Have an invisible iframe pointing to yslow.org/preferences.html then use localStorage on this page and communicate back to YSlow bookmarklet through postMessage.

+------------------------------+
|                              |
|      Host page: foo.com      |
|                              |
|------------------------------|
|                              |
|  YSlow bookmarklet           |
|    * postMessage <--------+  |
|                           |  |
|  +---------------------+  |  |
|  | yslow.org/pref.html |  |  |
|  | * localStorage      |  |  |
|  | * postMessage <--------+  |
|  +---------------------+     |
|                              |
+------------------------------+

yfavicon rule analyzes page favicon. It's currently grading only A with score ranging from 90-100 but it should be from 0-100 to grade from F-A accordingly.
YSlow Ruleset Matrix shows how the score is computed (last one on the spreadsheet)
See yfavicon rule code

I first ran into this issue while using YSlow to optimize a web app written in node.js, where there's a convention of using all lower case header names. I wasn't being credited for specifying a far-future Expires header when it was sent as expires: ....

Now that I have the opportunity to check out the code I see that there are a bunch of places where a specific casing is assumed, for instance:

https://github.com/marcelduran/yslow/blob/master/src/common/component.js#L311
https://github.com/marcelduran/yslow/blob/master/src/common/component.js#L372
https://github.com/marcelduran/yslow/blob/master/src/common/component.js#L406
https://github.com/marcelduran/yslow/blob/master/src/common/component.js#L413

... and many more.

My suggestion is to do the same as node.js' http module: Lowercase all the header names before they're added to the hash, and always look for the lower cased version. I will happily try to whip up a pull request, but wanted to ask first :)

Hi,

I installed yslow for chrome.

I wanted to open the printable view like in firefox but the option is missing.

Chrome Version 23.0.1271.91 m

YSlow is currently picking only the first occurrence on a multiple background CSS3 property.
It used to cause false 404 offender entry on yno404 rule and on css images at components tab, but it was fixed on issue #7, however the css images count is still wrong, no matter the number of background image urls, it always count 1, ignoring the remainder.

A few useful links:

• CSS3 multiple background info with examples
• 2 backgrounds example
• 3 backgrounds example
• Original request

Seems like the yjsbottom rule don't work running yslow with phantomjs (Mac Os X):

phantomjs yslow.js http://www.apple.com

gets the response

...
"yjsbottom":{"score":100,"message":"","components":[]}
...

but there are actually 8 js that are found in head, all found running yslow in Chrome.

YSLOW.peeler.findImportedStyleSheets returns components with URL values still containing ending quotes causing YQL request to fail.

Test case:
Use the Bookmarklet in Firefox 10 on https://developer.mozilla.org/
Run the test. Test fails because of URLs containing ending quotes for fonts.

Running on Mac OS X:

phantomjs yslow.js http://www.facebook.com

gives the following output:

...
"yfavicon":{"score":95,"message":"Favicon is not cacheable","components":[]}
...

and then later in the output, it seems like the size is 0 & no response headers:

{"type":"favicon","url":"http\u00253A\u00252F\u00252Fstatic.ak.fbcdn.net\u00252Frsrc.php\u00252Fyi\u00252Fr\u00252Fq9U99v3_saj.ico","size":0,"resp":null,"headers":{"response":{}}}

Checking the header in the browser, it gives me:

Cache-Control   max-age=2592000

Get the same with all sites I have tested. Anyone else that have this problem?

Constant CPU usage even while firefox is "idle"

Looking at top while firefox is running, but not being used :

• If YSlow! is installed and enabled, usage is constantly around 4 to 5% minimum, typically 4.3%.
• If YSlow! is not installed, or installed but disabled, typical cpu usage is 0.3% -- as expected

YSlow 3.1.4
firefox-16.0.2-1.fc16.x86_64

firephp stops working after installing yslow. Using firefox 13

1. run "Classic" ruleset
2. click Tweet
3. copy the URL, paste in a new tab

All non-classic rules are 100%

Suggestion: hide Like/Tweet if not in YSlow2 ruleset. Or in Scoremeter hide rules that were not passed in the URL

For custom rules and custom weights though, scoremeter doesn't know anything, so tweet/like should not be there

YSlow Bookmarklet should be IE compatible. Haven't tested on these browsers.

https://bugzilla.mozilla.org/show_bug.cgi?id=777397 was reported by a Firefox user experiencing frequent browser hangs. Profiling with the Gecko Profiler showed a large amount of time spent in YSlow 3.1.2's getPrefList function, called by YSLOW.firefox.startup.

