Junio C HamanoAug 4, 2016
  25. Michael HaggertyAug 13, 2016
  26. Jeff KingAug 14, 2016
  27. Michael HaggertyAug 3, 2016
  28. 3/8 xdl_change_compact(): rename i to endMichael Haggerty, Aug 3, 2016
  29. Jeff KingAug 4, 2016
  30. 8/8 diff: improve positioning of add/delete blocks in diffsMichael Haggerty, Aug 3, 2016
  31. Jacob KellerAug 3, 2016
  32. Michael HaggertyAug 3, 2016
  33. Jacob KellerAug 4, 2016
  34. Junio C HamanoAug 4, 2016
  35. Michael HaggertyAug 10, 2016
  36. Junio C HamanoAug 10, 2016
  37. Stefan BellerAug 3, 2016
  38. Michael HaggertyAug 3, 2016
  39. Stefan BellerAug 3, 2016
  40. Michael HaggertyAug 3, 2016
  41. Stefan BellerAug 4, 2016
  42. Michael HaggertyAug 10, 2016
  43. Jeff KingAug 4, 2016
  44. Stefan BellerAug 4, 2016
  45. Junio C HamanoAug 4, 2016
  46. Michael HaggertyAug 13, 2016
  47. Michael HaggertyAug 12, 2016
  48. Jeff KingAug 13, 2016
  49. Junio C HamanoAug 13, 2016
  50. Jacob KellerAug 14, 2016
  51. Stefan BellerAug 15, 2016
  52. Junio C HamanoAug 15, 2016
  53. Junio C HamanoAug 4, 2016
  54. Michael HaggertyAug 13, 2016
  55. Jeff KingAug 4, 2016
  56. Junio C HamanoAug 4, 2016
  57. Jeff KingAug 4, 2016

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.