octave.lang: several improvements including classdef and line continuation fixes.

* update syntax highlight for the new line-continuation rules, where
  anything after ... is considered a comment.  Also, \ can now only
  be used in double quoted strings so we just pick def:string.
* simplify the operators regex (both performance and readability)
* drop the over complicated and incorrect matches for decimal,
  floating-point, octal, and hexadecimal highlight.  These must have
  been copied from other file, since Octave does not support the "f",
  "u", or "l", syntax they cover.  It can all be simplified by using
  the definitions in def.lang.
* removed highlight of the variable "ans", because that's used at the
  interactive prompt, and not in scripts.
* expand list of keywords to include the new classdef system and unwrap
  the regexs inside a single keyword for readability.
* with octave.lang now including classdef keywords, drop all keywords
  from matlab.lang and simply include the matlab set from octave.lang.
* remove all number related from matlab.lang and pick def:decimal
  like octave.lang does.  Do not pick hexadecimal because Matlab has
  never supported it.

2 files changed, 76 insertions, 102 deletions

diff --git a/data/language-specs/matlab.lang b/data/language-specs/matlab.lang
index a12fca76..6343e450 100644
--- a/data/language-specs/matlab.lang
+++ b/data/language-specs/matlab.lang
@@ -6,7 +6,7 @@
  Authors: Søren Hauberg, Muthiah Annamalai, Carnë Draug
  Copyright (C) 2006 Søren Hauberg <soren@hauberg.org>
  Copyright (C) 2006 Muthiah Annamalai <gnumuthu@users.sf.net>
- Copyright (C) 2010 Carnë Draug <carandraug+dev@gmail.com>
+ Copyright (C) 2010-2015 Carnë Draug <carandraug+dev@gmail.com>
 
  GtkSourceView is free software; you can redistribute it and/or
  modify it under the terms of the GNU Lesser General Public
@@ -45,7 +45,8 @@
 
   <definitions>
 
-    <context id="block-comment" style-ref="comment" class="comment" class-disabled="no-spell-check">
+    <context id="block-comment" style-ref="comment" class="comment"
+             class-disabled="no-spell-check">
       <start>^\s*%{\s*$</start>
       <end>^\s*%}\s*$</end>
       <include>
@@ -54,34 +55,14 @@
       </include>
     </context>
 
-    <context id="line-comment" style-ref="comment" end-at-line-end="true" class="comment" class-disabled="no-spell-check">
+    <context id="line-comment" style-ref="comment" end-at-line-end="true"
+             class="comment" class-disabled="no-spell-check">
       <start>%</start>
       <include>
         <context ref="def:in-comment"/>
       </include>
     </context>
 
-    <context id="keyword">
-      <include>
-        <context style-ref="keyword">
-          <!--Only Matlab specific keywords. -->
-          <keyword>classdef</keyword>
-          <keyword>enumeration</keyword>
-          <keyword>events</keyword>
-          <keyword>methods</keyword>
-          <keyword>properties</keyword>
-        </context>
-        <!--
-          Octave and Matlab have already implemented most of each others keywords
-          but not all. Both Octave and Matlab are still missing keywords from the
-          other. As such, this block only has the keywords already implemented in
-          both. When adding a new keyword, make sure you place it on the right
-          place.
-        -->
-        <context ref="octave:octave-matlab-keyword" style-ref="keyword"/>
-      </include>
-    </context>
-
     <!-- FIXME: Matlab probably has much more builtin functions
      this is just the list that was already in the lang file before
      being split from octave.lang -->
@@ -156,13 +137,9 @@
         <context ref="octave:function-handle"/>
         <context ref="octave:storage-type"/>
         <context ref="octave:boolean"/>
-        <context ref="octave:decimal"/>
-        <context ref="octave:floating-point-number"/>
-        <context ref="octave:octal-number"/>
-        <context ref="octave:hex-number"/>
-        <context ref="octave:variables"/>
+        <context ref="def:decimal"/>
         <context ref="octave:reserved-constant"/>
-        <context ref="keyword"/>
+        <context ref="octave:octave-matlab-keyword"/>
         <context ref="builtin"/>
       </include>
     </context>
diff --git a/data/language-specs/octave.lang b/data/language-specs/octave.lang
index 4b276e84..d0e3c2e0 100644
--- a/data/language-specs/octave.lang
+++ b/data/language-specs/octave.lang
@@ -6,7 +6,7 @@
  Authors: Søren Hauberg, Muthiah Annamalai, Carnë Draug
  Copyright (C) 2006 Søren Hauberg <soren@hauberg.org>
  Copyright (C) 2006 Muthiah Annamalai <gnumuthu@users.sf.net>
- Copyright (C) 2010 Carnë Draug <carandraug+dev@gmail.com>
+ Copyright (C) 2010-2015 Carnë Draug <carandraug+dev@gmail.com>
 
