diff --git a/.travis.yml b/.travis.yml
index a4828240..e2bf9cb9 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -5,18 +5,6 @@ git:
 
 matrix:
   include:
-    - env:
-      - VIM_VERSION=v7.4
-      - MAKE_TARGET=test
-      - TEST_PROFILE=vim-profile-v7.4.txt
-    - env:
-      - VIM_VERSION=v8.0.0000
-      - MAKE_TARGET=test
-      - TEST_PROFILE=vim-profile-v8.0.txt
-    - env:
-      - VIM_VERSION=master
-      - MAKE_TARGET=test
-      - TEST_PROFILE=vim-profile-master.txt
     - env:
       - VIM_VERSION=installed
       - MAKE_TARGET="clean_compiled check js/test py/test test/node_position/test_position.out"
@@ -34,7 +22,7 @@ script:
   - uname -a
   - which -a vim
   - vim --cmd version --cmd quit
-  - make $MAKE_TARGET
+  - make --keep-going $MAKE_TARGET
 
 after_success:
   - covimerage write_coverage $TEST_PROFILE
diff --git a/autoload/vimlparser.vim b/autoload/vimlparser.vim
index 4b897ef7..55a316c9 100644
--- a/autoload/vimlparser.vim
+++ b/autoload/vimlparser.vim
@@ -239,7 +239,7 @@ function! s:iswhite(c)
 endfunction
 
 function! s:isnamec(c)
-  return a:c =~# '^[0-9A-Za-z_:#]$'
+  return a:c =~# '^[0-9A-Za-z_#]$'
 endfunction
 
 function! s:isnamec1(c)
@@ -3570,7 +3570,8 @@ function! s:ExprParser.parse_dot(token, left)
   endif
   let pos = self.reader.getpos()
   let name = self.reader.read_word()
-  if s:isnamec(self.reader.p(0))
+  let c = self.reader.p(0)
+  if c ==# ':' || c ==# '#'
     " XXX: foo is str => ok, foo is obj => invalid expression
     " foo.s:bar or foo.bar#baz
     return s:NIL
@@ -3979,7 +3980,18 @@ function! s:StringReader.read_nonwhite()
 endfunction
 
 function! s:StringReader.read_name()
-  let r = ''
+  " We come here for valid names only.
+  let r = self.getn(1)
+
+  " Colon only allowed at 2nd position, with scope identifiers.
+  let c = self.peekn(1)
+  if c ==# ':'
+    if r !~# '[vgslabwt]'
+      return r
+    endif
+    let r .= self.getn(1)
+  endif
+
   while s:isnamec(self.peekn(1))
     let r .= self.getn(1)
   endwhile
diff --git a/js/vimlfunc.js b/js/vimlfunc.js
index b4dbd260..a1ad4af3 100644
--- a/js/vimlfunc.js
+++ b/js/vimlfunc.js
@@ -78,11 +78,12 @@ var pat_vim2js = {
   "^[0-9A-Fa-f]$" : "^[0-9A-Fa-f]$",
   "^[0-9A-Za-z_]$" : "^[0-9A-Za-z_]$",
   "^[A-Za-z_]$" : "^[A-Za-z_]$",
-  "^[0-9A-Za-z_:#]$" : "^[0-9A-Za-z_:#]$",
+  "^[0-9A-Za-z_#]$" : "^[0-9A-Za-z_#]$",
   "^[A-Za-z_][0-9A-Za-z_]*$" : "^[A-Za-z_][0-9A-Za-z_]*$",
   "^[A-Z]$" : "^[A-Z]$",
   "^[a-z]$" : "^[a-z]$",
   "^[vgslabwt]:$\\|^\\([vgslabwt]:\\)\\?[A-Za-z_][0-9A-Za-z_#]*$" : "^[vgslabwt]:$|^([vgslabwt]:)?[A-Za-z_][0-9A-Za-z_#]*$",
+  "[vgslabwt]" : "[vgslabwt]",
   "^[0-7]$" : "^[0-7]$",
 }
 
diff --git a/js/vimlparser.js b/js/vimlparser.js
index 236abe61..33f78b26 100644
--- a/js/vimlparser.js
+++ b/js/vimlparser.js
@@ -78,11 +78,12 @@ var pat_vim2js = {
   "^[0-9A-Fa-f]$" : "^[0-9A-Fa-f]$",
   "^[0-9A-Za-z_]$" : "^[0-9A-Za-z_]$",
   "^[A-Za-z_]$" : "^[A-Za-z_]$",
-  "^[0-9A-Za-z_:#]$" : "^[0-9A-Za-z_:#]$",
+  "^[0-9A-Za-z_#]$" : "^[0-9A-Za-z_#]$",
   "^[A-Za-z_][0-9A-Za-z_]*$" : "^[A-Za-z_][0-9A-Za-z_]*$",
   "^[A-Z]$" : "^[A-Z]$",
   "^[a-z]$" : "^[a-z]$",
   "^[vgslabwt]:$\\|^\\([vgslabwt]:\\)\\?[A-Za-z_][0-9A-Za-z_#]*$" : "^[vgslabwt]:$|^([vgslabwt]:)?[A-Za-z_][0-9A-Za-z_#]*$",
+  "[vgslabwt]" : "[vgslabwt]",
   "^[0-7]$" : "^[0-7]$",
 }
 
