]> Adobe

masinter@adobe.com http://larry.masinter.net

Applications APPSAWG This specification (re)defines the multipart/form-data Internet Media Type, which can be used by a wide variety of applications and transported by a wide variety of protocols as a way of returning a set of values as the result of a user filling out a form. It obsoletes RFC 2388.

In many applications, it is possible for a user to be presented with a form. The user will fill out the form, including information that is typed, generated by user input, or included from files that the user has selected. When the form is filled out, the data from the form is sent from the user to the receiving application. The definition of multipart/form-data is derived from one of those applications, originally set out in and subsequently incorporated into HTML 3.2, where forms are expressed in HTML, and in which the form data is sent via HTTP or electronic mail. This representation is widely implemented in numerous web browsers and web servers. However, multipart/form-data is also used for forms that are presented using representations other than HTML (spreadsheets, PDF, etc.), and for transport using means other than electronic mail or HTTP; it is used in distributed applications which do not involve forms at all, or do not have users filling out the form. For this reason, this document defines a general syntax and semantics independent of the application for which it is used, with specific rules for web applications noted in context. The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14, RFC 2119 .

Within this specification, "percent-encoding" (as defined in ) is offered as a possible way of encoding characters in file names that are otherwise disallowed, including non-ASCII characters, spaces, control characters and so forth. The encoding is created replacing each non-ASCII or disallowed character with a sequence, where each byte of the UTF-8 encoding of the character is represented by a percent-sign (%) followed by the (case-insensitive) hexadecimal of that byte.

The representation and interpretation of forms and the nature of form processing is not specified by this document. However, for forms and form-processing that result in generation of multipart/form-data, some suggestions are included. In a form, there is generally a sequence of fields, where each field is expected to be supplied with a value, e.g. by a user who fills out the form. Each field has a name. After a form has been filled out, and the form's data is "submitted": the form processing results in a set of values for each field-- the "form data". In forms that work with multipart/form-data, field names could be arbitrary Unicode strings; however, restricting field names to ASCII will help avoid some interoperability issues (see ). Within a given form, ensuring field names are unique is also helpful. Some fields may have default values or presupplied values in the form itself. Fields with presupplied values might be hidden or invisible; this allows using generic processing for form data from a variety of actual forms.

The media-type multipart/form-data follows the model of multipart MIME data streams as specified in Section 5.1; changes are noted in this document. A multipart/form-data body contains a series of parts, separated by a boundary.

As with other multipart types, the parts are delimited with a boundary delimiter, constructed using CRLF, "--", the value of the boundary parameter. The boundary is supplied as a "boundary" parameter to the multipart/form-data type. As noted in Section 5.1, the boundary delimiter MUST NOT appear inside any of the encapsulated parts, and it is often necessary to enclose the boundary parameter values in quotes on the Content-type line.

Each part MUST contain a content-disposition header and where the disposition type is form-data. The content-disposition header MUST also contain an additional parameter of name; the value of the name parameter is the original field name from the form (possibly encoded; see ). For example, a part might contain a header:

Content-Disposition: form-data; name="user"

with the body of the part containing the form data of the user field.

For form data that represents the content of a file, a name for the file SHOULD be supplied as well, by using a filename parameter of the content-disposition header. The file name isn't mandatory for cases where the file name isn't available or is meaningless or private; this might result, for example, from selection or drag-and-drop or where the form data content is streamed directly from a device. If a filename parameter is supplied, the requirements of Section 2.3 for "receiving MUA" apply to recievers of multipart/form-data as well: Do not use the file name blindly, check and possibly change to match local filesystem conventions if applicable, do not use directory path information that may seems to be present. In most multipart types, the MIME headers in each part are restricted to US-ASCII; for compatibility with those systems, file names normally visible to users MAY be encoded using the percent-encoding method in , following how a "file:" URI might be encoded. NOTE: The encoding method described in , which would add a "filename*" paramter to the "Content-Disposition" header, MUST NOT be used. Some commonly deployed systems use multipart/form-data with file names directly encoded including octets outside the US-ASCII range. The encoding used for the file names is typically UTF-8, although HTML forms will use the charset associated with the form.