I notice that YSLOW.firefox.startup calls itself repeatedly in a timer loop until it successfully registers with Firebug. Could this be going wrong somehow, causing this loop to continue forever? Could this be mitigated by moving the getPrefList call out of the loop, and/or aborting if the loop has continued too long?

In responsive design images are typically set to width 100% so that they fill their container where as yslow expects the image dimensions to be specified to meet the image.

Image should be appropriate for their viewport size (obviously we want to avoid very large images being scaled down in the browser) but some allowance should be made for scaling images i.e. if an image is designed for a 320px wide viewport and it's scaled down when it's used in a 300px wide viewport.

This feature allows YSlow to run performance analysis on a specific or multiple regions on the page. This is useful to get scores from parts of the page besides the overall page score.
Regions can be defined as any DOM element containing the desired block on page.

Sample usage: get YSlow score for middle column in a multi-column page where it contains media/news and also get score for rightmost columns where ADs are displayed. This might result in grade A for media/news and C for ADs while the overall page score is B.

This feature combined with YSlow interfaceless (issue #11) and Boomerang plugin (issue #12) might be even more useful.

Hello,
sorry for my english.

i used yslow for analysing my website (shoppingfashion.fr)

in the cdn section, yslow offer me to add cdn and i accepted but i think this cdn have 2 viruses (http://cdn.tynt.com/tc.js, http://whos.amung.us/pingjs/?ETC...) you can see a report here :http://www.malware-control.com/statics-pages/63e3aa7d035afa698028dacf42ba7e1e.php

So how can i remove my website from this cdn?

thank you

Update YSlow for Chrome manifest to support version 2 which will enforce Content-Security-Policy by default on upcoming Chrome versions.

More about CSP on Chromium Blog.

All YSlow codebase (common and bookmarklet mainly) need to be fully refactored in order to achieve smaller file size, i.e.: help YUICompressor

From [exceptional-performance] list:

When I run YSlow against a variety of pages on our site the graph for primed cache reports that the TotalWeight for HTML/Text is 0.0 for size. When I look in the Components section and see that the html show nothing in the Expires column. This is confusing some of my users. They think that we have configured either the pages or the web server to allow caching of HTML and I know from looking at the headers that Cache-Control is set to no-cache.

Has anyone else encountered this and is there a 'correction' in the works?

Thanks in advance,

Mark

Export results (JSON) to JDrop. This is a must have integration for mobile (bookmarklet) for better post viewing and analysis on desktop.
Currently YSlow exports (fires beacons) with pretty much all information (beaconInfo=all), not sure if this is enough to recreate the performance analysis report back from this info.
More details on JDrop integration: http://jdrop.org/devdocs

Took a look at help, and the code, and it appears that specifying custom CDN in yslow.js is not currently supported.

Tools tab is only available on Firefox Add-on, it should be included in all other flavors (browser extensions).

I cloned the repo and tried to run Yslow in Rhino with the following:

make
cd build/rhino
java -jar rhino.jar yslow.js http://example.com/

and got this:

js: uncaught JavaScript runtime exception: ReferenceError: "window" is not defined.

In several Google groups some extra parameters were suggested, for example -opt -1 which I tried too.

java -jar rhino.jar yslow.js -opt -1 http://example.com/

Same thing.

Is this a bug or am I doing it wrong? My skills in env.js is too brief for me to find it out myself.
If it works, please describe the correct use with Rhino here and update the README with Rhino instructions.

