Author: Ipstenu (Mika Epstein)

Migrating A WordPress Site With wp-cli
This is … crazy simple. I wanted to move a site from ipstenu.org to ipstenu.com (yes, I own that too). While ipstenu.org is a Multisite network, ipstenu.com is where I put a ton of add-on domains. I was moving a site over and, as it was WordPress, did it in a matter of minutes.

Add the domain to the … domain

Since I don’t care to have multiple hosting accounts, and I’m the only one with SSH/FTP access, it’s safe enough for me to do this. There is a risk when you share multiple domains in one hosting account, that if one gets hacked they’re all vulnerable, but I consider it low in my situation. Every plugin is vetted, every file is checked, and then I went and gave each add-on it’s own FTP account. Neurotic? Thy name is me.

Anyway, I add the new domain to my hosting where I want it.

Create the new DB

I have to make a new database, and generally a new DB user, on the server too.

Export!

On existing hosting, I do this:
```
wp db export
```
That gives me my SQL file, thanks to WP-CLI. It’ll be named example_com.sql and will sit in my folder with .htaccess and everything else.

Copy!

I do it via SSH. I go to the new location and run this:
```
scp -r olduser@example.com:path/to/files/ .
```
Since I have ssh keys set up, it’s easy. If I don’t, I’ll put in the password, but that’s straightforward.

Edit wp-config.php

Now I have to point to the new DB. Sometimes I name it the same, but usually I don’t, so I’ll edit the DB name, DB user, and password.

Import the old DB

Ready?
```
wp db import example_com.sql
```
Boom. It’s all dumped in! Only two steps left!

Search & Replace

I love this one.
```
wp search-replace example.com newexample.com --dry-run
```
I ALWAYS dryrun test it. This is a serialization safe search, so rarely is it ever going to be an issue to just run, but it lets me make sure I don’t get any wonky results. I never have, so re-run without dry-run.

Cleanup!

Delete the SQL file, delete the old files on the old server.

Drink

I moved code. That’s shipping, right? Or is it margaritas are for migrations?
26 March, 2014
Category or Tag?
Everyone in WordPress has probably heard the advice of Lorelle VanFossen on the subject of Categories or Tags. I’m sure her 2005 post will remain, forever, one of her most popular posts, much like mine about why you shouldn’t use Multisite will be one of mine. And why is that? Well it’s simple.

She’s right.

She gives information in a clear, direct, concise way, and she’s right. She will always be right. No, sorry, if you disagree, you’re just wrong. Category abuse is like menu abuse in that it confuses your readers. If you’re organizing things in too many directions at once, the sense of location is lost and no one cares anymore.

But still, people ask which do you use and where and why? For me it’s really simple.
- Categories are organized
- Tags are free form
To me that means I need to organize my posts on a site to major topics. This is a tech blog, so I’m probably going to talk about how things are and how they do it, and how you do it. That means I really don’t need much more than three categories. I ended up with five, since I decided a CPT for presentations and another for videos wasn’t really needed. Categories categorize, though. It’s simple and straight forward. A category is a room in my house.

The tags are the items in my house and they can go in any room. They’re the minor topics of my site. Like my iPad can come into my bedroom or bathroom (we all do it), the post about plugins could be philosophical or explanatory. The little things versus the big things, as it were. The tags are all those little things scattered around the site.

But what good are they to anyone? They’re both aspects of organization, and they’re both somewhat useful to find old posts, but do your readers ever use them? It’s funny when you think about how much time we spend trying to make all these aspects of our site ‘perfect’ and how few people actually use them. Like I did a study on this site for related posts. I measured, using my analytics, how many people used them to click through and came out with a resounding “Less than a dozen.”

Same general test on another site? Over a hundred. And on that site, no one clicked on tags. They don’t really here either, but since I use categories to organize ‘sections’ of the site, those get used a lot everywhere. And if you can’t tell, this is all a lot of work. You’ve got to work hard to make the site flow right for your users and visitors. For your product, you have to consider what you’re sharing and selling, and how people logically get around. You need to study, watch, and experiment.

The real answer to category or tag is, of course, what works for you. But when you’re starting out, listen to Auntie Lorelle.
24 March, 2014
You’re Wrong about Open Source Development

We like to say that the ‘customer’ is always right. But when it comes to open-source products, the line between customer and developer is blurred. I joke that I’m not the owner of code, I’m the custodian, and by fielding questions from users and other developers, I turn that into a better product. There’s more and more calls for people like me by the way. A non-insignificant number of companies ask me “Is there someone else like you who would want to work for us?” because giving good support is hard, it’s a weird skill set, and it requires the ability to tell someone “I’m sorry, but that’s just not correct.”

