From c4f8495f11c52a6e44a8b13cd17d410283242772 Mon Sep 17 00:00:00 2001
From: Even Rouault <even.rouault@spatialys.com>
Date: Tue, 13 Dec 2022 18:49:48 +0100
Subject: [PATCH] GPKG: Add heuristics to try to detect corrupted RTree
 generated by GDAL 3.6.0, and disable use of the rtree

---
 autotest/ogr/ogr_gpkg.py                      | 39 +++++++++++++++++++
 .../gpkg/ogrgeopackagetablelayer.cpp          | 38 ++++++++++++++++++
 2 files changed, 77 insertions(+)

diff --git a/autotest/ogr/ogr_gpkg.py b/autotest/ogr/ogr_gpkg.py
index 9b9b339f0195..b3fbff88c5ce 100755
--- a/autotest/ogr/ogr_gpkg.py
+++ b/autotest/ogr/ogr_gpkg.py
@@ -7331,6 +7331,45 @@ def test_ogr_gpkg_background_rtree_build(filename):
     gdal.Unlink(filename)
 
 
+###############################################################################
+
+
+def test_ogr_gpkg_detect_broken_rtree_gdal_3_6_0():
+
+    filename = "/vsimem/test_ogr_gpkg_detect_broken_rtree_gdal_3_6_0.gpkg"
+
+    ds = gdaltest.gpkg_dr.CreateDataSource(filename)
+    lyr = ds.CreateLayer("foo")
+    for i in range(100):
+        f = ogr.Feature(lyr.GetLayerDefn())
+        f.SetGeometryDirectly(
+            ogr.CreateGeometryFromWkt("POINT(%d %d)" % (i % 10, i // 10))
+        )
+        lyr.CreateFeature(f)
+    ds = None
+
+    # Voluntary corrupt the RTree by removing the entry for the last feature
+    ds = ogr.Open(filename, update=1)
+    sql_lyr = ds.ExecuteSQL("DELETE FROM rtree_foo_geom WHERE id = 100")
+    ds.ReleaseResultSet(sql_lyr)
+    ds = None
+
+    with gdaltest.config_option("OGR_GPKG_THRESHOLD_DETECT_BROKEN_RTREE", "100"):
+        ds = ogr.Open(filename)
+        lyr = ds.GetLayer(0)
+        with gdaltest.error_handler():
+            gdal.ErrorReset()
+            lyr.SetSpatialFilterRect(8.5, 8.5, 9.5, 9.5)
+            assert (
+                "Spatial index (perhaps created with GDAL 3.6.0) of table foo is corrupted"
+                in gdal.GetLastErrorMsg()
+            )
+        assert lyr.GetFeatureCount() == 1
+        ds = None
+
+    gdal.Unlink(filename)
+
+
 ###############################################################################
 # Test ST_Area()
 
diff --git a/ogr/ogrsf_frmts/gpkg/ogrgeopackagetablelayer.cpp b/ogr/ogrsf_frmts/gpkg/ogrgeopackagetablelayer.cpp
index b23b98fb01c8..107bbb4ab88b 100644
--- a/ogr/ogrsf_frmts/gpkg/ogrgeopackagetablelayer.cpp
+++ b/ogr/ogrsf_frmts/gpkg/ogrgeopackagetablelayer.cpp
@@ -4170,6 +4170,44 @@ bool OGRGeoPackageTableLayer::HasSpatialIndex()
         m_osFIDForRTree = m_pszFidColumn;
     }
 
+    // Add heuristics to try to detect corrupted RTree generated by GDAL 3.6.0
+    // Cf https://github.com/OSGeo/gdal/pull/6911
+    if( m_bHasSpatialIndex )
+    {
+        const auto nFC = GetTotalFeatureCount();
+        if( nFC >= atoi(CPLGetConfigOption(
+                    "OGR_GPKG_THRESHOLD_DETECT_BROKEN_RTREE", "100000")) )
+        {
+            CPLString osSQL = "SELECT 1 FROM \"";
+            osSQL += SQLEscapeName(pszT);
+            osSQL += "\" WHERE \"";
+            osSQL += SQLEscapeName(GetFIDColumn());
+            osSQL += "\" = ";
+            osSQL += CPLSPrintf(CPL_FRMT_GIB, nFC);
+            osSQL += " AND \"";
+            osSQL += SQLEscapeName(pszC);
+            osSQL += "\" IS NOT NULL AND NOT ST_IsEmpty(\"";
+            osSQL += SQLEscapeName(pszC);
+            osSQL += "\")";
+            if( SQLGetInteger(m_poDS->GetDB(), osSQL, nullptr) == 1 )
+            {
+                osSQL = "SELECT 1 FROM \"";
+                osSQL += SQLEscapeName(m_osRTreeName);
+                osSQL += "\" WHERE id = ";
+                osSQL += CPLSPrintf(CPL_FRMT_GIB, nFC);
+                if( SQLGetInteger(m_poDS->GetDB(), osSQL, nullptr) == 0 )
+                {
+                    CPLError(CE_Warning, CPLE_AppDefined,
+                             "Spatial index (perhaps created with GDAL 3.6.0) "
+                             "of table %s is corrupted. Disabling its use. "
+                             "This file should be recreated or its spatial "
+                             "index recreated", m_pszTableName);
+                    m_bHasSpatialIndex = false;
+                }
+            }
+        }
+    }
+
     return CPL_TO_BOOL(m_bHasSpatialIndex);
 }