![](https://secure.gravatar.com/avatar/2259586bbe3f45ed8564afe9bd7a8439.jpg?s=120&d=mm&r=g)
2009/3/2 Hans Hagen
Yanrui Li wrote:
2009/2/25 Hans Hagen
: Hi,
As prelude to LuaTeX 0.40 there is now a beta 0.35. I uploaded a context beta that matches this version. The most important move is that we now have a new math subsystem. (Specially for Mojca: iwona seems to work ok). Apart from bringin unicode math this also gives a better performance. (Some info about what is involved can be found in mk.pdf on the website).
There might be ommisions so when you run into one, just let us know.
Hi,
Maybe there are some bugs in scrp-ini.lua and scrp-cjk.lua as follows:
1. In scrp-ini.lua, some flags of punctuations in the definition of "hash" are replaced by "chinese" flags, for example:
local hash = { ... ... [0xFF0C] = "full_width_close", -- , ... ... } ... ... for i=0x0FF00,0x0FFEF do hash[i] = "chinese" end ... ...
2. In scrp-cjk.lua, it seems that "insert_node_before" in some break & shrink & stretch functions should be "insert_node_after", for example:
local function nobreak_stretch(head,current) insert_node_before(head,current,make_penalty_node(10000)) insert_node_before(head,current,make_glue_node(0,inter_char_stretch,0)) end
we only insert nodes before, so in that case you need to look at another combination; so, like
[insert] is triggered only once, by the table entry
[class a] = { class_b = ... }
Thanks for your explanation, I see. There are some Chinese punctuations, such as “," (fullwidth comma), "!" (fullwidth exclamation mark) and so on, which are considered to be "chinese" but not "full_width_close" by scrp-cjk.lua. Is this deliberately doing? -- Best wishes, Li Yanrui * Gentoo Linux (~x86) * LuaTeX (snapshot-0.35.0-2009022702) * ConTeXt Minimals (2009.02.25 14:43)