  GtkSourceView is free software; you can redistribute it and/or
  modify it under the terms of the GNU Lesser General Public
@@ -44,24 +44,29 @@
     <style id="data-type"         _name="Data type"         map-to="def:type"/>
     <style id="storage-type"      _name="Storage Type"      map-to="def:type"/>
     <style id="boolean"           _name="Boolean"           map-to="def:boolean"/>
-    <style id="decimal"           _name="Decimal"           map-to="def:decimal"/>
-    <style id="floating-point"    _name="Floating Point"    map-to="def:floating-point"/>
-    <style id="base-n-integer"    _name="Base-N Integer"    map-to="def:base-n-integer"/>
     <style id="package-manager"   _name="Package Manager"   map-to="def:preprocessor"/>
     <style id="keyword"           _name="Keyword"           map-to="def:keyword"/>
-    <style id="variable"          _name="Variable"          map-to="def:identifier"/>
     <style id="reserved-constant" _name="Reserved Constant" map-to="def:special-constant"/>
     <style id="builtin"           _name="Builtin"           map-to="def:builtin"/>
   </styles>
 
   <definitions>
 
+    <!--
+      Many variable looking things such as NaN, NA, true, ones, and zeros,
+      are actually functions that take dimensions, and even class name as
+      argument.  However, are often used as a scalar variables since they
+      default to (1, 1) dimensions.  In such cases, when they are not
+      followed by parentheses, we highlight them as variables.
+    -->
+    <define-regex id="function-as-variable-suffix">\%](?!\s*\()</define-regex>
+
     <context id="block-comment" style-ref="comment" class="comment"
              class-disabled="no-spell-check">
       <start>^\s*[#%]{\s*$</start>
       <end>^\s*[#%]}\s*$</end>
       <include>
-        <context ref="block-comment"/>
+        <context ref="block-comment"/>  <!-- Nested block comments are real -->
         <context ref="def:in-comment"/>
       </include>
     </context>
@@ -75,12 +80,14 @@
     </context>
 
     <!--
-      Line continue in Octave are different from the one in def.lang. They can
-      either be  ... or \ optionally followed by whitespace only.
+      A backslash as line continuation can only be used within double
+      quoted strings.  An ellipsis as line continuation can be used
+      anywhere else to continue statements and anything that follows
+      a "..." is ignored, treated as a comment.
     -->
     <context id="line-continue" style-ref="comment">
-      <start>(\.\.\.|\\)(\s)*?$</start>
-      <end>^</end>
+      <start>\.\.\.</start>
+      <end>$</end>
       <include>
         <context ref="def:in-comment"/>
       </include>
@@ -90,7 +97,7 @@
              end-at-line-end="true" class="string"
              class-disabled="no-spell-check">
       <!--
-         To do not misdetect the transpose operator ' as the start of a string
+         To not misdetect the transpose operator ' as the start of a string
          we assert to not follow a variable name (letters, digits and underscores)
          or a closing bracket (round, square or curly) or a dot (to form the
          array transpose operator ".'" ).
@@ -110,20 +117,24 @@
       </include>
     </context>
 
-    <!--To see the list of operators, use the function __operators__-->
+    <!--
+      The Octave function __operators__ does not return an exhaustive list
+      of operators.  Do not forget that all operators can now be preceded
+      by "." to perform automatic broadcasting, including comparison.
+      Autoincrement operators are also not listed there.
+    -->
     <context id="operator" style-ref="operator">
       <match extended="true">
-        (\.)?\+{1,2}?(?!\+) |   # arithmetic operators
-        (\.)?\-{1,2}?(?!\-) |   # already counting
-        (\.)?\*{1,2}?(?!\*) |   # with auto increment,
-        (\.)?\/(?!\^) |         # element by element,
-        (\.)?\\(?!\^) |         # left division
-        (\.)?\^(?!\^) |         # and both exp
-        (?&lt;=[0-9a-zA-Z_)\]}])(\.)?' |                  # transpose operator
-        &lt;=? | &gt;=? | != | ~= | == | &lt;&gt; |       # comparison operators
-        &amp;{1,2}?(?!&amp;) | \|{1,2}?(?!\|) | ! | ~ |   # boolean operators
-        = |                                               # assignment operator
-        : |                                               # range operator
+        ## The operators, +, -, *, /, \, ^, and **, support both element by
+        ## element-by-element and automatic broadcasting (preceded by .),
+        ## compound assignment, as well both in one (".**=" is real).
+        (\.)?(\+|\-|\*|\/|\\|\^|\*\*) |
+
+        (?&lt;=[0-9a-zA-Z_)\]}])(\.)?' |  # transpose operator
+        \+\+ | \-\- |                     # auto-increment/decrement
+        \| | &amp; |                      # boolean operators
+        ! | ~ | = | &lt; | &gt; |         # comparison operators
+        :                                 # range operator
       </match>
     </context>
 
@@ -147,41 +158,13 @@
     </context>
 
     <context id="boolean" style-ref="boolean">
-      <!--
-        false and true are actually functions but let's ignore that
-        unless there are parentheses.
-      -->
-      <suffix>\b(?!(\s)*\()</suffix>
+      <suffix>\%{function-as-variable-suffix}</suffix>
       <keyword>false</keyword>
       <keyword>true</keyword>
     </context>
 
