beer planet » Tips

[Updated x3] Goodbye Outlook, I Don't Need You Anymore – Gmail Now Lets You Paste Images Directly From Clipboard

Artem Russakovskii — Mon, 13 Jun 2011 17:13:00 +0000

There used to be a time when you couldn't imagine your life without Microsoft's Outlook – web email clients were pathetic and non-functional, Internet access was scarce, and access to certain advanced features was only possible with a desktop application like Outlook.

Then, Gmail arrived and exploded the whole notion of desktop email clients forever, almost overnight. It was fast, robust, logical, and integrated – many things Outlook still isn't to this day (Outlook 2010 + IMAP is pure hell).

Slowly, Google brought out more and more features that made the fine line between web and desktop emailing thinner and thinner, and today, it finally disappeared, at least for me.

The final nail in the coffin turned out to be something I'd desired for a very long time – taking screenshots and pasting them directly into Gmail in the browser window. Inline image pasting from the clipboard if you will.

Update #1: Hrm, I may have jumped the gun here. While the image appears perfectly fine in the Create interface and can be resized, moved, etc, it seems to be sent out with formatting errors that prevent it from showing up properly on the recipient's side.

Essentially, what happens is the image is encoded inline, but for some reason the encoding is done wrong.

Here is what Gmail shows this image was encoded as (via "Show original"). I don't know if it's Gmail adding these or not, but there are a ton of =\r\n characters along with "3D" in there that break everything.

Here is what this image is supposed to look like with those things stripped (it works in Chrome and Firefox just fine).

We're close, really close.

Update #2: HowToGeek investigated this new capability with Chrome and found that while it doesn't work with the clipboard, it does work if you copy and paste images within Chrome itself, and it doesn't suffer from the bug above.

After digging into it, I found that it only works because rather than create the image inline, like Firefox does, Chrome just makes an tag with the src parameter pointing to the original url.

This means the image in the email will disappear if it's deleted and simply won't show up if you tried to copy something that requires a login.

Update #3 (6/13/11): According to the Gmail blog, you can finally paste images directly into your emails, albeit in Chrome only at this time!

I take a lot of screenshots of my desktop or certain parts of it when doing reviews or reporting bugs (with OneNote or SnagIt), and having to save the image, find it on the disk, and drop it into Gmail, while relatively easy, was still too cumbersome. In that regard, Outlook was great – I could quickly pump out a few screenshots and send off an email with all of them inline with text in only a few seconds. If I attempted to do that in Gmail, I either got nothing or a little image basically saying how full of fail I was.

It all changed with Firefox 4 (at least that's my theory) – I am now able to take screenshots and paste them straight into Gmail, right in the browser. Let's take a look:

Amazing, isn't it? This does not work for me in Chrome 12 or IE 9 – only Firefox 4 at the moment. I'm not sure how it works or why it started working only now, but I'm sure one of you will eventually come here and provide a sensible explanation.

Goodbye Outlook, I hardly knew ye.

P.S. I'm using Windows 7 in case it matters.

Similar Posts:

How To Dynamically Increase Memory Limits When Interfacing With WordPress Using XML-RPC (Windows Live Writer, Etc)

Artem Russakovskii — Sat, 06 Nov 2010 09:00:44 +0000

Today's snippet is tremendously helpful if you are using an XML-RPC WordPress interface to read and publish your articles and are running into 500 Server Error issues due to running out of memory, manifesting themselves in something like this error message: "Invalid Server Response – The response to the metaWeblog.newMediaObject method received from the weblog server was invalid".

For example, my regular PHP memory allocation is 32MB or so, but if I load up Windows Live Writer, my favorite publishing tool, and ask it to load 1000 of the latest blog posts, I will undoubtedly get a server error back.

One solution would be to increase the memory allocated to PHP to something higher, like 256MB, which is how I used to get around the issue. I say "get around" because it's not a good solution – if someone finds a page that uses a lot of memory on your site, they could easily kill not only your web server but the whole machine due to swap death. Keeping a lower memory limit allows you to run your web server, such as Apache, with more children, thus serving more requests without getting overloaded.

So, I've looked into the WordPress core and came up with what I think is a proper fix – dynamic memory limit tweaking when dealing with XML-RPC only. Here is the code – add it to your functions.php and you should be golden:

/**
* Dynamically increase allowed memory limit for XML-RPC only.
*
* @param array $methods
* @return array
*/
function my_xmlrpc_methods($methods) {
  ini_set('memory_limit', '256M');
  return $methods;
}
add_action('xmlrpc_methods', 'my_xmlrpc_methods');

The code is pretty self-explanatory: it hooks into a WordPress hook that fires only for XML-RPC requests and adjusts the memory limit on the fly. No more out of memory errors and I'm able to load 1000 posts in Windows Live Writer without problems.

