Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Improve performances #56

Merged
merged 9 commits into from
Jul 1, 2013
Merged

Improve performances #56

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
ea2aaf5
Lazy decompress data
dduponchel May 9, 2013
3754a98
Replace the INFLATE/DEFLATE implementations
dduponchel May 9, 2013
9656c5c
Update documentation
dduponchel May 9, 2013
6c75955
Tests : better display of unexpected errors.
dduponchel May 11, 2013
7f6387f
inflate/deflate : v0.1.6
dduponchel May 11, 2013
b84e00d
optimize memory consumption instead of cpu consumption
dduponchel May 11, 2013
54dcb7e
remove an old file
dduponchel May 17, 2013
a98f7ab
Avoid if possible strings manipulation.
dduponchel May 25, 2013
14217a0
performances : small gains on chrome
dduponchel Jun 30, 2013
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
77 changes: 62 additions & 15 deletions index.html
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -280,7 +280,6 @@ <h4 id="doc_file_name">file(name)</h4>
<dd>The file if any, <code>null</code> otherwise. The file has the following structure :
<ul>
<li><code>name</code> the absolute path of the file</li>
<li><code>data</code> the data contained in the file</li>
<li><code>options</code> the options of the file. The available options are :
<ul>
<li><code>base64</code>, boolean, cf <a href="#doc_file_name_data_options">file(name, data [,options])</a></li>
Expand All @@ -302,12 +301,11 @@ <h4 id="doc_file_name">file(name)</h4>
zip.file("file.txt", "content");

zip.file("file.txt").name // "file.txt"
zip.file("file.txt").data // "content"
zip.file("file.txt").asText() // "content"
zip.file("file.txt").options.dir // false

// utf8 example
var zip = new JSZip(zipFromAjaxWithUTF8);
zip.file("amount.txt").data // "&#226;&#130;&#172;15" <!-- something like that -->
zip.file("amount.txt").asText() // "&euro;15"
zip.file("amount.txt").asArrayBuffer() // an ArrayBuffer containing &euro;15
zip.file("amount.txt").asUint8Array() // an Uint8Array containing &euro;15
Expand Down Expand Up @@ -339,7 +337,7 @@ <h4 id="doc_file_regex">file(regex)</h4>
folder.file(/^file/); // array of size 2, the relative paths start with file

// arrays contain objects in the form:
// {name: "file2.txt", data: "content", dir: false}</pre>
// {name: "file2.txt", dir: false, asText : function () {...}, ...}</pre>

<h4 id="doc_file_name_data_options">file(name, data [,options])</h4>
<dl>
Expand All @@ -359,7 +357,7 @@ <h4 id="doc_file_name_data_options">file(name, data [,options])</h4>
<li><code>date</code> (Date) use it to specify the last modification date.
If not set the current date is used.</li>
<li><code>compression</code> (String), default null. If set, specifies the file compression method to use. If not, the default file compression is used, cf <a href="#doc_generate_options">generate(options)</a>.</li>
<li><code>optimizedBinaryString</code> (boolean), default false. Set it to true if (and only if) the input
<li><code>optimizedBinaryString</code> (boolean), default false. Set it to true if (and only if) the input is a string and
has already been prepared with a 0xFF mask.</li>
</ul>
</dd>
Expand Down Expand Up @@ -464,9 +462,16 @@ <h4 id="doc_generate_options">generate(options)</h4>
HTML5 note : when using type = "uint8array", "arraybuffer" or "blob", be sure to check if the browser supports it (you can use <a href="#jszip_support"><code>JSZip.support</code></a>).
This method will throw an exception otherwise.
<pre class="example">
content = zip.generate();
content = zip.generate(); // base64
location.href="data:application/zip;base64,"+content;</pre>

<pre class="example">
var blobLink = document.getElementById('blobLink');
blobLink.download = "hello.zip";
blobLink.href = window.URL.createObjectURL(
zip.generate({type:"blob"})
);</pre>

