Constructor Detail

• MultipartParser

  public MultipartParser(byte[] arr,
                         String delim)

  Constructor

Method Detail

• setDefaultCharset

  public void setDefaultCharset(String charset)

  Decodes parts in a named character set when the request does not name one itself.

  Opt-in, and it changes nothing for any caller that does not call it. Multipart is in live use by customer applications doing file upload, so the default stays what it has always been: the definition_charset part if the request carries one, and ISO-8859-1 otherwise. Altering that globally would change behaviour for deployed customers who did not ask for it.

  An in-band definition_charset still wins. This is a fallback for the request that says nothing, not an override of the request that does - including the Internet Explorer workaround, which distrusts the BROWSER's claim of UTF-8 and must not be undone by an application's own declaration.

  Both HTML-style and Java-style names are accepted ("UTF-8" and "UTF8", "iso-8859-2" and "8859_2"). The name is validated HERE rather than at decode time, and the reason is sharper than tidiness: MultipartToken.getDataString catches the decode failure and returns null, and getFormData then calls Hashtable.put with that null - which throws NullPointerException. So an unvalidated charset does not degrade the result, it fails the whole request from inside the parser, with a stack trace pointing at Hashtable.put rather than at whoever named the charset.

  Parameters:

  charset - The character set name, or null to keep the default.

  Throws:

  IllegalArgumentException - If the name is not a supported character set.

• parseAllTokens

  public int parseAllTokens()

  Parse all tokens into Vector of MultipartToken objects.

• getFormData

  public Hashtable<String,String> getFormData()

  Get form data hastable, using correct charset.