org.microproxy.http.HttpBodies

public final class HttpBodies extends Object

Reads and rewrites the bodies of buffered messages (FullHttpMessage) in filters, taking care of Content-Encoding and charsets so filters don't have to.


 if (o instanceof FullHttpResponse res && HttpBodies.isText(res) && HttpBodies.canDecode(res)) {
     String html = HttpBodies.text(res);
     HttpBodies.setText(res, html.replace("</body>", "<script src=/x.js></script></body>"));
 }

gzip, x-gzip, deflate (zlib-wrapped or raw) and br (with a bundled copy of the pure-Java Brotli codec jaybrotli) can be decoded, and so can zstd when the optional zstd-decoder module is on the class path. Others, such as dictionary codings (dcb, dcz), are reported by canDecode(org.microproxy.http.HttpMessage), and restrictAcceptEncoding(org.microproxy.http.HttpRequest) keeps servers from choosing them. Decoding is capped (64 MiB by default) to defuse compression bombs. Rewriting a body re-applies the original coding (Brotli at quality 5; zstd bodies are re-encoded as gzip, since there is no pure-Java zstd encoder) and drops validators (ETag, Content-MD5) that no longer match; Content-Length is fixed when the message is written.

Field Summary

Fields

Modifier and Type

Field

Description

static final int

BROTLI_QUALITY

The Brotli quality (0 to 11) that encode(java.lang.String, byte[]) and rewritten br bodies use, with jaybrotli's default strategies: on HTML and JSON of 10 KB to 1 MB it runs at about 10 to 35 MB/s, and its output is mostly 1 to 14% smaller than gzip's default level (at worst 2% larger).

static final Set<String>

DECODABLE

Content codings this class can decode: gzip, x-gzip, deflate and br, plus zstd when the zstd-decoder module is on the class path.

static final int

DEFAULT_MAX_DECODED_BYTES

Default cap on decoded body size.
Method Summary

Modifier and Type

Method

Description

static boolean

canDecode(HttpMessage message)

Whether every content coding of message can be decoded.

static Charset

charset(HttpMessage message, Charset fallback)

The charset parameter of Content-Type, or fallback when absent, unknown or malformed.

static List<String>

contentEncodings(HttpMessage message)

The content codings applied to message, outermost last (identity omitted).

static byte[]

decode(String coding, byte[] data, int max)

Removes one content coding from data, as decoded(FullHttpMessage, int) does for each coding of a message.

static byte[]

decoded(FullHttpMessage message)

The body with all content codings removed.

static byte[]

decoded(FullHttpMessage message, int maxDecodedBytes)

Removes all content codings with a limit on each decompression stage.

static byte[]

decodedPrefix(FullHttpMessage message, int maxBytes)

The first maxBytes of the body with all content codings removed, for previews and logs: unlike decoded(FullHttpMessage, int), a longer decoded body is cut short rather than an error, and only that much is ever decoded.

static byte[]

encode(String coding, byte[] data)

Applies one content coding to data: gzip (or x-gzip), deflate (zlib-wrapped) or br (at BROTLI_QUALITY), the codings this class can produce.

static boolean

isText(HttpMessage message)

Whether the media type is textual: text/*, JSON, XML, JavaScript or form data.

static String

mediaType(HttpMessage message)

The media type of Content-Type in lower case, without parameters, or "".

static void

removeContentEncoding(FullHttpMessage message)

Decodes the body and removes Content-Encoding, so later filters see plain bytes.

static void

restrictAcceptEncoding(HttpRequest request)

Removes content codings this class cannot decode (e.g.

static void

setDecoded(FullHttpMessage message, byte[] decodedBody)

Replaces the body with decodedBody, re-applying the message's content codings, and removes validators that described the old body.

static void

setText(FullHttpMessage message, String text)

Replaces the body with text in the message's charset (UTF-8 when none is declared).

static String

text(FullHttpMessage message)

The decoded body as text, in the message's charset (UTF-8 when none is declared).

Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Details
- DEFAULT_MAX_DECODED_BYTES
  
  public static final int DEFAULT_MAX_DECODED_BYTES
  
  Default cap on decoded body size.
  See Also:
  
  Constant Field Values
- DECODABLE
  
  public static final Set<String> DECODABLE
  
  Content codings this class can decode: gzip, x-gzip, deflate and br, plus zstd when the zstd-decoder module is on the class path.
- BROTLI_QUALITY
  
  public static final int BROTLI_QUALITY
  
  The Brotli quality (0 to 11) that encode(java.lang.String, byte[]) and rewritten br bodies use, with jaybrotli's default strategies: on HTML and JSON of 10 KB to 1 MB it runs at about 10 to 35 MB/s, and its output is mostly 1 to 14% smaller than gzip's default level (at worst 2% larger). Qualities 4 and 6 come out within 1% of it; the higher qualities and the alternative strategies trade too much speed for what they save.
  See Also:
  
  Constant Field Values
Method Details
- contentEncodings
  
  public static List<String> contentEncodings(HttpMessage message)
  
  The content codings applied to message, outermost last (identity omitted).
  
  Parameters:
  
  message - the message whose body and headers are inspected
  
  Returns:
  
  the lower-case content codings in application order, excluding identity
- canDecode
  
  public static boolean canDecode(HttpMessage message)
  
  Whether every content coding of message can be decoded.
  
  Parameters:
  
  message - the message whose body and headers are inspected
  
  Returns:
  
  whether all declared content codings are supported
- decoded
  
  public static byte[] decoded(FullHttpMessage message) throws IOException
  
  The body with all content codings removed.
  
