commit 40111e6fed154aeab93ea26a3e0a691745d13419
author Ben Wagner <bungeman@chromium.org> Fri Aug 15 13:24:02 2025 -0700
committer Pdfium LUCI CQ <pdfium-scoped@luci-project-accounts.iam.gserviceaccount.com> Fri Aug 15 13:24:02 2025 -0700
tree aad7a8657f1036a6e4d6dff40242b70b56e5caea
parent 14391854578644993d7ca4dee6e4cddf7fda08bd

Change UseTTCharmapUnicode to prefer MSSymbol

Prior to "Treat all Unicode character maps like MSUnicode" [0] only the
cmap(3,1) (Windows,Unicode BMP) was considered as Unicode. After that
change, all Unicode cmaps like cmap(0,*) (Unicode,*) are considered as
Unicode. This fixes a number of issues with embedded fonts that do not
have a cmap(3,1).

However, this caused a regression in the case of embedded fonts with
Unicode cmap that isn't cmap(3,1) which also have a cmap(3,0)
(Windows,Symbol). In this case the Unicode cmap is now chosen over the
Symbol cmap. The Symbol cmap has special rules to augment the cmap as
well as special rules around converting to Unicode which will not be
applied to a Unicode cmap.

This changes the rule to prefer cmap(3,1) as before, then any other
Unicode cmap so long as there is no cmap(3,0). This restores the
previous behavior while still using Unicode cmaps when they are
available.

[0] https://pdfium.googlesource.com/pdfium/+/5da0e90faa40079145dc146f70bf186d3ad5e93f

Bug: 438810722
Change-Id: I247bb739fcb39b177923a7d4f3821df70dbca4fe
Reviewed-on: https://pdfium-review.googlesource.com/c/pdfium/+/135030
Reviewed-by: Lei Zhang <thestig@chromium.org>
Commit-Queue: Ben Wagner <bungeman@google.com>
Reviewed-by: Ben Wagner <bungeman@google.com>

core/fpdfapi/font/cpdf_font.cpp

diff --git a/core/fpdfapi/font/cpdf_font.cpp b/core/fpdfapi/font/cpdf_font.cpp
index 9a98f3b..bd60942 100644
--- a/core/fpdfapi/font/cpdf_font.cpp
+++ b/core/fpdfapi/font/cpdf_font.cpp
@@ -34,6 +34,7 @@
 #include "core/fxge/cfx_fontmapper.h"
 #include "core/fxge/cfx_substfont.h"
 #include "core/fxge/fx_font.h"
+#include "core/fxge/fx_fontencoding.h"
 
 namespace {
 
@@ -429,11 +430,29 @@
 }
 
 bool CPDF_Font::UseTTCharmapUnicode(const RetainPtr<CFX_Face>& face) {
+  size_t charmap_unicode_index = 0;
+  bool charmap_unicode_found = false;
+  bool charmap_mssymbol_found = false;
   for (size_t i = 0; i < face->GetCharMapCount(); i++) {
-    if (face->GetCharMapEncodingByIndex(i) == fxge::FontEncoding::kUnicode) {
+    const int platform_id = face->GetCharMapPlatformIdByIndex(i);
+    const int encoding_id = face->GetCharMapEncodingIdByIndex(i);
+    const fxge::FontEncoding encoding = face->GetCharMapEncodingByIndex(i);
+    if (platform_id == 3 && encoding_id == 1) {
       face->SetCharMapByIndex(i);
       return true;
     }
+    if (platform_id == 3 && encoding_id == 0) {
+      charmap_mssymbol_found = true;
+      continue;
+    }
+    if (!charmap_unicode_found && encoding == fxge::FontEncoding::kUnicode) {
+      charmap_unicode_found = true;
+      charmap_unicode_index = i;
+    }
+  }
+  if (charmap_unicode_found && !charmap_mssymbol_found) {
+    face->SetCharMapByIndex(charmap_unicode_index);
+    return true;
   }
   return false;
 }