|
|
|
// Copyright Joyent, Inc. and other Node contributors.
|
|
|
|
//
|
|
|
|
// Permission is hereby granted, free of charge, to any person obtaining a
|
|
|
|
// copy of this software and associated documentation files (the
|
|
|
|
// "Software"), to deal in the Software without restriction, including
|
|
|
|
// without limitation the rights to use, copy, modify, merge, publish,
|
|
|
|
// distribute, sublicense, and/or sell copies of the Software, and to permit
|
|
|
|
// persons to whom the Software is furnished to do so, subject to the
|
|
|
|
// following conditions:
|
|
|
|
//
|
|
|
|
// The above copyright notice and this permission notice shall be included
|
|
|
|
// in all copies or substantial portions of the Software.
|
|
|
|
//
|
|
|
|
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
|
|
|
|
// OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
|
|
|
|
// MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN
|
|
|
|
// NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,
|
|
|
|
// DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
|
|
|
|
// OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
|
|
|
|
// USE OR OTHER DEALINGS IN THE SOFTWARE.
|
|
|
|
|
|
|
|
var Transform = require('_stream_transform');
|
|
|
|
|
|
|
|
var binding = process.binding('zlib');
|
|
|
|
var util = require('util');
|
|
|
|
var assert = require('assert').ok;
|
|
|
|
|
|
|
|
// zlib doesn't provide these, so kludge them in following the same
|
|
|
|
// const naming scheme zlib uses.
|
|
|
|
binding.Z_MIN_WINDOWBITS = 8;
|
|
|
|
binding.Z_MAX_WINDOWBITS = 15;
|
|
|
|
binding.Z_DEFAULT_WINDOWBITS = 15;
|
|
|
|
|
|
|
|
// fewer than 64 bytes per chunk is stupid.
|
|
|
|
// technically it could work with as few as 8, but even 64 bytes
|
|
|
|
// is absurdly low. Usually a MB or more is best.
|
|
|
|
binding.Z_MIN_CHUNK = 64;
|
|
|
|
binding.Z_MAX_CHUNK = Infinity;
|
|
|
|
binding.Z_DEFAULT_CHUNK = (16 * 1024);
|
|
|
|
|
|
|
|
binding.Z_MIN_MEMLEVEL = 1;
|
|
|
|
binding.Z_MAX_MEMLEVEL = 9;
|
|
|
|
binding.Z_DEFAULT_MEMLEVEL = 8;
|
|
|
|
|
|
|
|
binding.Z_MIN_LEVEL = -1;
|
|
|
|
binding.Z_MAX_LEVEL = 9;
|
|
|
|
binding.Z_DEFAULT_LEVEL = binding.Z_DEFAULT_COMPRESSION;
|
|
|
|
|
|
|
|
// expose all the zlib constants
|
|
|
|
Object.keys(binding).forEach(function(k) {
|
|
|
|
if (k.match(/^Z/)) exports[k] = binding[k];
|
|
|
|
});
|
|
|
|
|
|
|
|
// translation table for return codes.
|
|
|
|
exports.codes = {
|
|
|
|
Z_OK: binding.Z_OK,
|
|
|
|
Z_STREAM_END: binding.Z_STREAM_END,
|
|
|
|
Z_NEED_DICT: binding.Z_NEED_DICT,
|
|
|
|
Z_ERRNO: binding.Z_ERRNO,
|
|
|
|
Z_STREAM_ERROR: binding.Z_STREAM_ERROR,
|
|
|
|
Z_DATA_ERROR: binding.Z_DATA_ERROR,
|
|
|
|
Z_MEM_ERROR: binding.Z_MEM_ERROR,
|
|
|
|
Z_BUF_ERROR: binding.Z_BUF_ERROR,
|
|
|
|
Z_VERSION_ERROR: binding.Z_VERSION_ERROR
|
|
|
|
};
|
|
|
|
|
|
|
|
Object.keys(exports.codes).forEach(function(k) {
|
|
|
|
exports.codes[exports.codes[k]] = k;
|
|
|
|
});
|
|
|
|
|
|
|
|
exports.Deflate = Deflate;
|
|
|
|
exports.Inflate = Inflate;
|
|
|
|
exports.Gzip = Gzip;
|
|
|
|
exports.Gunzip = Gunzip;
|
|
|
|
exports.DeflateRaw = DeflateRaw;
|
|
|
|
exports.InflateRaw = InflateRaw;
|
|
|
|
exports.Unzip = Unzip;
|
|
|
|
|
|
|
|
exports.createDeflate = function(o) {
|
|
|
|
return new Deflate(o);
|
|
|
|
};
|
|
|
|
|
|
|
|
exports.createInflate = function(o) {
|
|
|
|
return new Inflate(o);
|
|
|
|
};
|
|
|
|
|
|
|
|
exports.createDeflateRaw = function(o) {
|
|
|
|
return new DeflateRaw(o);
|
|
|
|
};
|
|
|
|
|
|
|
|
exports.createInflateRaw = function(o) {
|
|
|
|
return new InflateRaw(o);
|
|
|
|
};
|
|
|
|
|
|
|
|
exports.createGzip = function(o) {
|
|
|
|
return new Gzip(o);
|
|
|
|
};
|
|
|
|
|
|
|
|
exports.createGunzip = function(o) {
|
|
|
|
return new Gunzip(o);
|
|
|
|
};
|
|
|
|
|
|
|
|
exports.createUnzip = function(o) {
|
|
|
|
return new Unzip(o);
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
// Convenience methods.