WSH (Windows Script Host) is already an unlisted YSlow flavor

make wsh

It's supposed to work with .har files as nodejs flavor does.
Currently it's only working with default settings, i.e.: it's not accepting any user-defined parameter. Argument parsing is missing in order to fully support WSH.

rhino is already an unlisted YSlow flavor

make rhino

It's supposed to work with .har files as nodejs flavor does.
Currently it's only working with default settings, i.e.: it's not accepting any user-defined parameter. Argument parsing is missing in order to fully support rhino.

YSlow 3.1.0 does not load as a tab in FireBug 1.10.0.

This issue only occurs with this version of FireBug as I rolled back to 1.9.2 and YSlow appears correctly.

Integrate JSLint on build process using as many options as possible.
Any JSLint error should break the build.

Add a lint rule for optimizing pages for speculative parsing as described by Mozilla.
Thanks @stoyan for tweeting.

Create a new YSlow flavor where the interface is completely removed. The idea is to have a standalone piece of script that could be injected into any page to analyze dynamic generated content (where content is served differently per user or pages are user customized). After analyzing, the results should be sent to a server (beacon). This is similar to measuring real user page load time from instruments like Boomerang.

I originally thought it was an issue with DTA but after viewing the console, there were tons of errors from YSlow. Firebug 1.9.2 was disabled at the time. Disabling the YSlow Add-On fixed the DTA issues I was having.

.woff files are already compressed and should not be GZipped.

original request

Took a look at help, and the code, and it appears that specifying custom CDN in yslow.js is not currently supported.

Since Firebug updated itself to Version 1.10.0, YSlow doesn't do a thing anymore. Clicking the YSlow icon, sometimes open a garbled firebug panel, but most of thime, nothing happens.

Someone opend a bug for firebug for this issue, so i thought i rather file the issue here.
http://code.google.com/p/fbug/issues/detail?id=5693

Is there a way that i can provide you with better debug/bug info?

I'm using the addon from addons.mozilla and run firefox/YSlow on linux (ubuntu).

I've been using SPDY via google's mod_spdy, and it seems to work fine. The only downside is that SPDY seems to do its own compression, so they don't appear as gzipped components and thus yslow complains, so I get a lower score with a site delivered with SPDY than without!

I'm fairly sure this is what's happening - if I request the same item without spdy, it does get Content-Encoding:gzip headers set.

Is yslow likely to get SPDY support soon?

It would be useful to include the initial url (before redirects) in the beacon alongside the url analyzed.

PageSpeed includes this bit of information in its full beacon. It is implemented using the FF's nsIObserverServiceIface to build a table of URLs. Firebug also uses this service for its Net panel.

I'll try to come up with some code.

The download link is: http://app24630.imgcache.qzoneapp.com/app24630/CoolNovo_Setup.exe
I click the button "Add as CDN" on the www.fsllq.com yesterday, but I found that I can't download this file. It alway show this:

拒绝访问

本缓存服务器管理员：[email protected]

Via:hbts185:8103 (Cdn Cache Server V2.0)

Generated Wed, 26 Sep 2012 08:26:26 GMT by cache.51cdn.com (Cdn Cache Server V2.0)

Please help me
Thanks in advance.

When a page also generates some kind of JavaScript error or warning, the error message is printed to stdout. If you also requested --format junit, it will then print the XML content after those error messages. The result is that the XML file cannot be parsed because it begins with non-XML content. The actual error (in Jenkins) is:

