Add //third_party/sqlite to dirs_to_snapshot, remove net_sql.patch

third_party/sqlite/src/test/ovfl.test

Index: third_party/sqlite/src/test/ovfl.test
diff --git a/third_party/sqlite/src/test/ovfl.test b/third_party/sqlite/src/test/ovfl.test
new file mode 100644
index 0000000000000000000000000000000000000000..075b1e43dd01e2c24518902d5e8784cf4cc0c5ba
--- /dev/null
+++ b/third_party/sqlite/src/test/ovfl.test
@@ -0,0 +1,49 @@
+# 2014 October 01
+#
+# The author disclaims copyright to this source code.  In place of
+# a legal notice, here is a blessing:
+#
+#    May you do good and not evil.
+#    May you find forgiveness for yourself and forgive others.
+#    May you share freely, never taking more than you give.
+#
+#***********************************************************************
+# This file implements regression tests for SQLite library.  The
+# focus of this file is testing the SQLITE_DIRECT_OVERFLOW_READ logic.
+#
+
+set testdir [file dirname $argv0]
+source $testdir/tester.tcl
+set testprefix ovfl
+
+# Populate table t2:
+#
+#   CREATE TABLE t1(c1 TEXT, c2 TEXT);
+#
+# with 2000 rows. In each row, c2 spans multiple overflow pages. The text
+# value of c1 ranges in size from 1 to 2000 bytes. The idea is to create
+# at least one row where the first byte of c2 is also the first byte of
+# an overflow page. This was at one point exposing an obscure bug in the
+# SQLITE_DIRECT_OVERFLOW_READ logic.
+#
+do_test 1.1 {
+  set c2 [string repeat abcdefghij 200]
+  execsql {
+    PRAGMA cache_size = 10;
+    CREATE TABLE t1(c1 TEXT, c2 TEXT);
+    BEGIN;
+  }
+  for {set i 1} {$i <= 2000} {incr i} {
+    set c1 [string repeat . $i]
+    execsql { INSERT INTO t1 VALUES($c1, $c2) }
+  }
+  execsql COMMIT
+} {}
+
+do_execsql_test 1.2 {
+  SELECT sum(length(c2)) FROM t1;
+} [expr 2000 * 2000]
+
+finish_test
+
+