-    <context id="decimal" style-ref="decimal">
-      <match>\b([1-9][0-9]*|0)([Uu]([Ll]|LL|ll)?|([Ll]|LL|ll)[Uu]?)?\b</match>
-    </context>
-
-    <context id="floating-point-number" style-ref="floating-point">
-      <match>\b([0-9]+[Ee][-]?[0-9]+|([0-9]*\.[0-9]+|[0-9]+\.)([Ee][-]?[0-9]+)?)[fFlL]?</match>
-    </context>
-
-    <context id="octal-number" style-ref="base-n-integer">
-      <match>\b0[0-7]+([Uu]([Ll]|LL|ll)?|([Ll]|LL|ll)[Uu]?)?\b</match>
-    </context>
-
-    <context id="hex-number" style-ref="base-n-integer">
-      <match>\b0[xX][0-9a-fA-F]+([Uu]([Ll]|LL|ll)?|([Ll]|LL|ll)[Uu]?)?\b</match>
-    </context>
-
-    <context id="variables" style-ref="variable">
-      <keyword>ans</keyword>
-    </context>
-
     <context id="reserved-constant" style-ref="reserved-constant">
-      <!--
-        Most of the constants are actually functions. Do not highlight as
-        constants if followed by parentheses.
-      -->
-      <suffix>\b(?!(\s)*\()</suffix>
+      <suffix>\%{function-as-variable-suffix}</suffix>
       <keyword>e</keyword>
       <keyword>eps</keyword>
       <keyword>(J|j|I|i)</keyword>
@@ -195,25 +178,34 @@
     </context>
 
     <context id="package-manager" style-ref="package-manager">
-      <!--
-        The list of pkg commands is optional to the match so that at least pkg
-        is highlighted when new commands get implemented.
-      -->
       <match extended="true">
-        \bpkg(?!(\s)*\()(\s)+
-        ((un)?install|update|(un)?load|list|(global|local)_list|describe|prefix|(re)?build)?\b
+        \%[
+        pkg\%{function-as-variable-suffix}
+        \ +
+        (install|uninstall|update|unload|load|list|global_list|local_list|
+         describe|prefix|rebuild|build)
+        \%]
       </match>
     </context>
 
     <context id="keyword">
       <include>
         <context style-ref="keyword">
-          <!--
-            Only Octave specific keywords. To see a complete list of keywords in
-            Octave, use __keywords__ at the Octave prompt.
-          -->
-          <keyword>end(_try_catch|_unwind_protect|for|function|if|parfor|switch|while)</keyword>
-          <keyword>unwind_protect(_cleanup)?</keyword>
+          <keyword>end_try_catch</keyword>
+          <keyword>end_unwind_protect</keyword>
+          <keyword>endclassdef</keyword>
+          <keyword>endenumeration</keyword>
+          <keyword>endevents</keyword>
+          <keyword>endfor</keyword>
+          <keyword>endfunction</keyword>
+          <keyword>endif</keyword>
+          <keyword>endmethods</keyword>
+          <keyword>endparfor</keyword>
+          <keyword>endproperties</keyword>
+          <keyword>endswitch</keyword>
+          <keyword>endwhile</keyword>
+          <keyword>unwind_protect_cleanup</keyword>
+          <keyword>unwind_protect_cleanup</keyword>
         </context>
         <context id="octave-matlab-keyword" style-ref="keyword">
           <!--
@@ -226,26 +218,34 @@
           <keyword>break</keyword>
           <keyword>case</keyword>
           <keyword>catch</keyword>
+          <keyword>classdef</keyword>
           <keyword>continue</keyword>
           <keyword>do</keyword>
-          <keyword>else(if)?</keyword>
+          <keyword>else</keyword>
+          <keyword>elseif</keyword>
           <keyword>end</keyword>
+          <keyword>enumeration</keyword>
+          <keyword>events</keyword>
           <keyword>for</keyword>
           <keyword>function</keyword>
           <keyword>if</keyword>
+          <keyword>methods</keyword>
           <keyword>otherwise</keyword>
           <keyword>parfor</keyword>
+          <keyword>properties</keyword>
           <keyword>return</keyword>
           <keyword>switch</keyword>
           <keyword>try</keyword>
           <keyword>until</keyword>
-          <keyword>vararg(in|out)</keyword>
+          <keyword>varargin</keyword>
+          <keyword>varargout</keyword>
           <keyword>while</keyword>
         </context>
       </include>
     </context>
 
     <context id="builtin" style-ref="builtin">
+      <!--  If any of these is between "." it's a struct.  -->
       <prefix>(?&lt;!\.)\%[</prefix>
       <suffix>\%](?!\.)</suffix>
 
@@ -1745,11 +1745,8 @@
         <context ref="function-handle"/>
         <context ref="storage-type"/>
         <context ref="boolean"/>
-        <context ref="decimal"/>
-        <context ref="floating-point-number"/>
-        <context ref="octal-number"/>
-        <context ref="hex-number"/>
-        <context ref="variables"/>
+        <context ref="def:decimal"/>
+        <context ref="def:hexadecimal"/>
         <context ref="reserved-constant"/>
         <context ref="package-manager"/>
         <context ref="keyword"/>