<pre class="example">
content = zip.generate({type:"string"});
for (var c = 0; c &lt; content.length; c++) {
Expand All @@ -484,6 +489,8 @@ <h4 id="doc_load_data_options">load(data, options)</h4>
<ul>
<li><code>base64</code> (boolean) <code>true</code> if the data is base64 encoded, <code>false</code> for binary. Default : <code>false</code>.</li>
<li><code>checkCRC32</code> (boolean) <code>true</code> if the read data should be checked against its CRC32. Default : <code>false</code>.</li>
<li><code>optimizedBinaryString</code> (boolean), default false. Set it to true if (and only if) the input is a string and
has already been prepared with a 0xFF mask.</li>
</ul>
</dd>
<dt>Returns : </dt>
Expand Down Expand Up @@ -514,7 +521,7 @@ <h4 id="doc_filter_predicate">filter(predicate)</h4>
<dd><code>predicate</code> (function) the predicate to use : <code>function (relativePath, file) {...}</code> It takes 2 arguments : the relative path and the file.
<ul>
<li><code>relativePath</code> (String) The filename and its path, reliatively to the current folder.</li>
<li><code>file</code> (Object) The file being tested. Like the result of <a href="#doc_file_name"><code>file(name)</code></a>, the file has the form <code>{name:"...", data:"...", options:{...}}</code>.</li>
<li><code>file</code> (Object) The file being tested. Like the result of <a href="#doc_file_name"><code>file(name)</code></a>, the file has the form <code>{name:"...", options:{...}, asText:function,...}</code>.</li>
<li>Return true if the file should be included, false otherwise.</li>
</ul>
</dd>
Expand All @@ -526,7 +533,7 @@ <h4 id="doc_filter_predicate">filter(predicate)</h4>
zip.file("readme.txt", "content");
zip.filter(function (relativePath, file){
// relativePath == "readme.txt"
// file = {name:"dir/readme.txt",data:"content",options:{...}}
// file = {name:"dir/readme.txt",options:{...},asText:function}
return true/false;
});</pre>

Expand All @@ -543,11 +550,13 @@ <h4 id="jszip_support">JSZip.support</h4>
</p>

<h3 id="zip_load_limits">Loading zip files, limitations</h3>
<h4>Not supported features</h4>
<p>
All the features of zip files are not supported.
Classic zip files will work but encrypted zip, multi-volume, etc are not supported
and the load() method will throw an <code>Error</code>.
</p>
<h4>ZIP64 and 32bit integers</h4>
<p>
ZIP64 files can be loaded, but only if the zip file is not "too big". ZIP64 uses 64bits integers
but Javascript represents all numbers as
Expand All @@ -559,17 +568,30 @@ <h3 id="zip_load_limits">Loading zip files, limitations</h3>
So if all the 64bits integers can fit into 32 bits integers, everything will be fine. If
it's not the case, you will have other problems anyway (see next limitation).
</p>
<h4>Performance issues</h4>
<p>
An other limitation comes from the browser (and the machine running the browser).
A compressed zip file of 10M is common and easily opened by desktop application,
but not in a browser. The processing of such a beast is likely to be painful :
the browser will eat hundreds of megabytes while using CPU like never.<br/>
If you use an old browser, things will be worse. For example, IE6 and IE7 are
quite slow to to execute the unit tests, and they completely freeze as soon as
they try to handle larger files.<br />
Conclusion : reading small files is OK, reading others is not.
A compressed zip file of 10MB is "easily" opened by firefox/chrome/opera/IE10 but
will crash older IE. Also keep in mind that strings in javascript are encoded in UTF-16 :
a 10MB ascii text file will take 20MB of memory.
</p>
<p>
If you're having performance issues, please consider the following :
<ul>
<li>Don't use IE &lt; 10. Everything is better with WebGL support.</li>
<li>
If you load the file from an ajax call, ask your XHR an ArrayBuffer.
Loading a string is asking for troubles.
</li>
<li>
If you want to get the content of an ASCII file as a string, consider using
<code>asBinary()</code> instead of <code>asText()</code>. The transformation
"binary string" -&gt; "utf string" is a consuming process.
</li>
</ul>
</p>
<h4>The output zip will differ from the input zip</h4>
<p>
Reading and generating a zip file won't give you back the same file.
Some data are discarded (file metadata) and other are added (subfolders).
</p>
Expand All @@ -586,6 +608,30 @@ <h3>1.0.0 <span>2013-02-14</span></h3>
<p>
First release after a long period without version.
</p>

<h2 id="migrating_guide">Migrating Guide</h2>
<h3>From 1.0.x to ?</h3>
<p>
<ul>
<li><code>JSZipBase64</code> has been renamed to <code>JSZip.base64</code>.</li>
<li>
The <code>data</code> attribute doesn't exist anymore :
use the getters <code>asText()</code>, <code>asBinary()</code>, etc
</li>
<li>
The compression/decompression methods now give their input type with the
<code>compressInputType</code> and <code>uncompressInputType<code> attributes.
</ul>

Example for the data attribute :
<pre>// before
zip.file("test.txt").data;
zip.files["test.txt"].data;

// after
zip.file("test.txt").asText();
zip.files["test.txt"].asText();
</p>
</div>

<div id="who" class="grid_3">
Expand All @@ -598,6 +644,7 @@ <h3>License</h3>

<h3>Version</h3>
<p>1.0.1</p>
<p>See the <a href="#migrating_guide">migrating guide</a> when updating the library !</p>

<h3>Todo</h3>
<ul>
Expand Down
Loading