  Parameters:
  
  message - the message whose body and headers are inspected
  
  Returns:
  
  the body bytes after all content codings are removed
  
  Throws:
  
  IOException - if a coding is unsupported, corrupt, or exceeds the default decoding limit
- decoded
  
  public static byte[] decoded(FullHttpMessage message, int maxDecodedBytes) throws IOException
  
  Removes all content codings with a limit on each decompression stage.
  
  Parameters:
  
  message - the message whose body and headers are inspected
  
  maxDecodedBytes - the maximum output size in bytes for each decompression stage
  
  Returns:
  
  the body bytes after all content codings are removed
  
  Throws:
  
  IOException - if a coding is unsupported, corrupt, or exceeds the decoding limit
- decodedPrefix
  
  public static byte[] decodedPrefix(FullHttpMessage message, int maxBytes) throws IOException
  
  The first maxBytes of the body with all content codings removed, for previews and logs: unlike decoded(FullHttpMessage, int), a longer decoded body is cut short rather than an error, and only that much is ever decoded.
  
  Parameters:
  
  message - the message whose body and headers are inspected
  
  maxBytes - the maximum prefix length; nonpositive values produce an empty result
  
  Returns:
  
  up to maxBytes decoded body bytes
  
  Throws:
  
  IOException - if a coding is unsupported or the decoded prefix cannot be read
- setDecoded
  
  public static void setDecoded(FullHttpMessage message, byte[] decodedBody) throws IOException
  
  Replaces the body with decodedBody, re-applying the message's content codings, and removes validators that described the old body.
  
  Parameters:
  
  message - the message to update
  
  decodedBody - the replacement bytes before applying content codings
  
  Throws:
  
  IOException - if a content coding is unsupported or encoding fails
- removeContentEncoding
  
  public static void removeContentEncoding(FullHttpMessage message) throws IOException
  
  Decodes the body and removes Content-Encoding, so later filters see plain bytes.
  
  Parameters:
  
  message - the message to decode and update
  
  Throws:
  
  IOException - if a coding is unsupported, corrupt, or exceeds the default decoding limit
- restrictAcceptEncoding
  
  public static void restrictAcceptEncoding(HttpRequest request)
  
  Removes content codings this class cannot decode (e.g. dcb) from the request's Accept-Encoding, keeping quality values, so the server picks one that filters can read. Leaves identity when nothing else remains.
  
  Parameters:
  
  request - the request whose accepted codings are restricted
- text
  
  public static String text(FullHttpMessage message) throws IOException
  
  The decoded body as text, in the message's charset (UTF-8 when none is declared).
  
  Parameters:
  
  message - the message whose body and headers are inspected
  
  Returns:
  
  the decoded body interpreted in the declared charset, or UTF-8 by default
  
  Throws:
  
  IOException - if a coding is unsupported, corrupt, or exceeds the default decoding limit
- setText
  
  public static void setText(FullHttpMessage message, String text) throws IOException
  
  Replaces the body with text in the message's charset (UTF-8 when none is declared).
  
  Parameters:
  
  message - the message to update
  
  text - the replacement text before charset encoding and content coding
  
  Throws:
  
  IOException - if a content coding is unsupported or encoding fails
- charset
  
  public static Charset charset(HttpMessage message, Charset fallback)
  
  The charset parameter of Content-Type, or fallback when absent, unknown or malformed.
  
  Parameters:
  
  message - the message whose Content-Type is inspected
  
  fallback - the charset to use when the parameter cannot be resolved
  
  Returns:
  
  the declared charset, or fallback if unavailable
- mediaType
  
  public static String mediaType(HttpMessage message)
  
  The media type of Content-Type in lower case, without parameters, or "".
  
  Parameters:
  
  message - the message whose Content-Type is inspected
  
  Returns:
  
  the lower-case media type without parameters, or an empty string if absent
- isText
  
  public static boolean isText(HttpMessage message)
  
  Whether the media type is textual: text/*, JSON, XML, JavaScript or form data.
  
  Parameters:
  
  message - the message whose media type is inspected
  
  Returns:
  
  whether the media type is recognized as textual
- decode
  
  public static byte[] decode(String coding, byte[] data, int max) throws IOException
  
  Removes one content coding from data, as decoded(FullHttpMessage, int) does for each coding of a message. Also used for gRPC messages, which are compressed one by one with the coding their grpc-encoding names.
  
  Parameters:
  
  coding - the coding, in lower case: gzip, x-gzip, deflate (zlib-wrapped or raw), br, or zstd when the module is present
  
  data - the encoded bytes
  
  max - the maximum decoded size in bytes
  
  Returns:
  
  the decoded bytes
  
  Throws:
  
  IOException - if the coding is unsupported, the data corrupt, or the result larger than max
- encode
  
  public static byte[] encode(String coding, byte[] data) throws IOException
  
  Applies one content coding to data: gzip (or x-gzip), deflate (zlib-wrapped) or br (at BROTLI_QUALITY), the codings this class can produce.
  
  Parameters:
  
  coding - the coding, in lower case
  
  data - the bytes to encode
  
  Returns:
  
  the encoded bytes
  
  Throws:
  
  IOException - if the coding is not one of those

Class HttpBodies

Field Summary

Method Summary

Methods inherited from class java.lang.Object

Field Details

DEFAULT_MAX_DECODED_BYTES

DECODABLE

BROTLI_QUALITY

Method Details

contentEncodings

canDecode

decoded

decoded

decodedPrefix

setDecoded

removeContentEncoding

restrictAcceptEncoding

text

setText

charset

mediaType

isText

decode

encode