Changeset 18379 for gsdl/trunk/perllib

Timestamp:

2009-01-14T13:47:04+13:00 (15 years ago)

Author:

ak19

Message:

Base64 encoding subroutine now replaces the plus and forward slash symbols that it may generate with the minus and underscore chars, respectively, in order to ensure that it doesn't generate unsafe filenames. Base64 decoding subroutine has been correspondingly modified.

File:

• gsdl/trunk/perllib/unicode.pm (modified) (3 diffs)

gsdl/trunk/perllib/unicode.pm

@@ -607 +607 @@
 sub url_to_filename {
     my ($text) =@_;
-    $text =~ s/%25/%/g if is_url_encoded($text);
+    $text =~ s/%25/%/g if &is_url_encoded($text);
     return $text;
 }
@@ -632 +632 @@
 sub base64_encode {
     my ($text) = @_;
-    if(!is_safe_filename_chars($text)) {  #if(!&is_base64_or_alphanumeric($text)) { # a subset
+    if(!&conforms_to_mod_base64($text)) {
     $text = &MIME::Base64::encode_base64($text);
+    # base64 encoding may introduce + and / signs,
+    # replacing them with - and _ to ensure it's filename-safe
+    $text =~ s/\+/\-/g; # + -> -
+    $text =~ s/\//\_/g; # / -> _
     }
     return $text;
 }
 
+# If the input fits the modified base64 pattern, this will try decoding it. 
+# Still, this method does not guarantee the return value is the 'original', only
+# that the result is where the base64 decoding process has been applied once.
+# THIS METHOD IS NOT USED at the moment. It's here for convenience and symmetry.
 sub base64_decode {
     my ($text) = @_;
-    # If the input fits the base64 pattern, this will try decoding it.
-    # Still, this does not guarantee the return value is the 'original'
-    if(&is_base64_or_alphanumeric($text)) {
+    if(&conforms_to_mod_base64($text)) {
+    # base64 encodes certain chars with + and /, but if we'd encoded it, we'd
+    # have replaced them with - and _ respectively. Undo this before decoding.
+    $text =~ s/\-/\+/g;      # - -> +
+    $text =~ s/\_/\//g;      # _ -> /
     $text = &MIME::Base64::decode_base64($text);
     }
@@ -648 +658 @@
 }
 
-# Returns through if the given string is compatible with base64 (which 
-# includes regular ASCII alphanumeric values). This method does not 
-# guarantee that base64_decoding will return anything meanigful, since 
-# this will return true for any simple alphanumeric ASCII string as well.
-sub is_base64_or_alphanumeric {
+# Returns true if the given string is compatible with a modified version
+# of base64 (where the + and / are replaced with - and _), a format which 
+# includes also regular ASCII alphanumeric values. This method does not 
+# guarantee that the given string is actually base64 encoded, since it will
+# return true for any simple alphanumeric ASCII string as well. 
+sub conforms_to_mod_base64 {
     my ($text) = @_;
-    # base 64 takes alphanumeric and [+/=] 
-    # (note that in some cases - and _ replace + and /)
-    return ($text =~ m/^[A-Za-z0-9\+\/\=]+$/);
-}
-
-# Returns true if the text is base64 chars or additionally contains - and/or _
-sub is_safe_filename_chars {
-    my ($text) = @_;
-    return ($text =~ m/^[A-Za-z0-9\+\/\=\-\_]+$/); #alphanumeric,[+/=-_]
+    # base 64 takes alphanumeric and [=+/], 
+    # but we use modified base64 where + and / are replaced with  - and _
+    return ($text =~ m/^[A-Za-z0-9\=\-\_]+$/); #alphanumeric and [=-_]
 }