How to get Dataverse running on a Windows computer for development

[zhul-pku]

[pdurbin]

As I said on the "Do you want to develop Dataverse on Windows or can you speak for someone who does?" thread I started at https://groups.google.com/d/msg/dataverse-community/Hs9j5rIxqPI/-q54751aAgAJ for now the answer is to use vagrant up on Windows. In 24340e9 I updated the dev guide to mention the use of Vagrant on Windows. The big problem is that if you clone the git repo on Windows, you end up with Windows line endings rather than Unix line endings such that the you have to use dos2unix on the installer and other files or you'll see crazy errors like /usr/bin/perl^M: bad interpreter as I mentioned in f4f0feb when updating the dev guide.

@akio-sone contributed some code for Windows users at https://github.com/akio-sone/windows-tools-for-dataverse_4_or_later but it looks like it hasn't been updated in a while.

I'd like to know if which of the following statements best applied to Windows developers:

• I want to install all the required software (Glassfish Postgres, Solr, etc.) directly on my Windows development computer.
• I want to install Vagrant and VirtualBox on my Windows development computer and have all the required required software (Glassfish Postgres, Solr, etc.) contained within a Vagrant virtual machine.
• I want to install Docker on my Windows development computer and have all the required required software (Glassfish Postgres, Solr, etc.) contained within a Docker image.

If a different statement applies better, that's fine. Please just let us know.

[zhul-pku]

Personally, the second statement best describes my needs.
(I want to install Vagrant and VirtualBox on my Windows development computer and have all the required required software (Glassfish Postgres, Solr, etc.) contained within a Vagrant virtual machine).

In this way, I can keep windows environment for other applications and daily work which I am very familiar with, and at the same time has a virtual linux developing and deploy environment.

For the first statement, it could be a lot of work to modify the codes to fit into windows environment. For the third statement, it could be a little difficult and it seems not necessary now.

[pdurbin]

@zhul-pku thanks. Yes I agree it would be a lot of work to get the installation process to work directly on Windows. By this I mean that the installation process assumes a Linux/Unix environment to run the Perl script which calls a number of Bash scripts.

It's fine, of course, that you find Docker difficult. I passed this along at #3937 (comment) . To be honest, while I've used Docker a bit here and there, I'm much more familiar with Vagrant. But don't let me bias anyone who comes along and answers my question above! 😄

Finally, Mike Mormando had this to say at https://groups.google.com/d/msg/dataverse-community/Hs9j5rIxqPI/OZzNrQpvAgAJ

I'd like very much to work on Windows, a lot of people are quite religious about it, I'm not one of them, but all I own are Windows machines, I do both Java and C#, and I play a game from time to time.

A lot of the developers I know would be lost without Visual Studio. I know it can be had on the Mac now, but it isn't really the same from what the few I know that have tried it have told me.
I'll look up Vagrant.

Thanks
Mike

[pdurbin]

I bumped into @zhul-pku on Monday and it sounds like she's sticking with Windows, as opposed to switching to Mac. She's going to think about which issue she'd like to tackle while she's in town.

[dlmurphy]

@pdurbin and I spent some time today trying to use Docker with Minishift on my Windows 7 laptop today. After lengthy wrestling with at least 4 different errors, I'm comfortable saying that this process is more complicated and fiddly than it's worth for me, though others' mileage may vary. I would love to get a Dataverse dev environment running on my PC but so far I've had very little luck with Vagrant or Docker.

[oscardssmith]

what about the bash subsystem? Does that help?

[pdurbin]

Are there any developers on Windows who would like to try a new Docker-based dev environment? I just created pull request #4168 and I would be happy for people to try it out!