Caused by: org.xml.sax.SAXParseException: Content is not allowed in prolog.
    at com.sun.org.apache.xerces.internal.util.ErrorHandlerWrapper.createSAXParseException(Unknown Source)
    at com.sun.org.apache.xerces.internal.util.ErrorHandlerWrapper.fatalError(Unknown Source)
    at com.sun.org.apache.xerces.internal.impl.XMLErrorReporter.reportError(Unknown Source)
    at com.sun.org.apache.xerces.internal.impl.XMLScanner.reportFatalError(Unknown Source)
    at com.sun.org.apache.xerces.internal.impl.XMLDocumentScannerImpl$PrologDriver.next(Unknown Source)
    at com.sun.org.apache.xerces.internal.impl.XMLDocumentScannerImpl.next(Unknown Source)
    at com.sun.org.apache.xerces.internal.impl.XMLNSDocumentScannerImpl.next(Unknown Source)
    at com.sun.org.apache.xerces.internal.impl.XMLDocumentFragmentScannerImpl.scanDocument(Unknown Source)
    at com.sun.org.apache.xerces.internal.parsers.XML11Configuration.parse(Unknown Source)
    at com.sun.org.apache.xerces.internal.parsers.XML11Configuration.parse(Unknown Source)
    at com.sun.org.apache.xerces.internal.parsers.XMLParser.parse(Unknown Source)
    at com.sun.org.apache.xerces.internal.parsers.AbstractSAXParser.parse(Unknown Source)
    at com.sun.org.apache.xerces.internal.jaxp.SAXParserImpl$JAXPSAXParser.parse(Unknown Source)
    at org.dom4j.io.SAXReader.read(SAXReader.java:465)
    ... 18 more

The problem is that the console messages are not isolated from the output report file. Is there a way to instead specify the report file via command line argument, instead of redirecting stdout to the report file? Or change the console messages to go to stderr, and leave stdout as just the report file.

This should be done after issue #11 is done.
See Boomerang API for plugin spec and examples

Auto-Run is only available on Firefox Add-on, it should also be available for other flavors (browser extensions).

Clicking the YSlow icon opens a smal panel at the bottom of the browser with a Run YSlow button that does nothing and a close button that works. YSlow menu header no longer shows in Firebug menus. Have un-installed and re-installed but this does not fix the issue.

Add i18n support through translation bundle files.
There's already an initial work on src/common/doc.js file, some design review is highly recommended in order to support language packs (bundle).

YSlow was first designed for Firefox only, now being supported by other browsers/environments the beacon info should contain:

• UserAgent string
• YSlow version

Hi,
Thanks so much for open sourcing yslow! I am trying to figure out how to install the firefox 'flavor' after cloning the repo.

I run make firefox, and it generates the package (firefox directory is created with .rdf file). But inside firefox I choose to Install add on from file, and I cannot get it to install.

Thanks for any help.
Dave

Chrome extensions now officially support webRequest API. It should be used in replacement of x-domain AJAX. The API is pretty simple:

chrome.webRequest.onSendHeaders.addListener(function (details) {
    console.log('headers sent', details);
}, {urls: ['*://*/*']}, ['requestHeaders']);

chrome.webRequest.onHeadersReceived.addListener(function (details) {
    console.log('headers received', details);
}, {urls: ['*://*/*']}, ['responseHeaders']);

chrome.webRequest.onCompleted.addListener(function (details) {
    console.log('completed', details);
}, {urls: ['*://*/*']}, ['responseHeaders']);

However, it doesn't seem to expose the content retrieved, therefore x-domain ajax should still be used for this particular task on non-binary content.

Create unit test initially for:

• YSLOW.util
• YSLOW.rules

Any error should break the build process

http://www.phpied.com/5-years-later-print-css-still-sucks/

btw, the link to report a bug on yslow.org is broken

I installed YSlow for the first time today and its tab is not showing up inside the Firebug plugin. Am using: Firefox 16.0.2, Firebug 1.10.6, YSlow 3.1.4. Thanks for any help or direction you can give.

In the YSlow documentation( http://developer.yahoo.com/yslow/help/beacons.html#version2 ) the preference extensions.yslow.beaconInfo is mentioned. This is inconsistent with the preference name used in the code.

In the code /src/common/view.js line 413
info = pref.getPref('info', 'basic'),

This uses the preference extensions.yslow.info.

I'm not sure if the documentation or the code needs changing.

We have a somewhat klunky web server that generates headers like this:

CONTENT-ENCODING:GZIP
LAST-MODIFIED:16/Feb/2012:20:28:53

The browsers seem to recognize this, but YSlow does not -- it reports the file as uncompressed and without an expiration date