tag 441934 +patch thank you The attached patch should care of "­". It does not work if the SOFT HYPHEN character appears verbatim in a properly-advertised Unicode document.
As of 0.5.3-2, the question-marks you reported for "‌" and "‍" appear when LC_CTYPE (or LANG) is set to "C". If I leave LC_CTYPE set to en_US.utf-8, the proper characters are emitted, but neither xterm nor urxvt render them. -Hilko
diff --git a/entity.c b/entity.c index fdd8f64..a7c9e99 100644 --- a/entity.c +++ b/entity.c @@ -44,6 +44,8 @@ conv_entity(unsigned int c) return " "; if (c == 0xa0) return NBSP; + if (c == 0xad) /* SOFT HYPHEN */ + return ""; if (c < 0x100) { /* Latin1 (ISO 8859-1) */ if (UseAltEntity) return alt_latin1[c - 0xa0];