On Mon, Nov 9, 2020 at 8:21 AM Marcel Fabian Krüger
Hi,
as recently discussed on the luatex list there is an issue regarding hyphenation ehn a frozen glyph appears next to an explicit hyphenation character. E.g.
\setbox0\hbox{X} X-\unhbox0 \bye
triggers an assertion error. I noticed that with the latest experimental code a TeX error is triggered instead but IMO this should not be an error. Therefore I suggest to fix this by treating glyph nodes failing is_simple_character just like non-glyph nodes, e.g. with
--- source/texk/web2c/luatexdir/lang/texlang.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/source/texk/web2c/luatexdir/lang/texlang.c b/source/texk/web2c/luatexdir/lang/texlang.c index 48a242fa0..e7635903a 100644 --- a/source/texk/web2c/luatexdir/lang/texlang.c +++ b/source/texk/web2c/luatexdir/lang/texlang.c @@ -1049,7 +1049,7 @@ void hnj_hyphenation(halfword head, halfword tail) if (explicit_hyphen == true) { /*tex we are not at the start, so we only need to look ahead */ halfword t = vlink(r) ; - if ((automatic_hyphen_mode_par == 0 || automatic_hyphen_mode_par == 1) && (t != null) && ((type(t) == glyph_node) && (character(t) != ex_hyphen_char_par))) { + if ((automatic_hyphen_mode_par == 0 || automatic_hyphen_mode_par == 1) && (t != null) && ((type(t) == glyph_node) && is_simple_character(t) && (character(t) != ex_hyphen_char_par))) { /*tex we have a word already but the next character may not be a hyphen too */ r = compound_word_break(r, char_lang(r)); if (compound_hyphen) { @@ -1063,7 +1063,7 @@ void hnj_hyphenation(halfword head, halfword tail) } } else { /*tex we jump over the sequence of hyphens */ - while ((t != null) && (type(t) == glyph_node) && (character(t) == ex_hyphen_char_par)) { + while ((t != null) && (type(t) == glyph_node) && is_simple_character(t) && (character(t) == ex_hyphen_char_par)) { r = t ; t = vlink(r) ; } -- 2.29.2
Best regards, Marcel
ok, I will see it this weekend. -- luigi