Support encoding markers: skip python/ruby encoding if present in first two lines

README.md

@@ -148,6 +148,19 @@ print "success!"
 <?php echo "success!\n"; ?>
 ```
 
+### File encodings
+
+Python and Ruby support comments to control the file encoding only within
+the first two lines of the file. When using sbang with encoding comments,
+put the shebang line for sbang first, then the encoding comment, and then
+the interpreter to use. E.g.:
+
+```sh
+#!/bin/sh /path/to/sbang
+# -*- coding: latin-1 -*-
+#!/very/long/path/to/python
+```
+
 ## How it works
 
 `sbang` is a very simple POSIX shell script. It looks at the first two

sbang

@@ -22,6 +22,14 @@
 #
 # See https://github.com/spack/sbang for more details.
 #
+# If using python- or ruby- style source code encoding, put that on the second
+# line of the file and the shebang line for the long path to the real
+# interpreter on the third line, like this:
+#
+#     #!/bin/sh /path/to/sbang
+#     # -*- coding: utf-8 -*-
+#     #!/long/path/to/real/interpreter with arguments
+#
 
 # Generic error handling
 die() {
@@ -46,19 +54,39 @@ elif [ ! -f "$script" ]; then
 fi
 
 # Search the first two lines of script for interpreters.
+# Search the third line if one of the first two is an encoding declaration
 lines=0
 while read -r line && [ $lines -ne 2 ]; do
     if [ "${line#\#!}" != "$line" ]; then
         shebang_line="${line#\#!}"
+        lines=$((lines+1))
     elif [ "${line#//!}" != "$line" ]; then      # // comments
         shebang_line="${line#//!}"
+        lines=$((lines+1))
     elif [ "${line#--!}" != "$line" ]; then      # -- lua comments
         shebang_line="${line#--!}"
+        lines=$((lines+1))
     elif [ "${line#<?php\ }" != "$line" ]; then  # php comments
         shebang_line="${line#<?php\ \#!}"
         shebang_line="${shebang_line%\ ?>}"
+        lines=$((lines+1))
+    else
+        # Check whether this is a python/ruby encoding line
+        # Strip leading whitespace, whitespace is allowed before encoding
+        leading_whitespace=${line%%[![:space:]]*}
+        line=${line#"$leading_whitespace"}
+
+        # encoding declaration starts with comment character
+        if [ "${line#\#}" != "$line" ]; then
+            case "$line" in
+                *"coding:"* | *"coding="*)
+                    ;;  # skip encoding line
+                *)
+                    lines=$((lines+1))
+                    ;;
+            esac
+        fi
     fi
-    lines=$((lines+1))
 done < "$script"
 
 # error if we did not find any interpreter

test/shebangs/python_with_encoding_emacs.py

@@ -0,0 +1,5 @@
+#!/usr/bin/env sbang
+# -*- coding: utf-8 -*-
+#!/usr/bin/python
+
+print("python")

test/shebangs/python_with_encoding_equals.py

@@ -0,0 +1,5 @@
+#!/usr/bin/env sbang
+# encoding=utf-8
+#!/usr/bin/python
+
+print("python")

test/shebangs/python_with_encoding_minimal.py

@@ -0,0 +1,5 @@
+#!/usr/bin/env sbang
+# coding: utf-8
+#!/usr/bin/python
+
+print("python")

test/shebangs/python_with_encoding_vim.py

@@ -0,0 +1,5 @@
+#!/usr/bin/env sbang
+# vim: set fileencoding=utf-8
+#!/usr/bin/python
+
+print("python")

test/shebangs/ruby_with_encoding.rb

@@ -0,0 +1,5 @@
+#!/usr/bin/env sbang
+# encoding: utf-8
+#!/usr/bin/ruby
+
+puts "ruby"

test/test-suite.sh

@@ -33,12 +33,17 @@ equals "/usr/bin/env perl -w -x shebangs/perl-w-env.pl" $SBANG shebangs/perl-w-e
 
 equals "/usr/bin/ruby -x shebangs/ruby.rb"              $SBANG shebangs/ruby.rb
 equals "/usr/bin/env ruby -x shebangs/ruby-env.rb"      $SBANG shebangs/ruby-env.rb
+equals "/usr/bin/ruby -x shebangs/ruby_with_encoding.rb"          $SBANG shebangs/ruby_with_encoding.rb
 
 equals "/usr/bin/perl5.32.0 -x shebangs/perl-ver.pl"    $SBANG shebangs/perl-ver.pl
 equals "/usr/bin/ruby2.7 -x shebangs/ruby-ver.rb"       $SBANG shebangs/ruby-ver.rb
 
 equals "/usr/bin/python shebangs/python.py"             $SBANG shebangs/python.py
 equals "/usr/bin/env python shebangs/python-env.py"     $SBANG shebangs/python-env.py
+equals "/usr/bin/python shebangs/python_with_encoding_emacs.py"   $SBANG shebangs/python_with_encoding_emacs.py
+equals "/usr/bin/python shebangs/python_with_encoding_minimal.py" $SBANG shebangs/python_with_encoding_minimal.py
+equals "/usr/bin/python shebangs/python_with_encoding_equals.py"  $SBANG shebangs/python_with_encoding_equals.py
+equals "/usr/bin/python shebangs/python_with_encoding_vim.py"     $SBANG shebangs/python_with_encoding_vim.py
 
 equals "/bin/sh shebangs/sh.sh"                         $SBANG shebangs/sh.sh
 equals "/bin/bash shebangs/bash.bash"                   $SBANG shebangs/bash.bash