In case you didn't know, it keeps the original download around after install (and maybe other files - it has been a while since I used it). Deleting those will free up quite a bit of room.
when it works, it works great. I always get the checksum failure messages, or errors in unrar returning a non 0 status code... eventually after multiple attempts of rm -rf ~/.ievms
and re-running the script will "sometimes" eventually work... it is the best option i know of for sure, but still not as easy as at least never been easy for me here...
Not much help for people running their dev environment in a VM since running two VMs at the same time is quite horrible/impossible even on recent powerful macs.
Put the vms on an SSD and max out your ram. I can run at least two windows vms on a late 2011 i7 MBP with SSD + an express card SSD for vms, and 16 gb of ram
So for a couple hundred dollars, you can run 2 vms at once, or sign up for BrowserStack for a year or two and run waaaay more. Neat setup, but you're not selling it too well.
The cheapest option they offer is $200 a year, my option has no ongoing cost, works offline or with a limited connection, works for locally hosted sites, works for resources I access through a VPN or tunnel, etc.
fwiw, local and VPN/tunneled work - they can set up a tunnel to your machine, at which point the data is entirely in your control (there's a "local testing" category in the support area). And I entirely agree with your earlier points, and there are probably other benefits to having that kind of SSD setup.
But did your setup cost $200? And that includes hardware and software. If you had waited a year, would you have saved $200? (depending on when you bought your SSDs, those alone could be a 'yes').
First, instead of fixing their stupid browser so it works like all the others, they provide "tools" to ask that you make your sites compliant with their stuff
Second, my site uses jQuery UI 1.10 and they suggest I update to 1.9.
Regarding web browsers, Microsoft is still sucking balls, year after year.
They offer a 3 month trial of BrowserStack (who also have a free trial, but presumably its less than 3 months?) but it requires Facebook login.
When I tweeted to @ie to ask why this is the case, @browserstack replied saying "Just a way to manage sign-ups especially duplicate".
How is it possible that companies like Microsoft and Browserstack, both of whom have existing "signup" functionality, and one of whom already operates an "online profile" system, require Facebook for a trial like this?
Because it's a pain to set up a new Facebook profile. Presumably enough of a pain that people won't be willing to do it every 3 months to avoid paying for BrowserStack.
I don't have a Facebook account, so while I think the concept of Browser Stack is awesome, I'm going to skip this entirely since setting up a Facebook account just to use it pisses me off...
I didn't sign up to instagram and spotify last weekend because they seemed to want my facebook details.
No, I do not want my music taste and photographs automatically shared with a random selection of friends, strangers, and businesses. There is a slight difference between who I am and who I want other people to think I am.
Oh, I can change them, for sure. And then they can change them without telling me, and share everything anyway.
Imagine my surprise at finding out that youtube was sharing my history. Along with my mostly technical interests, there was fails and a specific interest in philology!
This sort of thing pisses me off too. Until/unless a widely-adopted global user uniqueness authentication system is developed[1], we'll continue to see web properties use third-parties to provide a semblance of this utility.
For the moment, Facebook is rapidly cornering the market. Which is too bad, in my opinion.
[1]: Random Idea: Maybe ISPs can start issuing randomly-generated cryptographic key pairs along with an IP address. This way, a particular IP + public key would be distinguishable from the next user who gets that IP address, but would not be traceable to a particular customer (not any more so than an IP by itself), and would be used only until the customer requests a new IP.
The key would provide IP address authentication ("this IP has not been assigned to someone new since the public key is the same as before.") This would all happen transparently beneath the application layer (perhaps as an extension to TCP), so existing app-layer protocols could use it OotB. Other handy features might include detecting if a machine with a static IP has had its connection interrupted since your last visit.
Each keypair could be signed using the ISP's own private key as a protection against being spoofed. Customers would be asked to choose some sort of password/phrase/favorite number/etc. as part of their service signup, which would then be used by the ISP as part of the seed for keygen system.
Yes, their response sounds like evasion. Nothing is free. Having said that, their service is pretty good (based on the original 30 minute 'free trial').
It's the only option they support. Some people choose to have no association with Facebook, for a variety of reasons (usually privacy related) - and now Microsoft says "hey that Microsoft ID we made you create for all our other services, it doesn't work here!"
Yep, all this stuff was already available. We've been using BrowserStack for ages and the VMs have been available from MS for a long time. All they've done here is package it up in a shiny looking website that's basically begging developers not to forget about IE now that everybody writes for webkit & mozilla.
OK, maybe someone can explain. Have I been doing it wrong?
I have minimal Windows VMs built on VMWare to test with everything from ie6 to ie9 and a Windows 8 VM to test ie10. Same for Windows Safari and Firefox. With VMWare's workstation software you, effectively, get one browser per tab (well, one vm per tab) and testing is dead simple.
This wasn't so hard to setup at all. I can even remote-desktop into the machine hosting the VM's and test from another machine. The only cost were the Windows licenses, but we use Windows already, so that wasn't too bad.
Why would one want to pay to use these services, particularly when, if I understand it correctly, they offer static images of each browser as opposed to a real-time interaction?
Design teams that aren't techy enough for VMs, OSes that don't support emulating X (ie, Windows users wanting to test their mobile site on an iOS device), and testing a large number of variations come to mind.
For instance, how long (and how much HD space!) would it take you to set up emulators for all these? http://www.browserstack.com/list-of-browsers-and-platforms (note that these are in-your-browser remote desktop sessions, essentially. some services are static images, like BrowserShots.org, some are not)
For most people, I totally agree, especially with e.g. the IEVMS script. Pretty simple, low cost, and it probably covers all the variety you need, since it's not too hard to install multiple versions of Firefox and Chrome on most OSes. But exceptions exist, and I'd be willing to bet they're relatively large (absolute) numbers.
As a designer, I can't really use this. It misses some of the minor design details that I can see in a real IE8 browser. It's like looking at your jpegs at 40% compression in Photoshop.
'VMs for Mac and Linux coming soon.*'
On this page: http://www.modern.ie/virtualization-tools
Ie, no more screwing around with IEVMS and conversion scripts etc when trying to get a working copy of IE.
Alas latency makes Browserstack unusable for a lot of people.