|
|
|
|
// compress/decompress a string or buffer in one step.
|
|
|
|
exports.deflate = function(buffer, callback) {
|
|
|
|
zlibBuffer(new Deflate(), buffer, callback);
|
|
|
|
};
|
|
|
|
|
|
|
|
exports.gzip = function(buffer, callback) {
|
|
|
|
zlibBuffer(new Gzip(), buffer, callback);
|
|
|
|
};
|
|
|
|
|
|
|
|
exports.deflateRaw = function(buffer, callback) {
|
|
|
|
zlibBuffer(new DeflateRaw(), buffer, callback);
|
|
|
|
};
|
|
|
|
|
|
|
|
exports.unzip = function(buffer, callback) {
|
|
|
|
zlibBuffer(new Unzip(), buffer, callback);
|
|
|
|
};
|
|
|
|
|
|
|
|
exports.inflate = function(buffer, callback) {
|
|
|
|
zlibBuffer(new Inflate(), buffer, callback);
|
|
|
|
};
|
|
|
|
|
|
|
|
exports.gunzip = function(buffer, callback) {
|
|
|
|
zlibBuffer(new Gunzip(), buffer, callback);
|
|
|
|
};
|
|
|
|
|
|
|
|
exports.inflateRaw = function(buffer, callback) {
|
|
|
|
zlibBuffer(new InflateRaw(), buffer, callback);
|
|
|
|
};
|
|
|
|
|
|
|
|
function zlibBuffer(engine, buffer, callback) {
|
|
|
|
var buffers = [];
|
|
|
|
var nread = 0;
|
|
|
|
|
|
|
|
engine.on('error', onError);
|
|
|
|
engine.on('end', onEnd);
|
|
|
|
|
|
|
|
engine.end(buffer);
|
|
|
|
flow();
|
|
|
|
|
|
|
|
function flow() {
|
|
|
|
var chunk;
|
|
|
|
while (null !== (chunk = engine.read())) {
|
|
|
|
buffers.push(chunk);
|
|
|
|
nread += chunk.length;
|
|
|
|
}
|
|
|
|
engine.once('readable', flow);
|
|
|
|
}
|
|
|
|
|
|
|
|
function onError(err) {
|
|
|
|
engine.removeListener('end', onEnd);
|
|
|
|
engine.removeListener('readable', flow);
|
|
|
|
callback(err);
|
|
|
|
}
|
|
|
|
|
|
|
|
function onEnd() {
|
zlib: reduce memory consumption, release early
In zlibBuffer(), don't wait for the garbage collector to reclaim the zlib memory
but release it manually. Reduces memory consumption by a factor of 10 or more
with some workloads.
Test case:
function f() {
require('zlib').deflate('xxx', g);
}
function g() {
setTimeout(f, 5);
}
f();
Observe RSS memory usage with and without this commit. After 10,000 iterations,
RSS stabilizes at ~35 MB with this commit. Without, RSS is over 300 MB and keeps
growing.
Cause: whenever the JS object heap hits the high-water mark, the V8 GC sweeps
it clean, then tries to grow it in order to avoid more sweeps in the near
future. Rule of thumb: the bigger the JS heap, the lazier the GC can be.
A side effect of a bigger heap is that objects now live longer. This is harmless
in general but it affects zlib context objects because those are tied to large
buffers that live outside the JS heap, on the order of 16K per context object.
Ergo, don't wait for the GC to reclaim the memory - it may take a long time.
Fixes #4172.
