zaru 0.0.3 → 0.1.0

data/lib/zaru.rb

@@ -3,12 +3,14 @@
 class Zaru
   CHARACTER_FILTER = /[\x00-\x1F\/\\:\*\?\"<>\|]/u
   UNICODE_WHITESPACE = /[[:space:]]+/u
-  WINDOWS_RESERVED_NAMES = 
+  WINDOWS_RESERVED_NAMES =
     %w{CON PRN AUX NUL COM1 COM2 COM3 COM4 COM5
-       COM6 COM7 COM8 COM9 LPT1 LPT2 LPT3 LPT4 
+       COM6 COM7 COM8 COM9 LPT1 LPT2 LPT3 LPT4
        LPT5 LPT6 LPT7 LPT8 LPT9}
+  FALLBACK_FILENAME = 'file'
 
-  def initialize(filename)
+  def initialize(filename, options = {})
+    @padding = options[:padding] || 0
     @raw = filename.to_s.freeze
   end
 
@@ -18,28 +20,32 @@ class Zaru
     @normalized ||= @raw.strip.gsub(UNICODE_WHITESPACE,' ')
   end
 
-  # remove characters that aren't allowed cross-OS
+  # remove bad things!
+  # - remove characters that aren't allowed cross-OS
+  # - don't allow certain special filenames (issue on Windows)
+  # - don't allow filenames to start with a dot
+  # - don't allow empty filenames
   def sanitize
-    @sanitized ||= 
+    @sanitized ||=
       filter(normalize.gsub(CHARACTER_FILTER,''))
   end
 
-  # normalize unicode string and cut off at 255 characters
-  # TODO
+  # cut off at 255 characters
+  # optionally provide a padding, which is useful to
+  # make sure there is room to add a file extension later
   def truncate
-    @truncated ||= sanitize.chars.to_a.slice(0..254).join
+    @truncated ||= sanitize.chars.to_a.slice(0..254-@padding).join
   end
 
-  # convert back from multibyte string
   def to_s
     truncate
   end
-  
+
   # convenience method
-  def self.sanitize!(filename)
-    new(filename).to_s
+  def self.sanitize!(filename, options = {})
+    new(filename, options).to_s
   end
-  
+
   private
 
     def filter(filename)
@@ -53,11 +59,11 @@ class Zaru
     end
 
     def filter_blank(filename)
-      filename == '' ? 'file': filename 
+      filename.empty?? FALLBACK_FILENAME : filename
     end
 
     def filter_dot(filename)
-      filename.start_with?('.') ? "file#{filename}" : filename
+      filename.start_with?('.')? "#{FALLBACK_FILENAME}#{filename}" : filename
     end
-  
+
 end

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: zaru
 version: !ruby/object:Gem::Version 
-  hash: 25
+  hash: 27
   prerelease: 
   segments: 
   - 0
+  - 1
   - 0
-  - 3
-  version: 0.0.3
+  version: 0.1.0
 platform: ruby
 authors: 
 - Thomas Fuchs
@@ -15,7 +15,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2013-02-05 00:00:00 Z
+date: 2014-04-26 00:00:00 Z
 dependencies: []
 
 description: Zaru takes a given filename (a string) and normalizes, filters and truncates it, so it can be safely used as a filename in modern operating systems. Zaru doesn't remove Unicode characters when not necessary.