libclang.py: Introduce option sorting=none

0.34 -> 0.28 sec completion time on (examples/boost.cpp)

This was compared to soring=alpha

No sorting makes sense as vim will automatically sort the results as
soon as we type the first character. So spending a lot of time on
sorting may not be worth it.

In later patches I will try to speed up our own sorting, such that it
becomes less an issue. clang already provides a c function for this.
Though at the moment this function does not work for me (it crashes).

doc/clang_complete.txt

@@ -170,7 +170,8 @@ Default: ""
 
 					*clang_complete-sort_algo*
 					*g:clang_sort_algo*
-How results are sorted (alpha, priority). Currently only works with libclang.
+How results are sorted (alpha, priority, none). Currently only works with
+libclang.
 Default: "priority"
 
 					*clang_complete-complete_macros*

examples/boost.cpp

@@ -13,6 +13,6 @@
 
 int main()
 {
-	boost:
+	boost::matc
 	//     ^ Code complete here by typing ":"
 }

plugin/libclang.py

@@ -254,7 +254,7 @@ def WarmupCache():
 def getCurrentCompletions(base):
   global debug
   debug = int(vim.eval("g:clang_debug")) == 1
-  priority = vim.eval("g:clang_sort_algo") == 'priority'
+  sorting = vim.eval("g:clang_sort_algo")
   line = int(vim.eval("line('.')"))
   column = int(vim.eval("b:col"))
 
@@ -278,10 +278,10 @@ def getCurrentCompletions(base):
     regexp = re.compile("^" + base)
     results = filter(lambda x: regexp.match(getAbbr(x.string)), results)
 
-  if priority:
+  if sorting == 'priority':
     getPriority = lambda x: x.string.priority
     results = sorted(results, None, getPriority)
-  else:
+  if sorting == 'alpha':
     getAbbrevation = lambda x: getAbbr(x.string).lower()
     results = sorted(results, None, getAbbrevation)