> whenever I start a new project, I spin up a new VM with a small hard drive. That way, my development environments are even more isolated from each other than if I was using something like virtualenv.
+1 to this. Use the same OS for your dev environment and your production one. In many cases, this means Linux. This simplifies all the dependency setup trouble, and gives you *much* closer to the same setup in both environments. I tend to use vim over ssh to edit, but if you've got a favorite desktop editor, Nimret's suggestions of SSHFS or your virtualization software's shared folders will make that pretty seamless. On Mon, Jul 22, 2013 at 12:44 AM, Rohit Patnaik <[email protected]> wrote: > I personally run Windows, and I haven't had any trouble using VirtualBox > for all of my development needs. I just keep a Ubuntu Server 12.04 (LTS) > CD-image around, and, whenever I start a new project, I spin up a new VM > with a small hard drive. That way, my development environments are even > more isolated from each other than if I was using something like > virtualenv. In addition to having different versions of Python and > different Python libraries, I can have different system-wide C libraries as > well. > > As far as editing code, I either edit on Windows (using Sublime Text) and > scp it over to the Linux VM with WinSCP or, if it's a quick job, I just SSH > into the Linux VM and edit the code directly with vim. > > For me, at least, this setup allows me to have the best of both worlds: > Windows for games and Linux for development. > > -- Rohit Patnaik > > > On Sun, Jul 21, 2013 at 11:42 PM, Matt S. <[email protected]> wrote: > >> Interesting suggestions. Thanks! >> >> Having not had a Mac since about 2005, I'm curious if having one really >> is the silver bullet for having it all. As appealing as the walled garden >> sounds, the price tag keeps it on the periphery off my radar. >> >> Vagrant wins the awesome name and dope design award. >> >> For now it looks like a simple request to support will resolve my SciPy >> issue. But I'll be very interested to see if virtualbox or vagrant resolve >> my bifurcated (personal) computing crisis. >> >> >> On Fri, Jul 19, 2013 at 11:40 PM, Nick Bolten <[email protected]> wrote: >> >>> If you don't have root and have linux skills, gentoo prefix is a great >>> option http://www.gentoo.org/proj/en/gentoo-alt/prefix/. It basically >>> gives you a linux installation (though arguably the most difficult version >>> to use) in your home folder. I've used it many times to get scientific >>> python packages working (like installing RabbitMQ for Celery). >>> >>> -Nick >>> >>> >>> On 07/19/2013 11:24 PM, Nimret Sandhu wrote: >>> >>> I assume requesting 'sudo' access is not an option? even with being >>> locked down into specific commands? >>> >>> I run various versions of windows on linux via >>> https://www.virtualbox.org/ mostly for things that still only work on >>> windows (eg netflix) .. just got OS X snow leopard running on it (bought a >>> disk). Simultaneously. In hardware emulation. Trying to figure out how to >>> upgrade to Mountain Lion on there right now. Apple obviously doesn't make >>> it easy to do these sort of things. The shiny new mac laptop they got me at >>> work is nice and all (wayyyy more slick than windows) but I am a lot more >>> productive on linux and my linux laptop >> the mac hardware + price/perf >>> wise. >>> >>> if you want to step up from Virtualbox check out >>> http://www.vagrantup.com/ for firing up on demand VMs. you can also use >>> that with a specific VM image that contains specific distros/libraries etc >>> and redistribute/reuse/share it. >>> if u really want to get fancy, hook that up to rackspace, aws, etc for a >>> cloud on demand for HPC. >>> >>> btw you may also want to check out https://www.scientificlinux.org/ >>> >>> cheers, >>> - >>> Nimret Sandhu >>> http://www.nimret.org >>> >>> >>> On Fri, Jul 19, 2013 at 4:15 PM, Matt S. <[email protected]> wrote: >>> >>>> Feeling the pain as we speak... trying to figure out how to install >>>> SciPy in a virtualenv on a system that I don't have root privileges on (and >>>> therefore can't use a package manager rather than compile from source some >>>> significant dependencies). Funnily and thankfully though, this group, >>>> Continuum >>>> Analytics <https://store.continuum.io/cshop/anaconda/>, was mentioned >>>> the other day in the Training thread, and just moments ago I happened to be >>>> looped back to their site while the desperation of not making much progress >>>> was setting in. Turns out they make a Python distro called Anaconda? >>>> Looks very cool and pretty tricked out for number crunching and free to >>>> ALL and without any root requirement. I haven't tried it out but it looks >>>> like it could be a huge help to many. >>>> >>>> In the past, at times, I used the Enthought Python distro but got >>>> nervous that I was being too coddled by it (and b/c it's only free to >>>> academics). Nowadays I mostly go the DIY route (using virtualenv/pip as >>>> much as possible). But boy does it get dicey when you don't have root >>>> privileges (on Linux and therefore can't use a package manager), or you're >>>> on Windows (especially when trying to do the next task--see next), or >>>> you're trying to get a package with a lot of dependencies installed into a >>>> virtualenv (or non-standard place). >>>> >>>> Getting a little off topic now... >>>> >>>> Lately I've been experimenting with Ubuntu and trying to take a >>>> Windows diet (figuring out 8 was not my first choice). But it's damn hard >>>> to take the diet too far b/c neither system is perfect for all tasks (and >>>> lack of iTunes and Picasa in Linux was a big shock to my routine). But >>>> despite some inconvenience in my attempt to figure out how to optimally use >>>> multiple OSs for work and play, it's been good to find Git for Windows >>>> (msysGit) and delve into Pandas (which precipitated wanting to get SciPy >>>> installed). What's not been good is a lack of closure on figuring out why >>>> I can no longer paste from Windows into an X window. Realizing that I >>>> can't use Ctrl+C in ipython on Windows has been another pain. Out of >>>> curiosity, do any of you simultaneously use a Windows and Linux machine >>>> (without or without Wine, etc.)? >>>> >>>> Cheers, >>>> Matt >>>> >>> >>> >> >
