Require a word (including its subscript) to end in a number or letter.

[matthijs/master-project/report.git] / pret-lam.lua

diff --git a/pret-lam.lua b/pret-lam.lua
index f8ddb676088a7870fd94a979276d3283aa7b7dd6..feaa5e43934cbb157757b9152c38a6ea8f624c72 100644 (file)
--- 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.
@@ -221,12 +228,16 @@ vis.begin_of_inline = vis.begin_of_display
 vis.end_of_inline = vis.end_of_display
 
 function vis.flush_line(str,nested)
-    buffers.flush_result(vis.do_line(str), nested)   
+    buffers.flush_result(vis.do_line(str, false), nested)   
 end
 
-function vis.do_line(str)
-    result = {}
-    str = do_indent(str)
+function vis.do_line(str, no_indent)
+    local result = {}
+    if not no_indent then
+        -- Allow ignore of the indentation stuff when we're calling ourselves
+        -- for a partial line.
+        str = do_indent(str)
+    end
     while str ~= "" do
         local found = false
         local word, symbol
@@ -234,10 +245,17 @@ function vis.do_line(str)
         if text then
             table.insert(result, '\\strikethrough{')
             -- Recursively call ourselves to handle spaces gracefully.
-            result = array_concat(result, vis.do_line(text))
+            result = array_concat(result, vis.do_line(text, true))
             table.insert(result, '}')
             -- Eat the processed characters
             str = rest
+        elseif utf.match(str, "^%-%-") then
+            table.insert(result, '{\\italic{--')
+            -- Recursively call ourselves to handle spaces gracefully.
+            result = array_concat(result, vis.do_line(utf.sub(str, 3), true))
+            table.insert(result, '}}')
+            -- Done with this line
+            str = ''
         else
             -- See if the next token is a word
             word, str = take_word(str)