Xtext Validate code should trigger UML model creation using EMF

[farhan5248]

The goal is to create a UML model representing the feature file layer of files. I assume validate works against one file, so it needs to just create the one class in the model.

After this is implemented, it should then check that any classes and methods referenced exist as well. If they don't, a quick fix can create those methods or classes in layers 2 and 3. This will be more useful than a build all which I think will trigger code generation across multiple files which can be slower.

A Maven goal can still be used to select files by tag and only generate code for those

Fast validation is for syntax errors or line errors Normal validation is for semantic or scenario errors Slow validation is for layer errors such as something referenced in layer 1 doesn't exist in layer 2.

[farhan5248]

The goal is to make sure that a UML model can be created representing the feature file layer of files. I assume validate works against one file, so it needs to just create the one class in the model.

[farhan5248]

So I finally figured out how to actually include 3rd party jars in my Eclipse plug-in project. I'll need to use the Maven dependency plug-in later to copy stuff from the repo into the project for the build.properties. For now I'll just paste the extra jars in the lib directory.

[farhan5248]

This is taking a while because I'm still in jar hell trying to get a hold of all the compatible Eclipse jars I need to document how to take the jars from an Eclipse installation and put them in a local Maven repo. Basically you have to get the jars from the plugin directory. Then you have to make sure that the UML ones are compatible by downloading them from the Eclipse download site for the specific version of Eclipse you're using or just look in the local plugins directory

[farhan5248]

TODO Run feature test, the code is setup. I haven't run it yet but bet it will fail 😛

In general, I will make validation and transformer tests for each parser test:

1. In the transformer project a. First write the scenario b. open the feature file in the xtext editor and make a note of bugs/improvements c. Then make the keyword and test automation code d. update the transformation and validation code in the project itself
2. In the validator project a. convert the scenario into a validation scenario b. port over the validation logic for both the plug-in and service tests c. update the maven and xtext plug-ins with the latest validator

[farhan5248]

Validation TODO

Xtext

• [x] - replace \r\n as EOL with just \n which is what is the default when using UTF-8 in eclipse
• [x] - Background statement validation isn't happening.
• [x] - And has no validation
• [ ] - keywords in doc string break syntax validation (image below)
• [ ] - Validate file name is same as feature name

Validator

• [x] - Add container(project/application/service/batch-job/plugin), path/object-name type(page, file, configuration, resource, goal), path/section is as follows
• [x] - include positive and negative words like is and isn't present

[farhan5248]

Handle background and scenario outlines differently from previous attempts at work. Don't modify the Feature object itself, instead: Expand scenario outlines, loop over scenario for each row of data Append background, to all interactions if 1st is background

[farhan5248]

Working on "replace \r\n as EOL with just \n which is what is the default when using UTF-8 in eclipse" because the parsing is failing for any tests

[farhan5248]

That's done, now just trying to find a good UML editor to view the model when debugging

[farhan5248]

Finished the test automation code to test the mbt-transformer plugin. Will now work on the mbt-validator changes before testing the Scenario conversions.

[farhan5248]

2. In the validator project a. convert the scenario into a validation scenario b. port over the validation logic for both the plug-in and service tests c. update the maven and xtext plug-ins with the latest validator

Will now work on the mbt-validator changes (2 above) before testing the Scenario conversions.

[farhan5248]

mbt-validator code updated, adding more first layer tests.

[farhan5248]

• [ ] Test semi colons in the formatter. Convert Scenario formatting is buggy

[farhan5248]

Step conversion code done. Working on Background and Outlines next.

[farhan5248]

This is done, there are things left like applying scenario outline example data to data tables but that's only needed when I do MBT stuff so I'll skip it for now