Merge pull request #8 from aaronbell/main

Extending character set and correcting backslash

sources/build.py

@@ -19,7 +19,7 @@
 TEMP = Path('temp')
 EXPORT = Path('fonts/ttf')
 SRC_IMPORT = Path("sources/extensions")
-VERSION = "1.002"
+VERSION = "1.05"
 
 for font in IMPORT.glob("*.ttf"):
 
@@ -71,7 +71,6 @@
     # Step 2 - Some metadata changes to align with GF expectations
     # NAME table modifications
     sourceTTF["name"].removeNames(platformID=1)
-    sourceTTF["name"].removeNames(nameID=6,platformID=3, langID=1041)
 
     for platformID in [1033, 1041]:
 
@@ -99,6 +98,8 @@
             sourceTTF["name"].setName("This Font Software is licensed under the SIL Open Font License, Version 1.1. This license is available with a FAQ at: https://scripts.sil.org/OFL",13,3,1,platformID)
             sourceTTF["name"].setName("https://scripts.sil.org/OFL",14,3,1,platformID)
 
+    sourceTTF["name"].removeNames(nameID=6,platformID=3, langID=1041)
+
     # OS/2 Table modifications
     sourceTTF["OS/2"].fsType = 0
 
@@ -151,20 +152,6 @@
     newDSIG.signatureRecords = []
     finalVersion.tables["DSIG"] = newDSIG
 
-    # Irritatingly, pyftmerge does not calculate the average character width correctly. It appears to simply average the xAvgCharWidth value in the two font files. However, given the heavy weighting of wider Japanese characters, we need to recalculate the average and insert it back into the font. 
-
-    width_sum = 0
-    count = 0
-    for glyph_id in finalVersion['glyf'].glyphs:  # At least .notdef must be present.
-        width = finalVersion['hmtx'].metrics[glyph_id][0]
-        # The OpenType spec doesn't exclude negative widths, but only positive
-        # widths seems to be the assumption in the wild?
-        if width > 0:
-            count += 1
-            width_sum += width
-
-    avgCharWidth = int(round(width_sum / count))
-    finalVersion["OS/2"].xAvgCharWidth = avgCharWidth
 
     finalVersion.save(EXPORT / str(outputTTF).replace("BIZ-","BIZ"))