Improve fallback handling for CamelCase fonts

From early Canary trials we observe unfulfilled requests
for fonts named in CamelCase without spaces which do not
resolve using the Skia mechanisms. e.g. `TrebuchetMS`
can resolve as `Trebuchet MS`. We attempt to find these
fonts by applying a regex to the face name that inserts
a space between lowercase and uppercase characters.

Additionally, we observe that, like ArialBlack, some
requests for ArialNarrow specify a bad weight. Handling
for ArialNarrow and common variants is added.

This CL only affects PDFium running in Chrome if the
WinPdfUseFontProxy feature is enabled.

Tests: pdf_unittests.exe --gtest_filter=PDFiumFontWinTest.*
Bug: 344643689
Change-Id: Icca13fae5ff17d6abe303e983da8795da6a33c24
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5660721
Reviewed-by: Lei Zhang <thestig@chromium.org>
Commit-Queue: Alex Gough <ajgo@chromium.org>
Reviewed-by: Rick Byers <rbyers@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1321146}

pdf/BUILD.gn

@@ -200,6 +200,7 @@ if (enable_pdf) {
         "pdfium/pdfium_font_win.cc",
         "pdfium/pdfium_font_win.h",
       ]
+      deps += [ "//third_party/re2" ]
     }
 
     if (is_chromeos) {

pdf/pdfium/DEPS

@@ -2,6 +2,7 @@ include_rules = [
   "+components/services/font/public/cpp",
   "+skia/ext/font_utils.h",
   "+third_party/pdfium/public",
+  "+third_party/re2/src/re2/re2.h",
   "+ui/accessibility",
   "+ui/gfx/codec",
 ]

pdf/pdfium/pdfium_font_win.cc

@@ -26,6 +26,7 @@
 #include "skia/ext/font_utils.h"
 #include "third_party/blink/public/platform/web_font_description.h"
 #include "third_party/pdfium/public/fpdf_sysfontinfo.h"
+#include "third_party/re2/src/re2/re2.h"
 #include "third_party/skia/include/core/SkFontMgr.h"
 #include "third_party/skia/include/core/SkFontStyle.h"
 #include "third_party/skia/include/core/SkStream.h"
@@ -295,13 +296,25 @@ class SkiaFontMapper {
                                 const SkFontStyle& style) {
     // Some fonts are specified with weights that Skia can't provide.
     // pdf.js/tests/issue5801.pdf specifies ArialBlack but a weight of 390.
-    if (face == "ArialBlack" || face == "Arial Black") {
-      SkFontStyle force_black = SkFontStyle(SkFontStyle::Weight::kBlack_Weight,
-                                            style.width(), style.slant());
-      return manager_->matchFamilyStyle("Arial", force_black);
+    // Commonly seen patterns: `ArialBlack` `Arial Black` & `Arial-Black`.
+    if (base::StartsWith(face, "Arial")) {
+      if (base::EndsWith(face, "Black")) {
+        SkFontStyle black = SkFontStyle(SkFontStyle::Weight::kBlack_Weight,
+                                        style.width(), style.slant());
+        return manager_->matchFamilyStyle("Arial", black);
+      }
+      if (base::EndsWith(face, "Narrow")) {
+        SkFontStyle narrow = SkFontStyle(SkFontStyle::Weight::kThin_Weight,
+                                         style.width(), style.slant());
+        return manager_->matchFamilyStyle("Arial", narrow);
+      }
     }
-    if (face == "ComicSansMS") {
-      return manager_->matchFamilyStyle("Comic Sans MS", style);
+    // Some fonts are specified without spaces in their name e.g. `ComicSansMS`.
+    std::string with_spaces(face);
+    // s/{lower case letter}{uppercase letter}/l u/g.
+    if (re2::RE2::GlobalReplace(&with_spaces, "(\\p{Ll})(\\p{Lu})", "\\1 \\2") >
+        0) {
+      return manager_->matchFamilyStyle(with_spaces.c_str(), style);
     }
     return nullptr;
   }

pdf/pdfium/pdfium_font_win_unittest.cc

@@ -143,7 +143,7 @@ TEST_F(PDFiumFontWinTest, FallbackFontsHangeul) {
   DeleteFont(id);
 }
 
-TEST_F(PDFiumFontWinTest, FinalFixups) {
+TEST_F(PDFiumFontWinTest, FinalFixupsArialBlack) {
   // ArialBlack with a low weight needs to be forced to weight kBlack.
   FontId id = MapFont(390, false, FXFONT_DEFAULT_CHARSET, FXFONT_FF_ROMAN,
                       "ArialBlack");
@@ -151,4 +151,20 @@ TEST_F(PDFiumFontWinTest, FinalFixups) {
   DeleteFont(id);
 }
 
+TEST_F(PDFiumFontWinTest, FinalFixupsComicSansMS) {
+  // ComicSansMS -> Comic Sans MS.
+  FontId id = MapFont(FXFONT_FW_NORMAL, false, FXFONT_DEFAULT_CHARSET,
+                      FXFONT_FF_ROMAN, "ComicSansMS");
+  EXPECT_TRUE(id);
+  DeleteFont(id);
+}
+
+TEST_F(PDFiumFontWinTest, FinalFixupsTrebuchetMS) {
+  // TrebuchetMS -> Trebuchet MS.
+  FontId id = MapFont(FXFONT_FW_NORMAL, false, FXFONT_DEFAULT_CHARSET,
+                      FXFONT_FF_ROMAN, "TrebuchetMS");
+  EXPECT_TRUE(id);
+  DeleteFont(id);
+}
+
 }  // namespace chrome_pdf