12 years ago
|
|
|
var buf = Buffer.concat(buffers, nread);
|
|
|
|
buffers = [];
|
|
|
|
callback(null, buf);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
// generic zlib
|
|
|
|
// minimal 2-byte header
|
|
|
|
function Deflate(opts) {
|
|
|
|
if (!(this instanceof Deflate)) return new Deflate(opts);
|
|
|
|
Zlib.call(this, opts, binding.DEFLATE);
|
|
|
|
}
|
|
|
|
|
|
|
|
function Inflate(opts) {
|
|
|
|
if (!(this instanceof Inflate)) return new Inflate(opts);
|
|
|
|
Zlib.call(this, opts, binding.INFLATE);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
// gzip - bigger header, same deflate compression
|
|
|
|
function Gzip(opts) {
|
|
|
|
if (!(this instanceof Gzip)) return new Gzip(opts);
|
|
|
|
Zlib.call(this, opts, binding.GZIP);
|
|
|
|
}
|
|
|
|
|
|
|
|
function Gunzip(opts) {
|
|
|
|
if (!(this instanceof Gunzip)) return new Gunzip(opts);
|
|
|
|
Zlib.call(this, opts, binding.GUNZIP);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
// raw - no header
|
|
|
|
function DeflateRaw(opts) {
|
|
|
|
if (!(this instanceof DeflateRaw)) return new DeflateRaw(opts);
|
|
|
|
Zlib.call(this, opts, binding.DEFLATERAW);
|
|
|
|
}
|
|
|
|
|
|
|
|
function InflateRaw(opts) {
|
|
|
|
if (!(this instanceof InflateRaw)) return new InflateRaw(opts);
|
|
|
|
Zlib.call(this, opts, binding.INFLATERAW);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
// auto-detect header.
|
|
|
|
function Unzip(opts) {
|
|
|
|
if (!(this instanceof Unzip)) return new Unzip(opts);
|
|
|
|
Zlib.call(this, opts, binding.UNZIP);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
// the Zlib class they all inherit from
|
|
|
|
// This thing manages the queue of requests, and returns
|
|
|
|
// true or false if there is anything in the queue when
|
|
|
|
// you call the .write() method.
|
|
|
|
|
|
|
|
function Zlib(opts, mode) {
|
|
|
|
this._opts = opts = opts || {};
|
|
|
|
this._chunkSize = opts.chunkSize || exports.Z_DEFAULT_CHUNK;
|
|
|
|
|
|
|
|
Transform.call(this, opts);
|
|
|
|
|
|
|
|
// means a different thing there.
|
|
|
|
this._readableState.chunkSize = null;
|
|
|
|
|
|
|
|
if (opts.chunkSize) {
|
|
|
|
if (opts.chunkSize < exports.Z_MIN_CHUNK ||
|
|
|
|
opts.chunkSize > exports.Z_MAX_CHUNK) {
|
|
|
|
throw new Error('Invalid chunk size: ' + opts.chunkSize);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (opts.windowBits) {
|
|
|
|
if (opts.windowBits < exports.Z_MIN_WINDOWBITS ||
|
|
|
|
opts.windowBits > exports.Z_MAX_WINDOWBITS) {
|
|
|
|
throw new Error('Invalid windowBits: ' + opts.windowBits);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (opts.level) {
|
|
|
|
if (opts.level < exports.Z_MIN_LEVEL ||
|
|
|
|
opts.level > exports.Z_MAX_LEVEL) {
|
|
|
|
throw new Error('Invalid compression level: ' + opts.level);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (opts.memLevel) {
|
|
|
|
if (opts.memLevel < exports.Z_MIN_MEMLEVEL ||
|
|
|
|
opts.memLevel > exports.Z_MAX_MEMLEVEL) {
|
|
|
|
throw new Error('Invalid memLevel: ' + opts.memLevel);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (opts.strategy) {
|
|
|
|
if (opts.strategy != exports.Z_FILTERED &&
|
|
|
|
opts.strategy != exports.Z_HUFFMAN_ONLY &&
|
|
|
|
opts.strategy != exports.Z_RLE &&
|
|
|
|
opts.strategy != exports.Z_FIXED &&
|
|
|
|
opts.strategy != exports.Z_DEFAULT_STRATEGY) {
|
|
|
|
throw new Error('Invalid strategy: ' + opts.strategy);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (opts.dictionary) {
|
|
|
|
if (!Buffer.isBuffer(opts.dictionary)) {
|
|
|
|
throw new Error('Invalid dictionary: it should be a Buffer instance');
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
this._binding = new binding.Zlib(mode);
|
|
|
|
|
|
|
|
var self = this;
|
|
|
|
this._hadError = false;
|
|
|
|
this._binding.onerror = function(message, errno) {
|
|
|
|
// there is no way to cleanly recover.
