Reserve space for result in ::FilterContents

This changes the implementation for the specific bug listed and
proactively fixes it in the other overrides. The general bug here is
that if you concat a WideString in a tight loop without first
reserving space, each call will cause an allocation size change and
memcpy. This is very expensive and causes ClusterFuzz cases to
timeout.

BUG=chromium:863295

Change-Id: I6c1d900a31b98cd9ddcf91d1ec0f3973c9cdfa26
Reviewed-on: https://pdfium-review.googlesource.com/38110
Reviewed-by: Tom Sepez <tsepez@chromium.org>
Reviewed-by: Henrique Nakashima <hnakashima@chromium.org>
Commit-Queue: Ryan Harrison <rharrison@chromium.org>

1 files changed, 1 insertions, 0 deletions

diff --git a/fxbarcode/oned/BC_OnedEAN13Writer.cpp b/fxbarcode/oned/BC_OnedEAN13Writer.cpp
index 6818d432d6..1facf9ee26 100644
--- a/fxbarcode/oned/BC_OnedEAN13Writer.cpp
+++ b/fxbarcode/oned/BC_OnedEAN13Writer.cpp
@@ -62,6 +62,7 @@ bool CBC_OnedEAN13Writer::CheckContentValidity(const WideStringView& contents) {
 
 WideString CBC_OnedEAN13Writer::FilterContents(const WideStringView& contents) {
   WideString filtercontents;
+  filtercontents.Reserve(contents.GetLength());
   wchar_t ch;
   for (size_t i = 0; i < contents.GetLength(); i++) {
     ch = contents[i];