Yes, I tell people they’re wrong a lot. When I say it, I try to couch it in more friendly terms like “I understand why you’d think that, however because of XYZ the product chose to do ABC.” Or maybe even “That would be great, but historical support forces us to do that in a way that would remain backwards compatible. It would suck if we broke everyone going forward, right?” See the point here is that you’re not right but you’re not exactly wrong either, you’re just isolated in view.

Tunnel vision is something that happens to all of us. We look at the world from our perspective (yes, I was Captain Obvious there, I know), which means when most people remark that a project needs something, what they really mean is they need it. This is the part of passion that escalates into angry and vitriol remarkably fast, by the way, so if you’ve ever seen someone go from zero to abusive in three comments, that’s often what’s going on. They really want something to the point that they see red and can’t get out of their tunnel.

Getting back to the rational world is hard, especially if you don’t really understand what it means to develop open source. You may think the developers are ignorant of their users, or out of touch, or don’t care. After all, if open source allows anyone to contribute, why doesn’t a project do everything?

Well besides the fact that it can’t do everything, there are four main reasons a project doesn’t do things the way you it to. This doesn’t mean you’re right or wrong. Being wrong doesn’t mean you are wrong. It’s pretty hard to ever hear ‘wrong’ and not take it a little personally, though. Just keep in mind the reality that most Open Source developers are way more in touch with their users than people behind iOS or Microsoft Word. They just move at a different pace.

Support

The people who write the code have to support the people who don’t (or can’t). If they don’t want to support certain code, they shouldn’t have to. After all, what if they don’t feel confident that they can!? If you ask a developer to put in a feature they don’t use and don’t really understand, what happens when it breaks? I always tell people “You can’t support what you don’t know, and you can’t know what you don’t use.” This is why everyone who’s been through my WordPress training is pushed to actually use WordPress. Supporting something is so much easier if you use it. Thankfully all WordPress developers use the product every single day, so they know what it’s like.

Complications

The code is something everyone wants, but it’s too damn hard to code and remain backwards compatible, which is a huge deal for WordPress. A good example of this kind of thing would be WordPress Multisite’s shift from using /blogs.dir/ to /uploads/ for storing uploads. Doing this allowed us to dump MS Files and speed up WordPress because we’re no longer routing images through PHP (lots of benefits there). It came at the cost of losing the ‘hide’ effect of the /files/ URL, but you weren’t really fooling anyone about that anyway. Point being, we had to do this in a way that didn’t break everyone on an older design of WP! That took a lot of time!

Time

It takes a lot of time to get code right. So maybe they’re actually working on it, but it’s going to take a long time and it’s not done yet. Open Source moves at the speed of imagination and passion, so if a developer has the time and the itch, things get done. Some tasks are pure drudgery, which brings us to …

Feelings

If you’re stuck between writing code you like to do a feature you want and writing code someone else wants and you don’t have an investment in, you’re probably going to do what you want. This is the reason that’s hardest to understand, and it’s the one most people call ‘unprofessional’ because it boils down to “Oh you don’t like something so you’re not doing it?” If this was iOS or MS Word, yeah, you’d get fired. But this is Open Source, and the rules are a little different here. We make what we make out of that same passion you have to see what you want to see.

So … why not what I want?

Because not yet. Maybe never. But WordPress was built to be extendable, not to be everything for everyone all the time. And that’s the beauty of it. But that’s another post altogether.

21 March, 2014
I Don’t Care If You Use Multisite

I love Multisite. I think it’s awesome. It makes my personal life way easier, it helped me write a book and get a job, and it’s a totally cool aspect of WordPress that is still underdeveloped but a very hard worker. Multisite could be so much more, and it’s still that new kid on the block, but it’s growing every day.

When I tell people over and over again that they shouldn’t use WordPress Multisite, I do so knowing there’s a good time to break every reason I said not to use it. After all, why not use it?

But there’s one big question I don’t answer in the WordPress forums anymore, and that is “Should I use WordPress Multisite?” Oh I may clarify specific moments in what I said in other places (like you can duplicate content, but I don’t believe you should), but I won’t get into conversations of “Is this a good use case?” The reason why is that I cannot tell you, quickly, if you’re a good candidate for being a WP Multisite Admin, nor can I glance and say that your situation and environment is perfect for multisite. And the reason for this is that Multisite is complicated.

Actually that’s wrong. The reason is I don’t know, and neither do you.

