Today I learned

[NogDog]

cluelessPHP

Heh...I was just playing around with this a few days ago in a very preliminary look into a tool we're thinking of building (that I doubt will be in PHP, but it's still what I think in):

$miles = rad2deg(
  acos(
    sin(deg2rad($loc['from']['lat'])) * sin(deg2rad($loc['to']['lat'])) +
    cos(deg2rad( $loc['from']['lat'])) * cos(deg2rad($loc['to']['lat'])) *
    cos(deg2rad($loc['from']['long'] - $loc['to']['long']))
  )
) * 60 * 1.1515;

[cluelessPHP]

NogDog I'd prefer miles too, I don't get KM but that's what the tutorial was in

[Weedpacket]

...that the OED does not include the word "refactor".

[Weedpacket]

Weedpacket

...TIL that now it does define refactor. I wonder if it's because I pointed it out to them?

[cluelessPHP]

Today I learned when using geolocation to find something in a 10 mile radius, it's best not to use something over 100 miles away from current location..

[cluelessPHP]

Today I learned/finally sat down? And learnt how do make this actually work

RewriteEngine On
RewriteCond %{HTTPS} !=on
RewriteRule ^(.*)$ https://%{HTTP_HOST}%{REQUEST_URI} [L,R=301,NE] 

[dalecosp]

Well, that's good. HTTPS is where it's at these days.

Have you heard of LetsEncrypt?

[cluelessPHP]

dalecosp It's a new one to me

[cluelessPHP]

dalecosp I use it now, it was my thing I learnt today

[dalecosp]

http://www.commitstrip.com/en/2016/06/13/the-end-of-an-expensive-era/

[cluelessPHP]

Today I learned it's probably a good idea to back your code up a few times (ideally three)

[laserlight]

cluelessPHP don't you use a version control system?

[cluelessPHP]

Today I learnt there are two LA's in the US and apparently not everyone from there can understand me

Someone with a completely straight face asked me if I was speaking Gaelic at first

[cluelessPHP]

laserlight I really should in saying that I haven't been using comments either, currently I'm just debugging on my live server

[dalecosp]

laserlight

Is a VCS really a good substitute for a backup in your opinion? It has bearing on a question I've been asked here.

[NogDog]

dalecosp Is a VCS really a good substitute for a backup

Well, if it's something like Git and you push commits to GitHub or another remote location, then yes. If you're only committing it locally, then not so much. 🙂

[laserlight]

dalecosp If you have a local repository, then it is like a backup you manually make on your local computer: great for when you realise your latest code modification is a disaster, but useless if a disaster happens to your computer. If you have a remote repository, then it is like a backup you manually make on a remote computer: useful when a disaster happens to your computer. An advantage in both cases over a manual backup is that typically people make whole copies of their work when backing up manually, whereas a version control system would typically use file differencing algorithms and compression, and then it comes with a built-in log of history, and on the whole it is a less manual process so you're less likely to miss out something or include something you don't want to backup.

[dalecosp]

NogDog

Workstation -- > devel VM -- > production box.

Each is a separate machine (although devel and prod are both VM's).

I'm still a tad nervous though.

[NogDog]

dalecosp

The stuff I do at work uses Docker images, so when we update an environment, we build the appropriately named/tagged Docker image and push it to DockerHub. Then our AWS containers are configured to pull the images from there, so we do an "update" on the relevant AWS instance and tell it to "rebuild", and it pulls that image in. I know just enough about it all to make it happen, but my head spins when the guy who really understands it starts changing things and tries to explain to me what he's changing and why. 🙂

[dalecosp]

Haha, my head's spinning too.

Sort of.

I can't really imagine being able to deploy something so finished (and expected to stay that way?) that you image a disk to push it.

[laserlight]

dalecosp A Docker image isn't quite as set in stone as a disk image. If we leave the more specific benefits of containerisation aside, it's not all that different from say, a version control commit that is designated for production, that you then have your server automatically pull in the commit and make the update live when the commit is detected on some designated repository.

[Weedpacket]

I have an external drive synced to mirror the one my project repos are on, so I have a second local copy that is independent of the computer itself (I'm wondering if I should turn off write buffering so that I can yank the external from its USB cable without having to warn it first). A regular background task also syncs the repos themselves to cloud storage.