[Merge M63] Revert cleanup of IsHyphen and reimplement
The original version of this code was landed in
https://pdfium-review.googlesource.com/c/pdfium/+/13690/. A corner
case has been found that breaks.
In this CL I have reverted the changes in IsHyphen and implemented a
less aggressive cleanup that I have tested works as expected.
TBR=dsinclair@chromium.org
BUG=chromium:781804
Change-Id: I3b36f420834081fdd9e1ae17efc234b561b4df41
Reviewed-on: https://pdfium-review.googlesource.com/17950
Commit-Queue: Ryan Harrison <rharrison@chromium.org>
Reviewed-by: dsinclair <dsinclair@chromium.org>
Reviewed-on: https://pdfium-review.googlesource.com/18330
Reviewed-by: Ryan Harrison <rharrison@chromium.org>
diff --git a/core/fpdftext/cpdf_textpage.cpp b/core/fpdftext/cpdf_textpage.cpp
index 82778b1..0a7a39f 100644
--- a/core/fpdftext/cpdf_textpage.cpp
+++ b/core/fpdftext/cpdf_textpage.cpp
@@ -1218,25 +1218,29 @@
}
bool CPDF_TextPage::IsHyphen(wchar_t curChar) const {
- WideStringView curText;
- if (!m_TempTextBuf.IsEmpty())
- curText = m_TempTextBuf.AsStringView();
- else if (!m_TextBuf.IsEmpty())
+ WideStringView curText = m_TempTextBuf.AsStringView();
+ if (curText.IsEmpty())
curText = m_TextBuf.AsStringView();
- else
+
+ if (curText.IsEmpty())
return false;
- curText = curText.TrimmedRight(0x20);
- if (curText.GetLength() < 2)
- return false;
+ auto iter = curText.rbegin();
+ for (; iter != curText.rend() && *iter == 0x20; iter++) {
+ // Do nothing
+ }
- // Extracting the last 2 characters, since they are all that matter
- curText = curText.Right(2);
- if (!IsHyphenCode(curText.Last()))
- return false;
-
- if (FXSYS_iswalpha(curText.First() && FXSYS_iswalnum(curChar)))
- return true;
+ if (iter != curText.rend()) {
+ if (!IsHyphenCode(*iter))
+ return false;
+ iter++;
+ if (FXSYS_iswalpha(*iter) && FXSYS_iswalpha(*iter))
+ return true;
+ } else {
+ iter--;
+ if (!IsHyphenCode(*iter))
+ return false;
+ }
const PAGECHAR_INFO* preInfo;
if (!m_TempCharList.empty())