Do you remember, way back when you started with websites, and you first looked at WordPress and asked yourself “Is WordPress right for me?” Someone probably said “It can be.” or “Maybe.” Oh I’m sure someone said “Of course!” And another person said “No way!” But the point of all of those answers is that it depends.

When you consider using Multisite, or WordPress (or anything) for that matter, you should take stock of what you want to do. Does the product, natively, support those things out of the box? Do people complain that doing something is really hard? Do people say it’s weird, but works? Do people suggest plugins? None of those things are reasons to not use the product, whatever it is, but it’s getting an understanding as to what world you’re stepping into that is important.

But I know, I just know, one of you will post a comment here “Mika, I just want to know if I should use Multisite! Please tell me!”

No.

It takes me about an hour to determine if you and your current site are a candidate for Multisite (and yes, I have rules). Then there’s another two to three hours going over your specs (and probably pushing to get more details) and making sure this can’t be done easily with a standard WordPress install, that it can be done with WordPress, if there are extant plugins, and if there’s custom work that’s going to be needed.

Basically it’ll be about half a day spent to give you the answer you deserve, and even then, it’s not going to be the full answer because you’ve probably forgotten to tell me something like “I need non-logged in users to be able to upload files.” or “This will be on Windows IIS.”

Hopefully at this point you’re thinking that’s fair of me. Four hours of my time is worth around $1000 USD (my going rate is $250/hour when I’m available, which is rare), and while I’d be happy to do a flat-fee for that sort of thing, no one in the history of ever has come to me with a clean and precise list of what they need and why and how they plan to grow.

Which brings me back to why I just can’t do this anymore. I just don’t know. I don’t know you, I don’t know your skills. I don’t know the skills of your users and I don’t know your future plans. I don’t know if you’ve compared the options yet. I don’t know your feelings about custom code and a lot of plugins. I don’t know your requirements. I don’t know if you want to have multiple sites with no admins, or a network with a hundred super admins (please don’t do either of those things). I don’t know your SSL requirements, your domain mappings, your servers, or your host. And because I don’t know those, I can’t answer the question fairly.

I don’t care if you use it or if you use multiple separate sites. I do care that you understand that multisite is complex and I care a great deal that you’ve researched your options. I care that you ask questions like “If my database gets too big, can I split it with Multisite?” (Yes, use HyperDB or SharDB.) I care that you consider “Is a custom post type slower or faster than a separate site on a network?” (Neither, as it happens. They’re about the same.) I get really happy if you ask specific questions, actually. You know, the ones that tell us you’ve looked into this and thought about it.

The answer to “Should I use Multisite is?” is a question.

“Are you prepared to use Multisite?”

19 March, 2014
Embedding Videos from TVGuide
This one goes out to my fellow fansites.

There are a lot of you guys who use WordPress (probably that’s my fault) and like me, you bash your head in when a news source makes a video embedable, but only if you use their code. Why is this a problem? Most of us like using WP’s visual editor, and that means we have to switch to text mode, paste this in, and then never ever ever visit that page in the visual editor again:
```
<script src="http://player.ooyala.com/player.js?width=600&amp;amp;video_pcode=VlajQ6DTdv9-OYPHSJq6w4eU0Bfi&amp;amp;embedCode=NwdzM3aDp4BB3-MEdPemlMJK5XH7ZVdn&amp;amp;height=337&amp;amp;deepLinkEmbedCode=NwdzM3aDp4BB3-MEdPemlMJK5XH7ZVdn"></script>
```
Augh!

In part this is because WordPress uses TinyMCE for editing GUI-like, and switching between visual and text mode kills things. But on the other hand, how come we can just paste in a YouTube URL and that works? Well it’s a magical thing called Embeds, and basically WordPress sees that URL for YouTube and knows “OH, I should be a video and YouTube has a special code on it’s sever to tell me how to display it! Yay!”

Not everyone has that. TV Guide, CBS, etc. I’m looking at you. Where this makes sense is where they do not want you to embed a video because they want the traffic to themselves, which is fine. We can link. But where it’s daft is when they make it embeddable, give you that link, but it’s flipping ugly and can’t be used in the visual editor!

For TV Guide, at least, there’s a fix. Ooyala Video Browser.

While the search function doesn’t work ‘great’ right now, probably because I’m searching for their videos with the wrong keywords, but still I was able to embed a video like this:
```
[ooyala code="NwdzM3aDp4BB3-MEdPemlMJK5XH7ZVdn" player_id="undefined"]
```
The ‘code’ is the ’embedCode’ from the embed they gave me before.