Similar Posts:

How To Add Shortcode Support To WordPress Comments And Sidebar Widgets

Artem Russakovskii — Sat, 16 Oct 2010 03:20:26 +0000

WordPress has a great way of letting you use simple text tags called shortcodes to provide a whole bunch of functionality, including custom PHP code. In this article, I'm assuming that you already know what shortcodes do and how they operate (if you don't, head over here: Shortcode_API).

One glaring omission in the way shortcodes are set up by default is that they only get triggered in the content of your post, leaving the sidebar and comments out. I'm sure this is done for security, so that your readers can't screw something up by posting shortcodes they're not supposed to – after all, shortcodes are PHP snippets on the backend.

However, let's assume you really know what you're doing and you want, for example, to allow certain shortcodes in the comments and all shortcodes in the sidebar. It's your site and you should be able to do whatever you want with it. For instance, I wanted to provide a way to include QR codes using a [qr] shortcode syntax that I'd written both in the sidebar and in the comments, except in the comments that was the only shortcode I wanted to allow my readers to use.

Since these cases can get complicated, let's start simple and then progressively get deeper.

Allow all shortcodes in the sidebar

Assuming you want to enable shortcodes in the PHP Code Widget widgets (I believe this is the most common way to add PHP support to the sidebar), use this code:

1	add_filter('widget_execphp', 'do_shortcode');

If you have some other widgets in mind, you have to find out the right filter and use it just like above.

Allow only whitelisted shortcodes in the comments

Since I don't want all shortcodes to fire here, things get slightly trickier.

function init_common_shortcodes() {
  add_shortcode('qr', 'my_shortcode_qr');
}
 
function init_comment_shortcodes() {
  remove_all_shortcodes();
  init_common_shortcodes();
  add_filter('comment_text', 'do_shortcode');
}
 
init_common_shortcodes();
add_filter('comments_template', 'init_comment_shortcodes');

The code above will remove all shortcodes when it gets to comments and reinitialize only the whitelisted one called "[qr]", which is exactly what I wanted.

Note: If your code somehow uses shortcodes after the comments (such as a sidebar or a custom post query), you will need the trick in the next section.

Allow all shortcodes in the sidebar AND only whitelisted ones in the comments

At this point, things get slightly tricky because if I combine the previous 2 solutions into 1, by the time WordPress gets to the sidebar, it'll be left without any shortcodes at all – they've been removed by the code dealing with comments. Therefore, I need to save the shortcodes before dealing with comments and then restore them after the comments and before the sidebar.

function my_remove_all_shortcodes() {
  global $shortcode_tags;
  global $temp_shortcode_tags;
  $temp_shortcode_tags = $shortcode_tags;
  remove_all_shortcodes();
}
 
function my_restore_all_shortcodes() {
  global $shortcode_tags;
  global $temp_shortcode_tags;
  if(!empty($temp_shortcode_tags)) {
    $shortcode_tags = $temp_shortcode_tags;
  }
}
 
function init_common_shortcodes() {
  add_shortcode('qr', 'my_shortcode_qr');
}
 
function init_comment_shortcodes() {
  my_remove_all_shortcodes();
  init_common_shortcodes();
  add_filter('comment_text', 'do_shortcode');
}
 
add_filter('comments_template', 'init_comment_shortcodes');
add_filter('dynamic_sidebar', 'my_restore_all_shortcodes');
add_filter('widget_execphp', 'do_shortcode');
init_common_shortcodes();

Here is what the code does:

Initialize custom shortcodes that I want to add all over the place.
When WP gets to the comment block, it removes all shortcodes (again, for security) and re-initializes only the ones I want, but this time remembers them so that they can be restored later.
After the comments, when WP gets to the dynamic sidebar, it restores all the shortcodes and applies them to each PHP Code Widget.

And there you have it. Of course, the code may need tweaking for your own scenario but the ideas are all there. If you spot an error or know of an easier/better way, please let me know.

Similar Posts:

How To Update Eclipse From Galileo (3.5) To Helios (3.6) In-Place Without Reinstalling

Artem Russakovskii — Wed, 29 Sep 2010 23:47:21 +0000

As a developer, I both love and hate Eclipse for its chaotic nature, buggy and sometimes unusable interface, but at the same time incredible usefulness and ability to serve as a single tool for all of my development, be it C++, PHP, Java, Android, Perl, etc.

