summaryrefslogtreecommitdiffstats
path: root/chrome/browser/views/find_bar_win_browsertest.cc
diff options
context:
space:
mode:
Diffstat (limited to 'chrome/browser/views/find_bar_win_browsertest.cc')
-rw-r--r--chrome/browser/views/find_bar_win_browsertest.cc12
1 files changed, 8 insertions, 4 deletions
diff --git a/chrome/browser/views/find_bar_win_browsertest.cc b/chrome/browser/views/find_bar_win_browsertest.cc
index 22db8aa..3995293 100644
--- a/chrome/browser/views/find_bar_win_browsertest.cc
+++ b/chrome/browser/views/find_bar_win_browsertest.cc
@@ -347,11 +347,15 @@ IN_PROC_BROWSER_TEST_F(FindInPageControllerTest, FindCrash_Issue1341577) {
// This would crash the tab. These must be the first two find requests issued
// against the frame, otherwise an active frame pointer is set and it wont
// produce the crash.
+ // We used to check the return value and |ordinal|. With ICU 4.2, FiP does
+ // not find a stand-alone dependent vowel sign of Indic scripts. So, the
+ // exptected values are all 0. To make this test pass regardless of
+ // ICU version, we just call FiP and see if there's any crash.
+ // TODO(jungshik): According to a native Malayalam speaker, it's ok not
+ // to find U+0D4C. Still need to investigate further this issue.
int ordinal = 0;
- EXPECT_EQ(1, FindInPage(L"\u0D4C", FWD, IGNORE_CASE, &ordinal));
- EXPECT_EQ(1, ordinal);
- EXPECT_EQ(1, FindInPage(L"\u0D4C", FWD, IGNORE_CASE, &ordinal));
- EXPECT_EQ(1, ordinal);
+ FindInPage(L"\u0D4C", FWD, IGNORE_CASE, &ordinal);
+ FindInPage(L"\u0D4C", FWD, IGNORE_CASE, &ordinal);
// This should work fine.
EXPECT_EQ(1, FindInPage(L"\u0D24\u0D46", FWD, IGNORE_CASE, &ordinal));