summaryrefslogtreecommitdiff
path: root/tex/context/base/mkiv/l-lpeg.lua
diff options
context:
space:
mode:
authorHans Hagen <pragma@wxs.nl>2018-02-19 19:19:07 +0100
committerContext Git Mirror Bot <phg42.2a@gmail.com>2018-02-19 19:19:07 +0100
commit9bb459e9fedcad94579973ccd07ad9e78d0734c8 (patch)
treee01ac3174e8891a6e6004c88cb46f220312a2bfc /tex/context/base/mkiv/l-lpeg.lua
parent5321d35d68ce0acb71b46bddac1d8cab8eca3fcc (diff)
downloadcontext-9bb459e9fedcad94579973ccd07ad9e78d0734c8.tar.gz
2018-02-19 19:05:00
Diffstat (limited to 'tex/context/base/mkiv/l-lpeg.lua')
-rw-r--r--tex/context/base/mkiv/l-lpeg.lua8
1 files changed, 8 insertions, 0 deletions
diff --git a/tex/context/base/mkiv/l-lpeg.lua b/tex/context/base/mkiv/l-lpeg.lua
index fc1357afc..a7ebd567d 100644
--- a/tex/context/base/mkiv/l-lpeg.lua
+++ b/tex/context/base/mkiv/l-lpeg.lua
@@ -6,6 +6,10 @@ if not modules then modules = { } end modules ['l-lpeg'] = {
license = "see context related readme files"
}
+-- we can get too many captures (e.g. on largexml files) which makes me wonder
+-- if P(foo)/"" can't be simplfied to N(foo) i.e. some direct instruction to the
+-- lpeg virtual machine to ignore it
+
-- lpeg 12 vs lpeg 10: slower compilation, similar parsing speed (i need to check
-- if i can use new features like capture / 2 and .B (at first sight the xml
-- parser is some 5% slower)
@@ -17,6 +21,10 @@ if not modules then modules = { } end modules ['l-lpeg'] = {
-- move utf -> l-unicode
-- move string -> l-string or keep it here
+-- lpeg.B : backward without consumption
+-- lpeg.F = getmetatable(lpeg.P(1)).__len : forward without consumption
+
+
lpeg = require("lpeg") -- does lpeg register itself global?
local lpeg = lpeg