These two blog posts describe a way to debug failing regression tests using the same VM image that Travis-CI uses. It's a great idea, but the download link given there is out-of-date: the .box files they link to are 32-bit images, and Travis-CI now uses 64-bit images.
Where can I download the 64-bit images that Travis-CI now uses?
Update: Just in case it's useful: These days I use CircleCI for continuous integration, which offers easy-to-use ssh access to the build container. That makes debugging a troublesome CI setup way easier. Now there's no need to replicate the CI environment locally, as I was trying to do when I originally submitted this question.
We are no longer using Vagrant for our backend, and as such we aren't maintaining the Vagrant images. We're looking into a way of doing this, but for now you can email us at support AT travis-ci.org and we can spin up a debug VM for you if you need to debug an issue.
Related
This is my current work-flow:
I have a web-server on my development machine (on the same network). It houses my project & I use Notepad++ to make live-edits to the code. I make code edits from my laptop, refresh the page (which I am accessing thru hostname on my laptop) to see my PHP/HTML edits & when I'm satisfied, I merge to the master branch on GitHub.
In an effort to become more familiar with IDEs for PHP, and have some great debugging capabilities, I want to start using PhpStorm.
I thought of moving the web-server & GitHub Desktop to my laptop and just leaving the databases on the development machine, but that creates other issues.
My work-flow might not be modern. Could you help me understand how a new & similar work-flow could potentially be setup with PhpStorm in the mix? How have you seen it done?
I'm using Laravel which has a ton of files, so constant full-syncs instead of deltas would be too much time wasted.
I've tried to run the HelloWorld example from the get started guide in an eclipse python project but get errors.
Should I be able to do this or can I only use the virtual environment at the moment?
Mez63
Short answer -- we did not spend time on the Eclipse developer scenario. Probably better to stick to virtual environments unless you know well how to configure Eclipse.
I'm looking for ideas on how to easily build binaries for common platforms for a Golang project, for release on Github.
I already know how to manually do releases, using Github's instructions at Creating Releases. And I'm currently doing releases using aktau/github-release, but this requires manually logging into different machines (OSX, Linux, Windows) and doing the release.
Benedikt Lang has blogged about using Travis-CI (which I'm yet to experiment with). But I presume the public Travis will only build binaries for Linux.
Any suggestions?
You are most likely right about the Travis CI only building in Linux as go cross-compilation requires you to build from source and build other go executables. Like RoninDev suggested, I would suggest that you setup your own cross-compilation build environment by following the blog post:
http://dave.cheney.net/2012/09/08/an-introduction-to-cross-compilation-with-go
It is quite easy, and only takes about 10 minutes. After you have that, a build tool such as Jenkins will give you the controls to kick off a build for the desired platforms (Mac, Windows, Linux, etc) and then push out git releases for each one.
Thanks for all the suggestions and answers that were given. Cross-compilation was what I was looking for, and Dave Cheney's blog post was a good start.
However I found a better solution - laher/goxc - "a build tool for Go, with a focus on cross-compiling, packaging and deployment". It was inspired by Dave Cheney's work, and also includes deploy tools for Github - just what I was looking for. (For example, I used it to release soniah/awsenv)
I created goreleaser to do just that.
You can try it out if you want :)
I am in the need to set up eclipse in a way that I can connect to a SVN and checkout projects or files to a remote location. The remote location is Linux-based, the clients work with windows.
I read a few threads and it seems that it works on console with ssh+svn. But I am struggling badly to make this scenario run in eclipse.
Any hints? I appreciate your help.
Philipp
Your question sounds to me that you try to solve something, that we don't know yet. So I speculate here a little bit, and I will change my answer if the question gives indication that I was wrong.
(Part of your) development has to live on the server, so there are resources you have to use during development, which are necessary for development.
Possibly these resources are (only) necessary for testing (unit tests?), or for functional tests.
You have experience with Eclipse and want to use that.
So here are sketches of possible solutions that may work for you.
Using Eclipse on the server
You install an appropriate eclipse distro on the linux machine you have to develop on.
You install locally e.g. Cygwin with the XWin packages that allow you to start an X-windows server locally.
You open up an xterm locally (just to get the display variable correct).
You start from that xterm the eclipse installed on the Linux machine: ssh <user-id>#<ip-of-linux-server> <path to eclipse> -display $DISPLAY
Pros and cons
+ You work on the machine and have the display locally.
+ You are able to checkout directly on the machine, no need of a local copy.
- Your are not able to work without the connection to the Linux machine.
Using Eclipse locally
There are two variants, and both are valuable:
Have the sources on the server (only)
Have the sources locally
Sources on server, Eclipse locally
The easiest way is to mount the file system of the server, so you have access to them locally through a different drive letter. Ask your system administrator how that could be accomplished.
Pros and cons
+ Everything works as normal.
+ You don't have to install Subversion on the server.
- Latency for the remote file system may be annoying.
- You are only able to work with network connection to the server.
Sources locally, Eclipse locally
That is the normal way to do it. Install Eclipse with Subversion plugin as usual, checkout from the repository, work locally (even disconnected), commit your changes.
You are then able to test by doing a checkout on the server, build the system there, and do your unit and integration tests there.
Pros and cons
+ Easier to install and maintain.
- No tests during development without a build process in between.
- Tests can only be done with commited code, not with changes that are not commited.
My recommendation
I like the solution best with Eclipse on the server, so you use everything that is available on the server, and Eclipse under Linux is totally the same as under Windows. You don't have any steps in between for doing tests, everything is done locally (on the server).
See as well the following questions (and answers):
Is it possible to work on remote files in Eclipse?
PS: What I forgot: I think svn+ssh is just a different protocol of Subversion to do the checkout, update and commit. It is in no way different to using the protocols file://, svn://, http:// or even https://.
I am looking at getting a continuous integration/continuous deployment environment set up for my windows azure project and I was wondering if anyone had managed to (can point me in the right direction to) build and deploy a windows azure cloud service using powershell and Hudson and perhaps has sample scripts.
I can get the project to build using MSBuild64 (I'm running x64 Windows 2008 R2 Standard).
I know 32 bit development works, but assume 64 bit development is better as i understand it as problems will be ironed out on my local box as opposed to once deployed in the Azure environment which i believe is 64 bit. (Please feel free to correct my thinking here)
I assume i have to get the cspack.exe file to package the deployment first as in a manual deployment via the development portal.
Ideally i would like to deploy it locally (with the development simulation) run unit tests against it (perhaps against cloud storage for integration tests), deploy it to staging (run the acceptance/bdd) tests and then switch from staging to production.
Any help with anyones experience in this which will speed this research up for me would be appreciated
Many Thanks
Mark
This may help to get you started: http://blog.smarx.com/posts/building-running-and-packaging-windows-azure-applications-from-the-command-line
Also see http://scottdensmore.typepad.com/blog/2010/04/windows-azure-deployment-for-your-build-server-part-2-deploy-certs.html.
Mark, I am looking at something similar. The best solution at this point appears to be a custom solution that uses the REST APIs.
Erick