|
|
|
|
// continuing only obscures problems.
|
|
|
|
self._binding = null;
|
|
|
|
self._hadError = true;
|
|
|
|
|
|
|
|
var error = new Error(message);
|
|
|
|
error.errno = errno;
|
|
|
|
error.code = exports.codes[errno];
|
|
|
|
self.emit('error', error);
|
|
|
|
};
|
|
|
|
|
|
|
|
this._binding.init(opts.windowBits || exports.Z_DEFAULT_WINDOWBITS,
|
|
|
|
opts.level || exports.Z_DEFAULT_COMPRESSION,
|
|
|
|
opts.memLevel || exports.Z_DEFAULT_MEMLEVEL,
|
|
|
|
opts.strategy || exports.Z_DEFAULT_STRATEGY,
|
|
|
|
opts.dictionary);
|
|
|
|
|
|
|
|
this._buffer = new Buffer(this._chunkSize);
|
|
|
|
this._offset = 0;
|
|
|
|
this._closed = false;
|
|
|
|
|
|
|
|
this.once('end', this.close);
|
|
|
|
}
|
|
|
|
|
|
|
|
util.inherits(Zlib, Transform);
|
|
|
|
|
|
|
|
Zlib.prototype.reset = function reset() {
|
|
|
|
return this._binding.reset();
|
|
|
|
};
|
|
|
|
|
|
|
|
Zlib.prototype._flush = function(output, callback) {
|
|
|
|
var rs = this._readableState;
|
|
|
|
var self = this;
|
|
|
|
this._transform(null, output, function(er) {
|
|
|
|
if (er)
|
|
|
|
return callback(er);
|
|
|
|
|
|
|
|
// now a weird thing happens... it could be that you called flush
|
|
|
|
// but everything had already actually been consumed, but it wasn't
|
|
|
|
// enough to get over the Readable class's lowWaterMark.
|
|
|
|
// In that case, we emit 'readable' now to make sure it's consumed.
|
|
|
|
if (rs.length &&
|
|
|
|
rs.length < rs.lowWaterMark &&
|
|
|
|
!rs.ended &&
|
|
|
|
rs.needReadable)
|
|
|
|
self.emit('readable');
|
|
|
|
|
|
|
|
callback();
|
|
|
|
});
|
|
|
|
};
|
|
|
|
|
|
|
|
Zlib.prototype.flush = function(callback) {
|
|
|
|
var ws = this._writableState;
|
|
|
|
var ts = this._transformState;
|
|
|
|
|
|
|
|
if (ws.writing) {
|
|
|
|
ws.needDrain = true;
|
|
|
|
var self = this;
|
|
|
|
this.once('drain', function() {
|
|
|
|
self._flush(ts.output, callback);
|
|
|
|
});
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
this._flush(ts.output, callback || function() {});
|
|
|
|
};
|
|
|
|
|
|
|
|
Zlib.prototype.close = function(callback) {
|
|
|
|
if (callback)
|
|
|
|
process.nextTick(callback);
|
|
|
|
|
|
|
|
if (this._closed)
|
|
|
|
return;
|
|
|
|
|
|
|
|
this._closed = true;
|
|
|
|
|
|
|
|
this._binding.close();
|
|
|
|
|
|
|
|
var self = this;
|
|
|
|
process.nextTick(function() {
|
|
|
|
self.emit('close');
|
|
|
|
});
|
zlib: reduce memory consumption, release early
In zlibBuffer(), don't wait for the garbage collector to reclaim the zlib memory
but release it manually. Reduces memory consumption by a factor of 10 or more
with some workloads.
Test case:
function f() {
require('zlib').deflate('xxx', g);
}
function g() {
setTimeout(f, 5);
}
f();
Observe RSS memory usage with and without this commit. After 10,000 iterations,
RSS stabilizes at ~35 MB with this commit. Without, RSS is over 300 MB and keeps
growing.
Cause: whenever the JS object heap hits the high-water mark, the V8 GC sweeps
it clean, then tries to grow it in order to avoid more sweeps in the near
future. Rule of thumb: the bigger the JS heap, the lazier the GC can be.
