For an integration test I want to start nomad in a certain working directory (nomad needs to be able to write to that directory).
Now my question is how do I define this in a way, that I can pass a writeable directory to sh_test and the whole test still stays hermetic?
Or is there a general documentation how to handle integration tests?
Found TEST_UNDECLARED_OUTPUTS_DIR and seems to work like a charm. Only bad, that I cannot seem to retrieve it from a rule!?
Related
What are the persistence options for fitnesse files? So far it seems like a file system is the only thing supported. There does appear to be an out of date database plugin. Is there anything else that is supported (S3, database, etc.)? Is there a way to control where files are persisted if using the filesystem?
I believe there is very little in that area. The location of the files can be controlled using a command line option. See http://fitnesse.org/FitNesse.FullReferenceGuide.UserGuide.QuickReferenceGuide#FitNesseCommandLINE
-d /path/to/fitnesse/root
How I've used the FitNesse wiki is as a local development tool, with the pages on the file system. Once I'm satisfied with the tests I commit them to version control (e.g. git) so that they become part of the (integration) test pipeline setup (e.g. they are run as part of the CI/CD pipeline of the project).
There is a plugin I believe that will automatically commit any save actions to Git, but I've never used that. Saving each edit action just pollutes version control in my opinion. I only want to see tests after they have been checked/completed, and that tends not to be each save.
Working on a shared wiki environment (where I would expect a non-file system approach would fit in) you run into the same problem, I expect. Developing automated tests is a development task that requires some iterations before it is 'done', and not all attempts reach that 'done' state. So using shared storage for wiki persistence creates 'noise' in the test-set: which are the tests that form the current reference set that should pass and what is work in-progress.
If you are working on a larger project where new features are developed together with their automated tests it becomes even more important to know which test changes belong to which features/changes. Having tests on the file system, in version control, allows you to develop test in sync with code changes in the same branch. This is what I would recommend.
I have a scenario that requires me to revert to a clean snapshot before or after each Coded UI test method is executed. I have researched using the TFS Lab Management API (see http://blogs.microsoft.co.il/shair/2011/12/22/tfs-api-part-42-getting-started-with-lab-management-api/) to revert to a specific snapshot as part of the TestInitialize and/or TestCleanup method, but I can only get this to work when executed locally. When executed on a remote machine I get errors authenticating to the TFS service.
My other option is to somehow do a 'foreach test in testrun' into the build process template (LabDefaultTemplate.11.xaml). I have identified the area that I think this would fit best, but cannot find any documentation on running a loop on each test.
Is this something that is possible, or is there somehow a built in method to accomplish this that I have overlooked?
To do what you propose you should switch to Release Management and create a separate test run for each of your groupings, in your case each test. You can use RM to orchestrate looping through each of your runs and executing then.
http://nakedalm.com/execute-tests-release-management-visual-studio-2013/
However running a UI test should not break your application and I would suggest that either your tests are way too long, or there is some flaw in the design of your application.
let's say I am developing 2 applications for 2 different clients which are, using 2 different database-configurations.
When using Openshift and CakePHP it is considered good practise to not store the real connection-info in the configs, but instead to use environment-variables.
That way the GIT-Repo is also always clean of server-specific stuff.
That is all fine as long as I have ONE project but as soon as there is another one, I need to override my local env-vars according to the current project.
Is there any way to avoid this? Is it possible to set up env-vars on my local machine per directory or something like that?
I am running OSX with Mamp Pro.
This may not be the best solution, but It would work. Create a different user on your local machine and then change to that other user when you need to access those other environment variables.
I create a 'data' directory in my git repo and set it to ignore. This way anything in there will be saved in the repo and sent to openshift. I place a config.ini file with all the info that I don't want in my repo.
I then manually put that config.ini file in Openshift's persistent DATA directory by using winSCP. You only have to do this when you change your config.ini.
When my app runs it detects if it's local or on Openshift and loads the config.ini file from the correct directory.
I would be interested if somebody has a better idea.
HTH
My Jenkins job runs many tests that create log files. In case of failure, I want to look at the log of the failed test. I'd rather use Jenkins web-server to do it, even have a link in the email it sends me.
Is there any plugin that can do it? Or maybe another way?
You provide few details in your question, so it is impossible to give specific advice. In a general level: this is already possible. When your test framework creates JUnit XML files with test results, the test output can be included between the <failure> and </failure> tags. Usually test frameworks should take care of this automatically, so you are probably not using a test framework and are manually generating the XML files containing test results?
I recommend you adopt some test framework. It is usually well worth the effort.
I have a JEE6 project based on Glassfish 3.1.1 that is moving beyond the "one developer prototype" stage to being developed by a team.
Each member of the team will have their own local glassfish server. I don't want each of them to have to go through all the manual steps of setting up the JDBC connection pool, JMS services, jdbc security realm, etc via the admin console, as I did when first developing the prototype. It is error prone, and plus if I want to change something I have to tell everyone what to do. I want it to be done as part of the ant build, so that it is a one-clicker, and then if I have to change something I can just tell them to do a clean to blow away the domain and then run it again. So there would be an ant task to "config-glassfish" that would somehow configure the domain for them.
Despite extensive searching, I can't seem to find any step-by-step guide of how best to accomplish this. Anyone have a link?
Would it be best to attempt to capture the fully configured domain and store that in our src repository?
Or should I instead have ant issue "asadmin" commands to create and configure the domain?
You can do all of this with the sun-appserv-admin ant task. You can find more information here: http://docs.oracle.com/cd/E19316-01/820-4336/beaev/index.html
We struggle with this kind of thing at my work too, but only with a few developers. One thing I really like is that Glassfish has the concept of a resources.xml which will cover a lot of the config. I use this to pass around connection pool configs and JMS queues and it works really well, but it might not cover all your config needs. The contents of the file are pretty much snippets from the domain.xml, and I haven't figured out everything it can do yet. http://docs.oracle.com/cd/E19798-01/821-1751/ggoeh/index.html http://javahowto.blogspot.com/2011/02/sample-glassfish-resourcesxml.html
I haven't tried other ideas since the resources.xml solves my major pain points, but you could take your domain.xml and work through any issues brought up by copying it to another developer's domain, then do variable replacement on the part of the file that need it. That way you could have ant create the domain, then overwrite the domain.xml with the newly filled out one.
Maybe there is a way you could use asadmin backup-domain
One other idea would be Chef. http://wiki.opscode.com/display/chef/Home
I ended up just putting the domain.xml into the src repository, making an ant task to copy it over to the glassfish directory, and instructing other developers that when running that ant task, they should make sure glassfish is not running.
This worked for my case...