X-Git-Url: https://git.stderr.nl/gitweb?p=matthijs%2Fmaster-project%2Freport.git;a=blobdiff_plain;f=pret-lam.lua;h=feaa5e43934cbb157757b9152c38a6ea8f624c72;hp=43e8c8cc25f305ecf44b10e7d5a9b3b4311ca0ea;hb=HEAD;hpb=65f605d9f4c2420383c683c26c01be253644372b

diff --git a/pret-lam.lua b/pret-lam.lua
index 43e8c8c..feaa5e4 100644
--- a/pret-lam.lua
+++ b/pret-lam.lua
@@ -19,17 +19,20 @@ local colors = {
 local symbols = {
     -- Note, the space we replace with is a Unicode non-breaking space
     -- (U+00A0).
-    [' '] = {repr = ' '},
-    ['_'] = {repr = '\\_'},
-    ['->'] = {repr = '→'},
-    ['=>'] = {repr = '⇒'},
+    {symbol = ' ', repr = ' '},
+    {symbol = '_', repr = '\\_'},
+    {symbol = '->>', repr = '\\twoheadrightarrow'},
+    {symbol = '->', repr = '→'},
+    {symbol = '=>', repr = '⇒'},
     -- The default * sits very high above the baseline, \ast (u+2217) looks
     -- better.
-    ['*'] = {repr = '\\ast'},
-    ['~'] = {repr = '\\HDLine[width=.20 * \\the\\textwidth]'},
-    ['|'] = {repr = '\\char' .. utf.byte('|')},
-    -- Use ▶ from our roman font, since Iwona doesn't have the glyph
-    ['▶'] = {repr = '{\\rm{}▶}'},
+    {symbol = '*', repr = '\\ast'},
+    {symbol = '~', repr = '\\HDLine[width=.20 * \\the\\textwidth]'},
+    {symbol = '|', repr = '\\char' .. utf.byte('|')},
+    -- Use ▶ and ∧ from our roman font, since Iwona doesn't have the glyph
+    {symbol = '▶', repr = '{\\rm{}▶}'},
+    --{symbol = '∧', repr = '{$∧$}'},
+    {symbol = '∧', repr = '{\\rm{}∧}'},
 }
 
 -- Keywords that should be bold
@@ -63,9 +66,9 @@ end
 -- character.  We can do a lookup directly, since symbols can be different in
 -- length, so we just loop over all symbols, trying them in turn.
 local function take_symbol(str)
-    for symbol,props in pairs(symbols) do
+    for i,props in ipairs(symbols) do
         -- Try to remove symbol from the start of str 
-        symbol, newstr = utf.match(str, "^(" .. symbol .. ")(.*)")
+        symbol, newstr = utf.match(str, "^(" .. props.symbol .. ")(.*)")
         if symbol then
             -- Return this tokens repr, or just the token if it has no
             -- repr.
@@ -85,8 +88,12 @@ end
 -- the word taken.
 local function take_word(str)
         -- A word must always start with a-z (in particular, λ is not a valid
-        -- start of a word).
-        res, newstr = utf.match(str, "^([a-zA-Z][%a%d%+%-%,_]+)(.*)")
+        -- start of a word). A word must always end with a letter or a digit
+        res, newstr = utf.match(str, "^([a-zA-Z][%a%d%+%-%,_]*[%a%d]+)(.*)")
+        if not res then
+            -- The above does not catch single letter words
+            res, newstr = utf.match(str, "^([a-zA-Z])(.*)")
+        end
         return res, newstr or str
 end
 
@@ -206,7 +213,7 @@ function vis.begin_of_display()
     if in_block == 0 then
         -- Initially allow subscripts using _ or just appending a number (later,
         -- we will add extra patterns here.
-        submatches = {"^(%a*)_([%a%d,]+)$", "^(%a+)(%d[%d,]+)$"}
+        submatches = {"^(%a*)_([%a%d,]+)$", "^(%a+)(%d[%d,]*)$"}
         -- This stores all the bases we've encountered so far (to prevent
         -- duplicates). For each of them there will be a pattern in submatches
         -- above.