leak of post table?

[ebraminio]

Trying to make #1169 happen,

If I apply this, hb-shape's leak (on asan) will be resolved,

diff --git a/src/hb-ot-font.cc b/src/hb-ot-font.cc
index faacecb9..9a1c2e9d 100644
--- a/src/hb-ot-font.cc
+++ b/src/hb-ot-font.cc
@@ -160,8 +160,9 @@ hb_ot_get_glyph_name (hb_font_t *font HB_UNUSED,
                       char *name, unsigned int size,
                       void *user_data HB_UNUSED)
 {
-  const hb_ot_face_data_t *ot_face = (const hb_ot_face_data_t *) font_data;
-  return ot_face->post->get_glyph_name (glyph, name, size);
+  // const hb_ot_face_data_t *ot_face = (const hb_ot_face_data_t *) font_data;
+  // return ot_face->post->get_glyph_name (glyph, name, size);
+  return false;
 }

Apparently post's fini is not called

Any hint?

[behdad]

I think your printf didn't print because it had no newline and process was killed before it flushed. fini() was being called. Just that it was not freeing blob.