···<date: 2014-01-06, Monday>···<from: Stephan Hennig>···
> Am 05.01.2014 19:03, schrieb Philipp Gesang: > > > > It depends on whether the “liga” feature is active and whether > > you use base mode or node mode. > > Well, let me add a related question. To quote from the luaotfload manual: > > > node mode works by processing TeX's internal node list directly at > > the Lua end > > What callback does luaotfload use for the processing? The font handler is applied to both the pre_linebreak_filter and the hpack_filter. https://github.com/phi-gamma/luaotfload/blob/texlive2014/luaotfload.dtx#L1919 > And what is the > way to go for packages fighting for a callback? luatexbase.add_to_callback() without priority (append). > If callbacks are "just > chained", how can I be sure that luaotfload has already done its > processing before my package? Luaotfload inserts the callbacks into the chain at highest priority. As long as you do not specify a priority, you should be fine. > > E.g. in Context: [...] The values are: > > > > mode | liga > > | - + > > -----+-------- > > base | 256 2 > > node | 256 258 > > > > Run this with the bare fontloader, you get > > > > mode | liga > > | - + > > -----+-------- > > base | 256 2 > > node | 256 256 > > > > With luaotfload, however, it’s > > > > mode | liga > > | - + > > -----+-------- > > base | 256 258 > > node | 256 256 > > OK, what does that teach me? That you should be able to call the inverse operation (node.unprotect_glyphs()) on the node list to restore the original values. > I think I'll just look at (and follow) the > components field to parse ligatures and hope for the best ... I came up with this heuristic: http://tex.stackexchange.com/a/120529/14066 Best, Philipp
pgpTZW0E5FLGG.pgp
Description: PGP signature