@@ -426,7 +427,7 @@ function iswhite(c) {
 }
 
 function isnamec(c) {
-    return viml_eqregh(c, "^[0-9A-Za-z_:#]$");
+    return viml_eqregh(c, "^[0-9A-Za-z_#]$");
 }
 
 function isnamec1(c) {
@@ -3455,7 +3456,8 @@ ExprParser.prototype.parse_dot = function(token, left) {
     }
     var pos = this.reader.getpos();
     var name = this.reader.read_word();
-    if (isnamec(this.reader.p(0))) {
+    var c = this.reader.p(0);
+    if (c == ":" || c == "#") {
         // XXX: foo is str => ok, foo is obj => invalid expression
         // foo.s:bar or foo.bar#baz
         return NIL;
@@ -3883,7 +3885,16 @@ StringReader.prototype.read_nonwhite = function() {
 }
 
 StringReader.prototype.read_name = function() {
-    var r = "";
+    // We come here for valid names only.
+    var r = this.getn(1);
+    // Colon only allowed at 2nd position, with scope identifiers.
+    var c = this.peekn(1);
+    if (c == ":") {
+        if (!viml_eqregh(r, "[vgslabwt]")) {
+            return r;
+        }
+        r += this.getn(1);
+    }
     while (isnamec(this.peekn(1))) {
         r += this.getn(1);
     }
diff --git a/py/vimlfunc.py b/py/vimlfunc.py
index 8cc75216..e0f42efb 100644
--- a/py/vimlfunc.py
+++ b/py/vimlfunc.py
@@ -74,11 +74,12 @@ class AttributeDict(dict):
   "^[0-9A-Fa-f]$" : "^[0-9A-Fa-f]$",
   "^[0-9A-Za-z_]$" : "^[0-9A-Za-z_]$",
   "^[A-Za-z_]$" : "^[A-Za-z_]$",
-  "^[0-9A-Za-z_:#]$" : "^[0-9A-Za-z_:#]$",
+  "^[0-9A-Za-z_#]$" : "^[0-9A-Za-z_#]$",
   "^[A-Za-z_][0-9A-Za-z_]*$" : "^[A-Za-z_][0-9A-Za-z_]*$",
   "^[A-Z]$" : "^[A-Z]$",
   "^[a-z]$" : "^[a-z]$",
   "^[vgslabwt]:$\\|^\\([vgslabwt]:\\)\\?[A-Za-z_][0-9A-Za-z_#]*$" : "^[vgslabwt]:$|^([vgslabwt]:)?[A-Za-z_][0-9A-Za-z_#]*$",
+  "[vgslabwt]" : "[vgslabwt]",
   "^[0-7]$" : "^[0-7]$",
 }
 
diff --git a/py/vimlparser.py b/py/vimlparser.py
index 760cdc50..29816b72 100644
--- a/py/vimlparser.py
+++ b/py/vimlparser.py
@@ -74,11 +74,12 @@ class AttributeDict(dict):
   "^[0-9A-Fa-f]$" : "^[0-9A-Fa-f]$",
   "^[0-9A-Za-z_]$" : "^[0-9A-Za-z_]$",
   "^[A-Za-z_]$" : "^[A-Za-z_]$",
-  "^[0-9A-Za-z_:#]$" : "^[0-9A-Za-z_:#]$",
+  "^[0-9A-Za-z_#]$" : "^[0-9A-Za-z_#]$",
   "^[A-Za-z_][0-9A-Za-z_]*$" : "^[A-Za-z_][0-9A-Za-z_]*$",
   "^[A-Z]$" : "^[A-Z]$",
   "^[a-z]$" : "^[a-z]$",
   "^[vgslabwt]:$\\|^\\([vgslabwt]:\\)\\?[A-Za-z_][0-9A-Za-z_#]*$" : "^[vgslabwt]:$|^([vgslabwt]:)?[A-Za-z_][0-9A-Za-z_#]*$",
+  "[vgslabwt]" : "[vgslabwt]",
   "^[0-7]$" : "^[0-7]$",
 }
 
@@ -372,7 +373,7 @@ def iswhite(c):
     return viml_eqregh(c, "^[ \\t]$")
 
 def isnamec(c):
-    return viml_eqregh(c, "^[0-9A-Za-z_:#]$")
+    return viml_eqregh(c, "^[0-9A-Za-z_#]$")
 
 def isnamec1(c):
     return viml_eqregh(c, "^[A-Za-z_]$")
@@ -2752,7 +2753,8 @@ def parse_dot(self, token, left):
             return NIL
         pos = self.reader.getpos()
         name = self.reader.read_word()
-        if isnamec(self.reader.p(0)):
+        c = self.reader.p(0)
+        if c == ":" or c == "#":
             # XXX: foo is str => ok, foo is obj => invalid expression
             # foo.s:bar or foo.bar#baz
             return NIL
@@ -3085,7 +3087,14 @@ def read_nonwhite(self):
         return r
 
     def read_name(self):
-        r = ""
+        # We come here for valid names only.
+        r = self.getn(1)
+        # Colon only allowed at 2nd position, with scope identifiers.
+        c = self.peekn(1)
+        if c == ":":
+            if not viml_eqregh(r, "[vgslabwt]"):
+                return r
+            r += self.getn(1)
         while isnamec(self.peekn(1)):
             r += self.getn(1)
         return r
diff --git a/test/test_slice_var.ok b/test/test_slice_var.ok
new file mode 100644
index 00000000..61176042
--- /dev/null
+++ b/test/test_slice_var.ok
@@ -0,0 +1,5 @@
+(let = s '123')
+(let = a 1)
+(let = b (- 1))
+(echo (slice s a b))
+(echo (slice s a b))
diff --git a/test/test_slice_var.vim b/test/test_slice_var.vim
new file mode 100644
index 00000000..774cc101
--- /dev/null
+++ b/test/test_slice_var.vim
@@ -0,0 +1,5 @@
+let s = '123'
+let a = 1
+let b = -1
+echo s[a : b]
+echo s[a:b]