Better fix for [2478607]: colours leaking information about trap status.

There is still a bug in that 'x'ing over a former trap will describe the
feature as 'floor' rather than 'trap'. This is because _describe_feature()
looks at the actual grid, not the envmap. However, there is currently no
perfect way to prevent this: map_cell only stores an appearance, not the
actual feature layer. So if a monster was on a trap when you last saw
it, and then the trap gets disarmed, there is no way to know that there
was once a trap there. The correct fix is to have map_cell remember the
object, feature and monster layers separately, but since this will break
savefiles and require some work, I'm not fixing it yet.


git-svn-id: https://crawl-ref.svn.sourceforge.net/svnroot/crawl-ref/trunk@9020 c06c8d41-db1a-0410-9941-cceddc491573

1 files changed, 7 insertions, 1 deletions

diff --git a/crawl-ref/source/view.cc b/crawl-ref/source/view.cc
index 765509a497..dd380b4107 100644
--- a/crawl-ref/source/view.cc
+++ b/crawl-ref/source/view.cc
@@ -689,7 +689,13 @@ screen_buffer_t colour_code_map( const coord_def& p, bool item_colour,
         return (LIGHTGREEN);
 #endif
 
-    const dungeon_feature_type grid_value = grd(p);
+    dungeon_feature_type grid_value = grd(p);
+    if (!see_grid(p))
+    {
+        const int remembered = get_envmap_obj(p);
+        if (remembered < NUM_REAL_FEATURES)
+            grid_value = static_cast<dungeon_feature_type>(remembered);
+    }
 
     unsigned tc = travel_colour ? _get_travel_colour(p) : DARKGREY;