1 files changed, 33 insertions, 0 deletions
diff --git a/layout/reftests/fonts/README b/layout/reftests/fonts/README
new file mode 100644
index 0000000000..fd2410568f
--- /dev/null
+++ b/layout/reftests/fonts/README
@@ -0,0 +1,33 @@
+
+Notes about fonts in this directory
+
+Ahem.ttf - ACID3 test font
+
+markXXX.ttf and markXXX.otf
+
+These fonts are autogenerated with FontForge using the Python script mark-generate.py.  
+See the comments in that file for more information on how to run the script.
+
+The markX.ttf and markX.otf fonts contain a single glyph for the X character consisting
+of three stacked boxes.  The mark2X.ttf and mark2X.otf files also contain just a glyph for the
+character X but the glyph is similar to the space mark character.  The markXmark2Y.ttf has
+two glyphs, the first glyph for X and the second glyph for Y.
+
+The markA.eot and markB.eot files are for EOT-related testing.  The markA.eot file is just 
+a copy of markA.ttf while markB.eot is a valid EOT file embedding the contents of markB.ttf
+with a null root string (so it can be used in IE without domain-specific restrictions).
+
+
+PositioningTest1.ttf, PositioningTest2.ttf
+
+These fonts were generated with FontForge and then OpenType tables added with MS VOLT.
+The glyphs are from Charis SIL (see http://scripts.sil.org/CharisSILFont), released under the
+Open Font License.
+
+In both fonts, the lowercase 'o' glyph should appear raised when rendered. Font 1 uses a GPOS
+positioning rule to achieve this, so it tests whether vertical deltas are being handled
+correctly. Font 2 uses a GSUB substitution to replace the 'o' with a glyph that has already
+been offset within the em-square, so no vertical delta is involved.
+
+Finally, the capital 'O' character maps directly to the shifted 'o' glyph, so this can be
+used for comparison without requiring any OpenType layout support.