I am using the output parameter of google test like
--gtest_output=xml:Reports\fooReport.xml
However some XML attributes like author and project won´t be exported as the tool gtest2html might expect.
Is there a property like 'author' available in gtest at all? If so, where would I assign it?
Explanation
In GoogleTest, those fields that you see in gtest2html like 'author', 'project', etc. are not available in the xml output by default. Those are custom field expected to be found by gtest2html.
However, you can add them to the same xml element using RecordProperty function provided by GoogleTest.
The final section of the documentation explains:
Calling RecordProperty() outside of the lifespan of a test is allowed. If it's called outside of a test but between a test suite's SetUpTestSuite() and TearDownTestSuite() methods, it will be attributed to the XML element for the test suite. If it's called outside of all test suites (e.g. in a test environment), it will be attributed to the top-level XML element.
So, in order to make your xml output compatible with gtest2html, you need those fields in the outermost xml element (which has the tag "testsuites").
Solution
So, in main or in a test environment (you don't have to use one, but it is strongly recommended in the documentation), before any test is started, you need to add below call:
::testing::Test::RecordProperty ("project", "MyProject");
This will then add this to the xml output:
<testsuites ... project="MyProject" name="AllTests">
(I know it is quite late as an answer, but hopefully this will help someone)
Related
Tailwind offers a feature where you can give an input element the peer class and then, in a sibling element use the peer-invalid: pseudoclass to apply conditional stylings. This is commonly used to provide helper text when a form field is not valid.
I don't see how this can be tested in JS-DOM though, as JS-DOM doesn't have access to the actual CSS when it is just rendering components in unit tests.
The information I have seen about getting JS-DOM to use CSS at all is kinda sketchy, so I was wondering if it's even worth trying to get it to evaluate a bunch of tailwind classes in a NextJS/Jest/RTL project. I'm using Tailwind 3 so it's not even like I have a big file full of generated classes I could try and pass to JS-DOM :/
So can it be done? (Note the hard fact please mods!!!)
(And, somewhat more subjectively, should it be done? Or is there a better way to approach this?)
I'm writing API tests using Postman. I'm organizing them into folders by endpoint, and subfolders by test cases within the endpoint folders. There are multiple cases for each endpoint and for each case there are post calls that set up data prior the the csubject-endpoint call that I'm making assertions against.
I already have 100s of calls in this suite. The test runner, unfortunately, does not provide the folder names in its output, so it's difficult to see at a glance which particular case I am looking at when, for example, it reports a test fail.
Is there a convenient way to obtain the folder names for a given call in its test script? With this, I could prepend the case name to the test name, and that would make my tests vastly more readable.
I think a variable containing the current request path in the folder hierarchy would be the best, but for now I didn't find such a feature.
Instead I may suggest such a workaround solution:
In each folder prerequest scripts you set a variable:
pm.environment.set("folder", "folder1/folder1.1/")
the value of folder variable you have to maintain separately for each folder.
Then on a collection level you put a collection test like this:
pm.test("location: " + pm.environment.get("folder"), true)
After running your collection in the Runner you will get the output
from collection tests at the beginning of the test results for each request
showing the folder location.
The effort of setting folder variables pays off when you estimate the results of complex tests. I used to change the names of the requests but it is even more complicated.
UPDATE:
You can also find the info in results hovering over a gray shortcut of path on the left of the request status. A tooltip displays a full path, what in fact eliminates the need of the above solution if you only want to observe the results, but the solution can be useful if you want to make some logs containing the path.
I don't think that there is anything like that from within the application - The closet I can see is the pm.info.requestName function which references the request name that the test belongs too.
This is a basic use case but you could add this to the test name to give you a 'quick glance' and what was run against what request.
pm.test(`${pm.info.requestName} - Status code is 200`, () => {
pm.response.to.have.status(200)
})
If you take a look at Newman it might have something within the summary object that you could extract, in a script, to get the folder name but I've never done this.
Closest thing right now would be this:
I believe I included this in my logged bugs / feature requests out to their team.
I have seen the [DebuggerNonUserCode] and [ExcludeFromCodeCoverage] attributes in resources and other SO questions about exlcuding code from coverage statistics, and wanted to know if it was possible to automatically add this attribute to the classes in the code generated by the Entity Framework using .NET 4.0.
Also would it need to be class level or could it be on the diagram.Designer.cs level, needing one attribute for all code generated by that diagram?
Since partial classes (which Entity Framework creates) merge attributes, extended functionality in other partial classes are also excluded if the attribute is class level in the template, it will have to be applied at the method level.
The best way that I've found to do this is using T4 (as recommended in #Craig Stuntz's answer) to:
include: using System.Diagnostics.CodeAnalysis; at the top of the file
Then apply [ExcludeFromCodeCoverage] to getters, setters and Factory methods by searching for:
#>get
#>set
Template_FactoryMethodComment
and placing them in the appropriate place.
This was made a lot easier using Tangible's T4 editor Extension for VS.
This is my first attempt and it seems to work, "your milage may vary", so complete a test run to make sure everything's working as necessary.
Trying to figure out the best way to accomplish this. I have inhereted a Django project that is pretty well done.
There are a number of pre coded modules that a user can include in a page's (a page and a module are models in this app) left well in the admin (ie: side links, ads, constant contact).
A new requirement involves inserting a module of internal links in the same well. These links are not associated with a page in the same way the other modules, they are a seperate many to many join - ie one link can be reused in a set across all the pages.
the template pseudo code is:
if page has modules:
loop through modules:
write the pre coded content of module
Since the links need to be in the same well as the modules, I have created a "link placeholder module" with a slug of link-placeholder.
The new pseudo code is:
if page has modules:
loop through modules:
if module.slug is "link-placeholder":
loop through page.links and output each
else:
write pre-coded module
My question is where is the best place to write this output for the links? As I see it, my options are:
Build the out put in the template (easy, but kind of gets messy - code is nice and neat now)
Build a function in the page model that is called when the "link placeholder is encountered) page.get_internal_link_ouutput. Essentially this would query, build and print internal link module output.
Do the same thing with a custom template tag.
I am leaning towards 2 or 3, but it doesn't seem like the right place to do it. I guess I sometimes get a little confused about the best place to put code in django apps, though I do really like the framework.
Thanks in advance for any advice.
I'd recommend using a custom template tag.
Writing the code directly into the template is not the right place for that much logic, and I don't believe a model should have template-specific methods added to it. Better to have template-specific logic live in template-specific classes and functions (e.g. template tags).
I'm creating a DB web application with grails for my company and found myself in the need to change the default scaffolding templates.
So far so good, everything gets generated with the modified templates (controllers, views, ..).
Now, however, I get a request to create some "composite screens" with functionalities and a layout that differ from the overwritten templates.
So now my question is: is it possible in grails to create one or more templates (next the the default one) and pass this template name as an argument to the generate-* commands?
Thanks in advance!
EDIT: Adding the template name to the generate commands was just an idea, if it's possible to do this a different way, I'll be happy too.
Grails commands are scripts in grails/scripts. If you follow its logic you will see two things.
1) There is only one parameter passed to the script → domain.
2) Class for generating views is DefaultGrailsTemplateGenerator. You can analyse sourcecode and check what this class offers.
Update
Link to DefaultGrailsTemplateGenerator in GitHub.
I am not sure about the generate command parameters, but if you add another .gsp page into scaffolding directory, I believe it will try to run it through generation process.
So for example I used to have a show.gsp page as well as showBasic.gsp page, which showed fewer properties.