The form data for a form field might include multiple files. suggested that multiple files for a single form field be transmitted using a nested multipart/mixed part. This usage is deprecated. To match widely deployed implementations, multiple files MUST be sent by supplying each file in a separate part, but all with the same name parameter. Receiving applications intended for wide applicability (e.g. multipart/form-data parsing libraries) SHOULD also support the older method of supplying multiple files.

Each part MAY have an (optional) content-type, which defaults to text/plain. If the contents of a file are to be sent, the file data SHOULD be labeled with an appropriate media type, if known, or application/octet-stream.

In the case where the form data is text, the charset parameter for the text/plain Content-Type MAY be used to indicate the character encoding used in that part. For example, a form with a text field in which a user typed "Joe owes <eu>100" where <eu> is the Euro symbol might have form data returned as:

--AaB03x content-disposition: form-data; name="field1" content-type: text/plain;charset=UTF-8 content-transfer-encoding: quoted-printable Joe owes =E2=82=AC100. --AaB03x

In practice, many widely deployed implementations do not supply a charset parameter in each part, but, rather, they rely on the notion of a "default charset" for a multipart/form-data instance. Subsequent sections will explain how the default charset is established.

Some form processing applications (including HTML) have the convention that the value of a form entry with entry name _charset_ and type hidden is automatically set when the form is opened; the value is used as the default charset of text field values (see form-charset in ). In such cases, the value of the default charset for each text/plain part without a charset parameter is the supplied value. For example:

--AaB03x content-disposition: form-data; name="_charset_" iso-8859-1 --AaB03x-- content-disposition: form-data; name="field1" ...text encoded in iso-8859-1 ... AaB03x--

Previously, it was recommended that senders use a Content-Transfer-Encoding encoding (such as quoted-printable) for each non-ASCII part of a multipart/form-data body, because that would allow use in transports that only support a 7BIT encoding. This use is deprecated for use in contexts that support binary data such as HTTP. Senders SHOULD NOT generate any parts with a Content-Transfer-Encoding header. Currently, no deployed implementations that send such bodies have been discovered.

The multipart/form-data media type does not support any MIME headers in the parts other than Content-Type, Content-Disposition, and (in limited circumstances) Content-Transfer-Encoding. Other headers MUST NOT be included and MUST be ignored.

Normally, MIME headers in multipart bodies are required to consist only of 7-bit data in the US-ASCII character set. While suggested that non-ASCII field names be encoded according to the method in , this practice doesn't seem to have been followed widely. This specification makes three sets of recommendations for three different states of workflow.

For broadest interoperability with existing deployed software, those creating forms SHOULD avoid non-ASCII field names. This should not be a burden, because in general the field names are not visible to users. The field names in the underlying need not match what the user sees on the screen. If non-ASCII field names are unavoidable, form or application creators SHOULD use UTF-8 uniformly. This will minimize interoperability problems.

Some applications of this specification will supply a character encoding to be used for interpretation of the multipart/form-data body. In particular, HTML 5 uses: The content of a '_charset_' field, if there is one. the value of an accept-charset attribute of the <form> element, if there is one, the character encoding of the document containing the form, if it is US-ASCII compatible, otherwise UTF-8. Call this value the form-charset. Any text, whether field name, field value, or (text/plain) form data which is uses characters outside the ASCII range MAY be represented directly encoded in the form-charset.

While this specification provides guidance for creation of multipart/form-data, parsers and interpreters should be aware of the variety of implementations. File systems differ as to whether and how they normalize Unicode names, for example. The matching of form elements to form-data parts may rely on a fuzzier match. In particular, some multipart/form-data generators might have followed the previous advice of and used the "encoded-word" method of encoding non-ASCII values:

encoded-word = "=?" charset "?" encoding "?" encoded-text "?="

Others have been known to follow , to send unencoded UTF-8, or even strings encoded in the form-charset. For this reason, interpreting multipart/form-data (even from conforming generators) may require knowing the charset used in form encoding, in cases where the _charset_ field value or a charset parameter of a text/plain Content-Type header is not supplied.

Form processors given forms with a well-defined ordering SHOULD send back results in order (note that there are some forms which do not define a natural order.) Intermediaries MUST NOT reorder the results. Form parts with identical field names MUST NOT be coalesced.

