[truetype] Apply MVAR hasc, hdsc and hlgp metrics to current FT_Face metrics. Instead of setting typo or win metrics as the new FT_Face metrics indiscriminately, apply only typo deltas to the currently active FT_Face metrics. This prevents line height differences when e.g. the default outlines were used as the regular face and instances for everything else. * src/truetype/ttgxvar.c (tt_apply_mvar): Implement.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108
diff --git a/ChangeLog b/ChangeLog
index aac45a4..870972c 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,5 +1,16 @@
2019-02-02 Nikolaus Waxweiler <madigens@gmail.com>
+ [truetype] Apply MVAR hasc, hdsc and hlgp metrics to current FT_Face metrics.
+
+ Instead of setting typo or win metrics as the new FT_Face metrics
+ indiscriminately, apply only typo deltas to the currently active FT_Face
+ metrics. This prevents line height differences when e.g. the default
+ outlines were used as the regular face and instances for everything else.
+
+ * src/truetype/ttgxvar.c (tt_apply_mvar): Implement.
+
+2019-02-02 Nikolaus Waxweiler <madigens@gmail.com>
+
[sfnt] Use typo metrics if OS/2 fsSelection USE_TYPO_METRICS bit is set.
If the OS/2 table exists and fsSelection bit 7 (USE_TYPO_METRICS) is set,
diff --git a/src/truetype/ttgxvar.c b/src/truetype/ttgxvar.c
index a2e6dd1..787d686 100644
--- a/src/truetype/ttgxvar.c
+++ b/src/truetype/ttgxvar.c
@@ -1325,6 +1325,9 @@
{
GX_Blend blend = face->blend;
GX_Value value, limit;
+ FT_Short mvar_hasc_delta = 0;
+ FT_Short mvar_hdsc_delta = 0;
+ FT_Short mvar_hlgp_delta = 0;
if ( !( face->variation_support & TT_FACE_FLAG_VAR_MVAR ) )
@@ -1359,6 +1362,14 @@
/* since we handle both signed and unsigned values as FT_Short, */
/* ensure proper overflow arithmetic */
*p = (FT_Short)( value->unmodified + (FT_Short)delta );
+
+ /* Treat hasc, hdsc and hlgp specially, see below. */
+ if ( value->tag == MVAR_TAG_HASC )
+ mvar_hasc_delta = (FT_Short)delta;
+ else if ( value->tag == MVAR_TAG_HDSC )
+ mvar_hdsc_delta = (FT_Short)delta;
+ else if ( value->tag == MVAR_TAG_HLGP )
+ mvar_hlgp_delta = (FT_Short)delta;
}
}
@@ -1366,25 +1377,38 @@
{
FT_Face root = &face->root;
-
- if ( face->os2.version != 0xFFFFU )
- {
- if ( face->os2.sTypoAscender || face->os2.sTypoDescender )
- {
- root->ascender = face->os2.sTypoAscender;
- root->descender = face->os2.sTypoDescender;
-
- root->height = root->ascender - root->descender +
- face->os2.sTypoLineGap;
- }
- else
- {
- root->ascender = (FT_Short)face->os2.usWinAscent;
- root->descender = -(FT_Short)face->os2.usWinDescent;
-
- root->height = root->ascender - root->descender;
- }
- }
+ /*
+ * Apply the deltas of hasc, hdsc and hlgp to the FT_Face's ascender,
+ * descender and height attributes, no matter how they were originally
+ * computed.
+ *
+ * (Code that ignores those and accesses the font's metrics values
+ * directly is already served by the delta application code above.)
+ *
+ * The MVAR table supports variations for both typo and win metrics.
+ * According to Behdad Esfahbod, the thinking of the working group was
+ * that no one uses win metrics anymore for setting line metrics (the
+ * specification even calls these metrics "horizontal clipping
+ * ascent/descent", probably for their role on the Windows platform in
+ * computing clipping boxes), and new fonts should use typo metrics, so
+ * typo deltas should be applied to whatever sfnt_load_face decided the
+ * line metrics should be.
+ *
+ * Before, the following led to different line metrics between default
+ * outline and instances, visible when e.g. the default outlines were
+ * used as the regular face and instances for everything else:
+ *
+ * 1. sfnt_load_face applied the hhea metrics by default.
+ * 2. This code later applied the typo metrics by default, regardless of
+ * whether they were actually changed or the font had the OS/2 table's
+ * fsSelection's bit 7 (USE_TYPO_METRICS) set.
+ */
+ FT_Short current_line_gap = root->height - root->ascender +
+ root->descender;
+ root->ascender = root->ascender + mvar_hasc_delta;
+ root->descender = root->descender + mvar_hdsc_delta;
+ root->height = root->ascender - root->descender +
+ current_line_gap + mvar_hlgp_delta;
root->underline_position = face->postscript.underlinePosition -
face->postscript.underlineThickness / 2;