I can also use this:
```
[ooyala code="NwdzM3aDp4BB3-MEdPemlMJK5XH7ZVdn" player_id=&"VlajQ6DTdv9-OYPHSJq6w4eU0Bfi"]
```
Using ‘video_pcode’ for ‘player_id’

So this wasn’t super obvious, and personally I would love it if Ooyala had an oembed and then a ‘share this on WordPress’ link for things, but seeing as Automattic (and Pete! Hi Pete!) helped with the plugin, this may be as good as it gets for now. Maybe they could marry it into the Media Uploader, below ‘Insert from URL’ – have ‘Insert from Ooyala’. Dreams. I haz them.

There’s a lot to this plugin that I didn’t need, though. After all, since I don’t have an account, I don’t need all the search stuff, so I stripped it down to just this (which is an mu-plugin):
```
/**
 * Ooyala Shortcodes
 * Usually for crap like TV Guide
 * Example: [ooyala video_pcode="VlajQ6DTdv9-OYPHSJq6w4eU0Bfi" width="222" embedCode="NwdzM3aDp4BB3-MEdPemlMJK5XH7ZVdn"]
 */
add_shortcode( 'ooyala', 'ooyala_shortcode' );

function ooyala_shortcode( $atts ) {
	extract(shortcode_atts(array(
		'width' => '500',
		'video_pcode' => '',
		'embedcode' => '',
		), $atts
	));

	$width = (int) $width;
	$height = floor( $width*9/16 );
		
	if ( !is_feed() ) {
		$output = '<script src="http://player.ooyala.com/player.js?video_pcode='.$video_pcode.'&amp;width='.$width.'&amp;deepLinkEmbedCode='.$embedcode.'&amp;height='.$height.'&amp;embedCode='.$embedcode.'"></script>';
	} elseif ( $options['show_in_feed']  ) {
		$output = __('[There is a video that cannot be displayed in this feed. ', 'ooyalavideo').'<a href="'.get_permalink().'">'.__('Visit the blog entry to see the video.]','ooyalavideo').'</a>';
	}
	return $output;
}
```
And that is that!
17 March, 2014
Not The Bug I Expected
“Comments aren’t working,” he said and I winced. I hate when people say ‘broken’ or ‘not working’ because I don’t magically know what they mean by ‘working.’ But this guy was a coworker whom I like and trust so I asked what he meant by not working, and he said “When you enter a comment, the page reloads, but there’s no comment.”

“Okay,” I muttered to myself. “Touché, Mika.” Because that, indeed, was ‘not working.’ We turned on debugging and got a strange error:
```
WordPress database error: [Duplicate entry '0' for key 'PRIMARY']
INSERT INTO `wp_comments` (`comment_post_ID`,`comment_author`,`comment_author_email`,`comment_author_url`,`comment_author_IP`,`comment_date`,`comment_date_gmt`,`comment_content`,`comment_karma`,`comment_approved`,`comment_agent`,`comment_type`,`comment_parent`,`user_id`) VALUES (507,'lori','admin@example.com','','12.248.40.138','2014-02-07 23:07:08','2014-02-07 23:07:08','test',0,'1','Mozilla/5.0 (Macintosh; Intel Mac OS X 10.9; rv:26.0) Gecko/20100101 Firefox/26.0','',0,1)
```
And then it was a mess of ‘Trying to get property of a non-object…’ and ‘headers already sent.’

After some kicking of the usual suspects (reinstall WP, remove plugins, theme, etc) I realized I was stupid and the problem was that the table wasn’t updating the primary entry. It should start at 1 and move up to 2 and so on. Since everything else was working, I compared the table to a working one … and they were not the same. Here’s a good one:

There’s had comment_id as the name, and the rest of the values were blank. So I restored the values and it worked! Except … I was getting a new error.
```
Notice: Undefined property: stdClass::$comment_ID in /home/wp-includes/comment.php on line 155

Warning: Cannot modify header information - headers already sent by (output started at /home/wp-includes/comment.php:155) in /home/wp-includes/pluggable.php on line 896
```
I stared at that for a moment and after I had given up and proposed we reinstall WP and copy tables over, my brain kicked in and I said “Wait a second, isn’t that comment_ID and not id?” I changed it on a whim and shouted “HOLY **** I AM GOOD!”

Yes, it was in all caps. Yes, it was said to my coworker, who said I have his vote.

That was, by far, the weirdest error I’ve seen in months. And the moral of all this? You can’t know everything, but if you read and pay attention and are willing to experiment, you can solve pretty much anything.
14 March, 2014