[gentium-commits] [SCM] gentium updated: r62 - trunk/test/type1/Scripts

[Pavel Far??]

Author: pavel
Date: 2022-03-27 18:46:17 +0000 (Sun, 27 Mar 2022)
New Revision: 62

Modified:
   trunk/test/type1/Scripts/Makefile
   trunk/test/type1/Scripts/README
Log:
add command for showing diff of *.enc files for Type1 and TrueType

For reasons for this see Remarks in README.


Modified: trunk/test/type1/Scripts/Makefile
===================================================================
--- trunk/test/type1/Scripts/Makefile	2022-03-27 15:32:31 UTC (rev 61)
+++ trunk/test/type1/Scripts/Makefile	2022-03-27 18:46:17 UTC (rev 62)
@@ -1,4 +1,5 @@
 AFM_PATH = '../../../gentium/fonts/afm/public/gentium-tug'
+ENC_PATH = '../../../gentium/fonts/enc/dvips/gentium-tug'
 BUILD_PATH = '../../../gentium/source/fonts/gentium-tug/type1'
 
 # variables for target 'diff'
@@ -65,3 +66,10 @@
 split-kerns-clean:
 	# delete temporarty files created by split-kerns
 	rm *kerns*.afm
+
+diff-encodings:
+	# show differences between encodings for ttf end Type 1 fonts, suppress exit code
+	-cd $(ENC_PATH); for i in *-ttf*.enc; do \
+		echo '*' $$i $${i/-ttf/}; \
+		diff $$i $${i/-ttf/}; \
+	done

Modified: trunk/test/type1/Scripts/README
===================================================================
--- trunk/test/type1/Scripts/README	2022-03-27 15:32:31 UTC (rev 61)
+++ trunk/test/type1/Scripts/README	2022-03-27 18:46:17 UTC (rev 62)
@@ -59,3 +59,40 @@
 
 diff-clean
   Delete the files generated by 'make diff'.
+
+diff-encodings
+  Show differences of the encoding files for ttf fonts and Type 1 fonts (if
+  there are special ttf encodings). They use some characters under different
+  names and it has effect for splitting the extra kerning pairs into
+  individual files.
+
+
+Remarks
+=======
+
+I must look if the differences in the encoding files for Type 1 and TrueType
+fonts have some effect for metrics of Gentium fonts.
+
+I'm not sure now about some issues and I must look at them later.
+
+I think that some kerning pairs might be lost when generating the tfm files
+and using some name (such as dcroat) that is not in the used encoding. It
+might be checked by converting tfm to pl and looking there.
+
+I think that the encoding file is also used when generating e.g. the final pdf
+document. Is might result in some glyphs missing (but probably not bad
+metrics) if using either Type 1 or TrueType fonts.
+
+I remember that I generated metrics for TeX from one font and I used glyphs
+from completely different font (for some testing). I think that I used two
+different encoding files (in addition to other things). However it was very
+long time ago.
+
+It is also probable that these characters have different names in Type 1 and
+TrueType fonts. And if so, what are advantages and disadvantages and should
+this be changed?
+
+I found that the differences have some effect, because I compared the extra
+kerning pairs splitted to individual encoding files and the kerns for encoding
+T1 weren't the same for the old and the new fonts. Four kerning pairs with
+dcroat.sc were in T1 for the old fonts, but in 'rest' for the new fonts.