Support the | character in pret-lam.

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

diff --git a/pret-lam.lua b/pret-lam.lua
index aff6988d8af774ef8f1e1021f954ed3f34e786d9..9bf5b96bd390d922b2e1ec8f58ce4956706d5dd5 100644 (file)
--- a/pret-lam.lua
+++ b/pret-lam.lua
@@ -6,7 +6,7 @@
 
 local utf = unicode.utf8
 
-vis = buffers.newvisualizer("lam")
+local vis = buffers.newvisualizer("lam")
 
 local colors = {
     "prettytwo",
@@ -24,9 +24,9 @@ local symbols = {
     -- better.
     ['*'] = {repr = '\\ast'},
     ['~'] = {repr = '\\sim'},
+    ['|'] = {repr = '\\char' .. utf.byte('|')},
 }
 
-
 -- Keywords that should be bold
 local keywords = {
     ['case'] = {},
@@ -63,7 +63,7 @@ end
 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_]+)(.*)")
+        res, newstr = utf.match(str, "^([a-zA-Z][%a%d%+%-%,_]+)(.*)")
         return res, newstr or str
 end
 
@@ -90,21 +90,49 @@ local function do_subscripts(word)
         if not bases[base] then
             -- Register that we've added this base
             bases[base] = true
-            -- Add a pattern for this base
-            submatches[#submatches+1] = "^(" .. base .. ")([%a%d,]+)$"
+            -- Add a patterns for this base. First, the base with a single
+            -- letter or number subscript.
+            submatches[#submatches+1] = "^(" .. base .. ")([%a%d])$"
+            -- Seconde, the base with a longer prefix that includes at least
+            -- one of +-, (to catch things like ri+1, but not return).
+            submatches[#submatches+1] = "^(" .. base .. ")([%a%d]*[%-%+%,]+[%a%d%-%+%,]*)$"
         end
     end
     return word
 end
 
+local in_block = 0
+
+-- Mark the begin of a block of lambda formatted buffers or expressions. This
+-- means that, until you call end_of_block again, the subscript bases are
+-- shared. For example, if you have \lam{y1} some text \lam{yn} within a
+-- single block, the yn will properly get subscripted. Be sure to call
+-- end_of_block again!
+--
+-- Blocks can be partially nested, meaning that the block
+-- won't be closed until end_of_block was called exactly as often as
+-- begin_of_block. However, subscripts from the inner block can still
+-- influence subscripts in the outer block.
+function vis.begin_of_block()
+    vis.begin_of_display()
+    in_block = in_block + 1
+end
+
+-- Ends the current block
+function vis.end_of_block()
+    in_block = in_block - 1
+end
+
 function vis.begin_of_display()
-    -- Initially allow subscripts using _ or just appending a number (later,
-    -- we will add extra patterns here.
-    submatches = {"^(.*)_([%a%d,]+)$", "^(.*[^%d])(%d+)$"}
-    -- 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.
-    bases = {}
+    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,]+)$", "^(%a+)(%d+)$"}
+        -- 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.
+        bases = {}
+    end
 end
 
 -- Make things work for inline typeing (e.g., \type{}) as well.