#regexes

[daily-devops]

I like regexes because, like pure functional programming, saying you like them gives you a sense of superiority over others

[daily-devops]

I don't know who you are. I don't know what you want. If you are looking for technical help, I can tell you I don't have any time. But what I do have are a very particular set of regexes. Regexes I have acquired over a very long career. Regexes that are a nightmare for people like you to debug. If you leave me alone now, that'll be the end of it. I will not look for you, I will not pursue you, but if you don't, I will look for you, I will find you and I will use them in your code.

[osgux]

[nokob-liveblogs-webcomics]

Dude, I was looking at my Notepad++ regexes earlier today because my images weren't working, and doesn't this look like an emoticon? (\.[^]/.]+) [Edit: Not sure why the backslash isn't showing up, at least for me? Weird.]

Wait, I guess it works better in a fixed-width font: (\.[^]/.]+)

Hahaha, it's like… a happy bear or something? I dunno, but whatever it is, it's pretty cute. :D

So, I write all my posts in one big Word doc (per webcomic), as I've mentioned before. And to get them into Tumblr, I copy the post into a new document and save as "Web Page, Filtered", which converts it to HTML (with relatively little of the mostly unnecessary stuff that Word sticks in its generated HTML), so I can then paste it into the Tumblr post thingy.

But there's an intermediate step, which is where the regexes come in: I open up the HTML in Notepad++ and apply a bunch of find/replaces to it with a macro. This lets me create my own sort of code to add formatting and stuff to my posts.

So if I want to include an image in my post, for example, I can just save it to a particular Dropbox folder, then put [img]filename[/img] in my Word doc, and that'll be automatically converted to an <img> tag that refers to the right image. :D (I used a lot of square brackets in the syntax, like BBcode, since I'm used to that from posting on the MSPA forums.)

Take, for instance, this cool gif I did earlier. I type [img]ava_entering_spaceship.gif[/img], and a couple find/replaces turn that into <img alt="ava_entering_spaceship" src="https://dl.dropboxusercontent.com/u/40585720/Liveblog/Avas_Demon/Images/ava_entering_spaceship.gif">. One of those regexes is where that emoticon thing came from: (\.[^]/.]+)\[/img\], which looks for a period, followed by one or more of anything that isn't ], /, or ., followed by [/img]. (I'm not sure why I've got ] instead of [; it might just be a mistake, but I guess it doesn't really matter.) This regex also saves the first part so it can be referred to in the replacement, which is simply \1">, or the saved bit followed by ">.

In this case, it finds .gif[/img] and converts it to .gif">. But wait, you might say, why not just turn [/img] into "> and avoid a lot of fuss? Well, I reply smugly, the answer is that most of the time when I want to add an image, it's a png, so I actually turn [/img] into .png">. So I can just say, for example, [img]shy_drawer[/img] to refer to this screenshot, rather than [img]shy_drawer.png[/img]. But of course, I don't want to end up with ava_entering_spaceship.gif.png, so I have to detect the cases where I specify the file extension explicitly.

Incidentally, if you're curious, the problem I was having with the images turned out to be because I just upgraded from Notepad++ version 5.something to the latest version, 6.4.3. (It's much better for this, too. :D I just wish I'd realized there were new versions sooner.) The specific culprit was \[img\]([^[.]+), which looks for [img] and gets the filename after it, sans extension. More specifically, it was [^[.]: after the upgrade, I had to move the bracket over, making it [^.[], for it to be recognized as a proper regular expression.

[animpressivelookingscalar]

regexes

a necessary evil. they're *easy* to write once you know how they work - but hold onto your ass if you try to read them back again. usually you have to match text in one go, and do all your manipulation in one line. this makes regexes a maintainability nightmare. they only really work against regular languages. for example, don't try to use regexes, ever, with HTML. it will never work reliably. (unless you're using a modern language with backreferences, but still.) i think jordan sissel's "grok" library is a good way out - at least you can break up your regexes into macros, then match with macros that actually make sense. but really, fuck the man