One of the biggest problems with Eclipse is that there is no clear upgrade path from major versions, for example 3.5->3.6. What I ended up having to do for years is back up the old release, download and unpack the new release, and then try to migrate all the settings by importing and exporting left and right. Not so pleasant.

Turns out, as of Eclipse 3.3 (though I've only tried it with Eclipse 3.5), it's possible to successfully complete an in-place upgrade without doing any wiping or voodoo hackery. I've successfully completed this on 2 of my machines and now need to share it with the rest of the world.

Step 1. Go to Help > Install New Software… > Available Software Sites or Windows > Preferences > Install/Update > Available Software Sites. I personally prefer the former way because it results in fewer clicks but both end up in the same place.

Step 2. Everywhere you see 3.5, press Edit and change it to 3.6. Similarly, change "galileo" to "helios". For example, you should end up with at least these (if you don't have these, just add them):

http://download.eclipse.org/eclipse/updates/3.6
http://download.eclipse.org/releases/helios

Note: you may also need to update other repositories (for example, Galileo was compatible with CDT 6, while Helios needed CDT 7).

Step 3. Press OK, then go to Help > Check for Updates. You should see the progress in the Progress tab – let the magic happen, restart when prompted, and after a few minutes, you should have a working upgraded version of Eclipse.

Similar Posts:

How To Display Just The HTTP Response Code In Command Line Curl

Artem Russakovskii — Fri, 11 Jun 2010 05:49:05 +0000

Today, I was looking for a quick way to see HTTP response codes of a bunch of urls. Naturally, I turned to the curl command, which I would usually use like this:

curl -IL "URL"

This command would send a HEAD request (-I), follow through all redirects (-L), and display some useful information in the end. Most of the time it's ideal:

curl -IL "http://www.google.com"
 
HTTP/1.1 200 OK
Date: Fri, 11 Jun 2010 03:58:55 GMT
Expires: -1
Cache-Control: private, max-age=0
Content-Type: text/html; charset=ISO-8859-1
Server: gws
X-XSS-Protection: 1; mode=block
Transfer-Encoding: chunked

However, the server I was curling didn't support HEAD requests explicitly. Additionally, I was really only interested in HTTP status codes and not in the rest of the output. This means I would have to change my strategy and issue GET requests, ignoring HTML output completely.

Curl manual to the rescue. A few minutes later, I came up with the following, which served my needs perfectly:

curl -sL -w "%{http_code} %{url_effective}\\n" "URL" -o /dev/null

Here is a sample of what comes out:

curl -sL -w "%{http_code} %{url_effective}\\n" "http://www.amazon.com/Kindle-Wireless-Reading-Display-Generation/dp/B0015T963C?tag=androidpolice-20" -o /dev/null
 
200 http://www.amazon.com/Kindle-Wireless-Reading-Display-Generation/dp/B0015T963C

Here, -s silences curl's progress output, -L follows all redirects as before, -w prints the report using a custom format, and -o redirects curl's HTML output to /dev/null.

Here are the other special variables available in case you want to customize the output some more:

url_effective
http_code
http_connect
time_total
time_namelookup
time_connect
time_pretransfer
time_redirect
time_starttransfer
size_download
size_upload
size_header
size_request
speed_download
speed_upload
content_type
num_connects
num_redirects
ftp_entry_path

Is there a better way to do this with curl? Perhaps, but this way offers the most flexibility, as I am in control of all the formatting.

Similar Posts:

How To Show Hidden/Invisible Files In Total Commander, Both Locally And On An FTP Server

Artem Russakovskii — Sat, 06 Mar 2010 08:56:49 +0000

Today I have 2 tips for Total Commander users:

how to display hidden local directories and files and
how to display hidden FTP directories and files

Really, Total Commander should just control this setting in one place but, unfortunately, it is not the case.

I usually prefer when my file manager shows me everything I have, so that I can be more in control and see the hidden directories, such as .svn or $Recycle.Bin, and files, such as .bashrc or pagefile.sys.

How To Display Hidden Local Directories And Files

go to Configuration -> Options… -> Display
put a check next to the "Show hidden/system files (for experts only)"

How To Display Hidden FTP Directories And Files

This one is a bit trickier – you would think the same option as above controls this, and you would be wrong.

Furthermore, you would think that it's an option in the FTP manager (Ctrl-F), and you would be wrong again.

Here is how to do it:

connect to the FTP server you need
select Net -> FTP Show Hidden Files

Weird choice for an option, isn't it?