A side effect of a bigger heap is that objects now live longer. This is harmless
in general but it affects zlib context objects because those are tied to large
buffers that live outside the JS heap, on the order of 16K per context object.
Ergo, don't wait for the GC to reclaim the memory - it may take a long time.
Fixes #4172.
12 years ago
|
|
|
};
|
|
|
|
|
|
|
|
Zlib.prototype._transform = function(chunk, output, cb) {
|
|
|
|
var flushFlag;
|
|
|
|
var ws = this._writableState;
|
|
|
|
var ending = ws.ending || ws.ended;
|
|
|
|
var last = ending && (!chunk || ws.length === chunk.length);
|
|
|
|
|
|
|
|
if (chunk !== null && !Buffer.isBuffer(chunk))
|
|
|
|
return cb(new Error('invalid input'));
|
|
|
|
|
|
|
|
// If it's the last chunk, or a final flush, we use the Z_FINISH flush flag.
|
|
|
|
// If it's explicitly flushing at some other time, then we use
|
|
|
|
// Z_FULL_FLUSH. Otherwise, use Z_NO_FLUSH for maximum compression
|
|
|
|
// goodness.
|
|
|
|
if (last)
|
|
|
|
flushFlag = binding.Z_FINISH;
|
|
|
|
else if (chunk === null)
|
|
|
|
flushFlag = binding.Z_FULL_FLUSH;
|
|
|
|
else
|
|
|
|
flushFlag = binding.Z_NO_FLUSH;
|
|
|
|
|
|
|
|
var availInBefore = chunk && chunk.length;
|
|
|
|
var availOutBefore = this._chunkSize - this._offset;
|
|
|
|
var inOff = 0;
|
|
|
|
|
|
|
|
var req = this._binding.write(flushFlag,
|
|
|
|
chunk, // in
|
|
|
|
inOff, // in_off
|
|
|
|
availInBefore, // in_len
|
|
|
|
this._buffer, // out
|
|
|
|
this._offset, //out_off
|
|
|
|
availOutBefore); // out_len
|
|
|
|
|
|
|
|
req.buffer = chunk;
|
|
|
|
req.callback = callback;
|
|
|
|
|
|
|
|
var self = this;
|
|
|
|
function callback(availInAfter, availOutAfter, buffer) {
|
|
|
|
if (self._hadError)
|
|
|
|
return;
|
|
|
|
|
|
|
|
var have = availOutBefore - availOutAfter;
|
|
|
|
assert(have >= 0, 'have should not go down');
|
|
|
|
|
|
|
|
if (have > 0) {
|
|
|
|
var out = self._buffer.slice(self._offset, self._offset + have);
|
|
|
|
self._offset += have;
|
|
|
|
// serve some output to the consumer.
|
|
|
|
output(out);
|
|
|
|
}
|
|
|
|
|
|
|
|
// exhausted the output buffer, or used all the input create a new one.
|
|
|
|
if (availOutAfter === 0 || self._offset >= self._chunkSize) {
|
|
|
|
availOutBefore = self._chunkSize;
|
|
|
|
self._offset = 0;
|
|
|
|
self._buffer = new Buffer(self._chunkSize);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (availOutAfter === 0) {
|
|
|
|
// Not actually done. Need to reprocess.
|
|
|
|
// Also, update the availInBefore to the availInAfter value,
|
|
|
|
// so that if we have to hit it a third (fourth, etc.) time,
|
|
|
|
// it'll have the correct byte counts.
|
|
|
|
inOff += (availInBefore - availInAfter);
|
|
|
|
availInBefore = availInAfter;
|
|
|
|
|
|
|
|
var newReq = self._binding.write(flushFlag,
|
|
|
|
chunk,
|
|
|
|
inOff,
|
|
|
|
availInBefore,
|
|
|
|
self._buffer,
|
|
|
|
self._offset,
|
|
|
|
self._chunkSize);
|
|
|
|
newReq.callback = callback; // this same function
|
|
|
|
newReq.buffer = chunk;
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
// finished with the chunk.
|
|
|
|
cb();
|
|
|
|
}
|
|
|
|
};
|
|
|
|
|
|
|
|
util.inherits(Deflate, Zlib);
|
|
|
|
util.inherits(Inflate, Zlib);
|
|
|
|
util.inherits(Gzip, Zlib);
|
|
|
|
util.inherits(Gunzip, Zlib);
|
|
|
|
util.inherits(DeflateRaw, Zlib);
|
|
|
|
util.inherits(InflateRaw, Zlib);
|
|
|
|
util.inherits(Unzip, Zlib);
|