Many web applications use the "application/x-url-encoded" method for returning data from forms. This format is quite compact, e.g.:

name=Xavier+Xantico&verdict=Yes&colour=Blue&happy=sad&Utf%F6r=Send

However, there is no opportunity to label the enclosed data with content type, apply a charset, or use other encoding mechanisms. Many form-interpreting programs (primarily web browsers) now implement and generate multipart/form-data, but an existing application might need to optionally support both the application/x-url-encoded format as well.

This specification provides no specific mechanism by which multipart/form-data can be associated with the form that caused it to be transmitted. This separation is intentional; many different forms might be used for transmitting the same data. In practice, applications may supply a specific form processing resource (in HTML, the ACTION attribute in a FORM tag) for each different form. Alternatively, data about the form might be encoded in a "hidden field" (a field which is part of the form but which has a fixed value to be transmitted back to the form-data processor.)

Please update the Internet Media Type registration of multipart/form-data to point to this document, using the template in . In addition, please update the registrations of the "name" and "form-data" parameters in the "Content Disposition Parameters" registry to both point to this document.

Applications which receive forms and process them must be careful not to supply data back to the requesting form processing site that was not intended to be sent. It is important when interpreting the filename of the Content-Disposition header to not overwrite files in the recipient's file space inadvertently. User applications that request form information from users must be careful not to cause a user to send information to the requestor or a third party unwillingly or unwittingly. For example, a form might request 'spam' information to be sent to an unintended third party, or private information to be sent to someone that the user might not actually intend. While this is primarily an issue for the representation and interpretation of forms themselves (rather than the data representation of the form data), the transportation of private information must be done in a way that does not expose it to unwanted prying. With the introduction of form-data that can reasonably send back the content of files from a user's file space, the possibility arises that a user might be sent an automated script that fills out a form and then sends one of the user's local files to another address. Thus, additional caution is required when executing automated scripting where form-data might include a user's files. Files sent via multipart/form-data may contain arbitrary executable content, and precautions against malicious content are necessary. The considerations of Sections 2.3 and 5 with respect to the filename parameter of the Content-Disposition header also apply to its usage here. All form processing software should treat user supplied form-data with sensitivity, as it often contains confidential or personally identifying information. There is widespread use of form "auto-fill" features in web browsers; these might be used to trick users to unknowingly send confidential information when completing otherwise innoccuous tasks. Multipart/form-data does not supply any features for checking integrity, ensuring confidentiality, avoiding user confusion, or other security features; those concerns must be addressed by the form-filling and form-data-interpreting applications.

This section is the media type registration. multipart form-data boundary none Common use is BINARY. In limited use (or transports that restrict the encoding to 7BIT or 8BIT each part is encoded separately using Content-Transfer-Encoding . See of this document. This document makes several recommendations for interoperability with deployed implementations, including . This document. Numerous web browsers, servers, and web applications. None: Fragment identifiers are not defined for this type. None: no deprecated alias names, magic numbers, file extensions or Macintosh ssssfile type codes. Author of this document. COMMON none Author of this document. IETF N/A

&rfc.2119; &rfc.2046; &rfc.2047; &rfc.2231; &rfc.2183; &rfc.3986; &rfc.1867; &rfc.2388; &rfc.5987; &rfc.6838; &I-D.file; &w3c.html5; &w3c.html32;

The handling of non-ASCII field names changed-- no longer recommending the RFC 2047 method, instead suggesting senders send UTF-8 field names directly, and file names directly in the form-charset. The handling of multiple files submitted as the result of a single form field (e.g. HTML's <input type=file multiple> element) results in each file having its own top level part with the same name parameter; the method of using a nested multipart/mixed from is no longer recommended for creators, and not required for receivers as there are no known implementations of senders. The _charset_ convention and use of an explicit form-data charset is documented. 'boundary' is a required parameter in Content-Type. The relationship of the ordering of fields within a form and the ordering of returned values within multipart/form-data was not defined before, nor was the handling of the case where a form has multiple fields with the same name. Editorial: Removed obsolete discussion of alternatives in appendix. Update references. Move outline of form processing into Introduction.

There are numerous alternative ways in which form data can be encoded; many are listed in section 5.2. The multipart/form-data encoding is verbose, especially if there are many fields with short values. In most use cases, this overhead isn't significant. More problematic is the ambiguity introduced because implementations did not follow because it used "may" instead of "MUST" when specifying encoding of field names, and for other unknown reasons, so now, parsers need to be more complex for fuzzy matching against the possible outputs of various encoding methods.