Apparently, at least as of Total Commander 7.50a, this option persists while Total Commander is running but it is not saved when you restart it.

Similar Posts:

The Most Awesome VPN Tip: How To Make Windows Automatically Use Your Local WiFi/LAN Connection Directly For Requests That Don't Need To Go Through VPN

Artem Russakovskii — Tue, 02 Mar 2010 00:45:05 +0000

This tip can also be filed in the "post with the longest title that kind of makes sense but needs more explanation" category.

If you use a VPN (Virtual Private Network), this tip is for you.

The Problem

you connect to a VPN to get access to your work/whatever network
your connection is fast but the VPN connection is balls slow
you try to stream a bit of online radio, go to a website, watch a video, or do anything, which is automatically routed through the VPN connection but everything TAKES AGES because the VPN connection is the limiting factor
so not only are you frustrated by hiccupping radio, stuttering video, and a never disappearing progress bar but you're also making your already slow link slower, so now remote desktop and other activity for which you've VPN'ed in the first place is taking even longer

Familiar situation? You bet.

The Solution

This solution is for Windows users only but I'm sure similar settings are available in MacOS, Linux, etc.

Here in particular I am using Windows 7 but this solution applies to Windows XP and Vista as well (the menus look only slightly different for those).

1. Open up the VPN Connection's Properties

These properties can be reached by right clicking on the VPN connection in the list of networks and selecting Properties.

2. Go to the TCP Properties

On Windows XP, you might see only the "Internet Protocol (TCP/IP)", while on Windows 7 you will see both IPv4 and IPv6. You will most likely need to select IPv4 then.

3. Head to the Advanced settings

4. Uncheck the "Use default gateway on remote network" checkbox

5. Restart the VPN connection

This should do the trick.

The New Behavior

The new behavior from what I can tell is:

use a local connection, such as your WiFi or wired connection, so your radio, streaming video, websites, etc should now bypass the VPN connection
if the request above fails, Windows should only then fall back on the VPN connection

When Not To Use This Tweak

Do not use this tweak if you rely on the VPN to always encrypt all your network traffic.

Temporarily disable this tweak if normally you are OK with not encrypting all of your traffic but are currently using an unsecure network. Examples include a WiFi connection that does not prompt for a password or a public network, such as an Internet cafe.

Enjoy your faster VPN experience!

Similar Posts:

How To View A Specific SVN Revision In Your Browser

Artem Russakovskii — Sun, 21 Feb 2010 06:04:17 +0000

This is a quick recipe that I found pretty interesting and relatively unknown.

Everyone who uses SVN knows that most repositories are set up to allow viewing of their contents via a web browser. For example, here's the trunk of WP Plugins SVN: http://plugins.svn.wordpress.org/ and here is the current trunk version of a specific file, let's say http://plugins.svn.wordpress.org/stats/trunk/readme.txt.

The Problem

However, what if you wanted to view a specific revision of a file or directory in your browser?

Let's say I wanted revision 100,000 of http://plugins.svn.wordpress.org/stats/trunk/readme.txt

Normally, on a command line, you'd do something like

svn co http://plugins.svn.wordpress.org/stats/trunk/readme.txt stats
cd stats;
svn up -r100000 readme.txt

or simply

svn export -r100000 http://plugins.svn.wordpress.org/stats/trunk/readme.txt

However, how would you do this in your browser? Trying something like ?r=100000 or ?revision=100000 parameters doesn't work.

The Solution

Here is the syntax to get a specific revision of a file or directory:

Take the root of the repository (in our case, http://plugins.svn.wordpress.org)
Append this string !svn/bc/REVISION_NUMBER right after the repository root and before the directory or file path.

Going back to our example:

http://plugins.svn.wordpress.org/!svn/bc/100000/stats/
shows the stats directory exactly as it was at revision 100,000 and
http://plugins.svn.wordpress.org/!svn/bc/100000/stats/trunk/readme.txt
shows our readme.txt at 100,000.

And that's how you do it, folks. Happy revisioning!

Update: this post is now sponsored by Wirefly Learn where you can find cell phone news and reviews, as well as comparisons and videos.

Similar Posts:

How To Remove Inline Hardcoded Recent Comments Sidebar Widget Style From Your WordPress Theme

Artem Russakovskii — Sun, 31 Jan 2010 09:45:15 +0000

The Problem

If you use the Recent Comments sidebar widget in your WordPress installation, it's possible that you want to customize this widget's style.

You will quickly find, however, that as soon as you add the widget to your sidebar, it injects the following inline, hardcoded CSS into the containing page (using !important to make things worse):

The code above comes from recent_comments_style() (found in wp-includes/default-widgets.php), which is in turn called by WP_Widget_Recent_Comments() in the same file (this is just an old-style PHP4 constructor – same as PHP5's __construct()), which is triggered when the Recent Comments widget is used:

add_action( 'wp_head', array(&$this, 'recent_comments_style') );

This leaves a bad taste in my mouth because:

the style is automatically forced into any page that embeds the widget
it uses !important to override any existing styles
it cannot be configured from the widget
this is a core WordPress widget and not some widget added by a random plugin. WordPress should not contain such bad practices.

After seeing this, I raised a new issue #11928 in the WordPress bug tracker but in the meantime, of course, I could go and hack the default-widgets.php core file to solve this problem. That, of course, would be a horrible thing to do – the next upgrade could wipe my changes or result in a conflict (if upgrading via svn). In general, one should never need to modify core files.

The Solution

Fortunately, there is a solution that doesn't involve modifying core files. Remember that we are talking about WordPress here, which gained its popularity due to the ease of customization.

Normally, we would just need to remove the recent_comments_style hook, which would prevent it from ever firing and injecting the CSS. In fact, somebody over at webstractions.com already took a crack at this approach with the following code, which may have worked for a while but doesn't anymore:

function remove_wp_widget_recent_comments_style() {
   if ( has_filter('wp_head', 'wp_widget_recent_comments_style') ) {
      remove_filter('wp_head', 'wp_widget_recent_comments_style' );
   }
}
add_filter( 'wp_head', 'remove_wp_widget_recent_comments_style', 1 );

This code doesn't work in the current version of WordPress because recent_comments_style() belongs to the WP_Widget_Recent_Comments class, and we need a tiny bit more effort to get through to the object of that class.

The following snippet, suggested by Andrew Nacin in a comment to the bug I opened up, does the job and works in WordPress 2.9.1:

add_action('widgets_init', 'my_remove_recent_comments_style');
function my_remove_recent_comments_style() {
	global $wp_widget_factory;
	remove_action('wp_head', array($wp_widget_factory->widgets['WP_Widget_Recent_Comments'], 'recent_comments_style'));
}

This code first gets access to the right object and then passes it to the remove_action() function, along with the right function name recent_comments_style. Stick it intoyour functions.php and it should do the trick.

And there you have it.

Similar Posts:

[WordPress Plugin Development] How To Include CSS and JavaScript Conditionally And Only When Needed By The Posts

Artem Russakovskii — Wed, 13 Jan 2010 17:15:00 +0000

Introduction

In this tutorial, I am going to introduce a WordPress technique that I believe was unpublished until I raised the question a few days ago on the WordPress forums.

In short, the problem I was trying to solve was plugins unnecessarily loading their JavaScript and CSS on *every* page of the blog, even when doing so would achieve absolutely nothing and the plugin wouldn't do any work.

Update #1: I have posted a follow-up in response to some comments received around the web.

Update #2: There is a solution that can be considered a compromise as it works well for loading JavaScript but doesn't handle CSS.

I briefly mentioned this approach here but but Scribu decided to expand on it by providing a nice Jedi-themed tutorial. It is available here.

Let me explain using this example:

a code formatter plugin only does something useful when it sees a [code] shortcode in any post on the page.
most of your posts do not contain the [code] shortcode as you don't include code snippets that often or you only started using this particular plugin recently.
the plugin, however, loads the CSS and JS (which are most likely GeSHi and take up loads of space) on every page.
these CSS and JS do absolutely nothing on most page loads.
bandwidth is wasted, extra DNS and HTTP requests are processed, the browser is slowed down, and for what? For no good reason, other than the plugin author didn't know how to achieve this conditional loading.

If you think about it, there are many plugins that only do something once in a blue moon. Table of contents, text manipulators, galleries, sliders, etc, etc. If only they loaded their frontend code strictly when necessary, most page loads would suddenly become much lighter.

So what can we do to solve this? Let's look at a few techniques.

Loading CSS And JS In Place?

Here's one, albeit pretty bad, solution - only print the CSS and JS includes if and when you determine somewhere in the middle of loading the posts that the scripts and styles are indeed needed, then set a flag to avoid printing them again. In our example, that would be when the plugin detects the [code] shortcode.

This, however, is a mediocre solution because, while it's not a bad idea to load Javascript in the footer, CSS should be loaded in the header, otherwise the page might look unformatted until the CSS is reached.

Additionally, it's not the cleanest and most robust solution because you shouldn't be writing