-
Notifications
You must be signed in to change notification settings - Fork 5
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
GPII-1336: update schemaMiddleware
to work with more than just POST data...
#5
Conversation
…ee `README.md` for details about the list of components and their purpose.
…of local copy/paste code.
…on to better reflect how it works.
…scapes slashes and tildes in path variables.
…r pull request feedback.
…Z-Schema improvements (and advice from the package author).
…it is supported by Z-Schema.
…e installation and test cycle.
…e schema directory. Also added test content to exercise this.
… required there and breaks things on the server side.
…ad request) instead of 500 (server error).
… exposed password data.
…hould include information about the JSON schema it uses.
…ma resolve dependencies.
…ter logging when an invalid schema is supplied.
…oken all-tests.js.
… `validator` output using hints from the schema itself.
…ost-2.0.0 binder transform functions and deletion relay fix.
When we get around to this, we should review all-tests.js together. All of the tests required from all-tests.js pass individually, but fail when run together. I suspect a similar problem in gpii-express-user as well. |
…d dependencies based on output of `yarn outdated`.
…pendencies based on output of `yarn outdated`.
@amb26, once the gpii-pouchdb pull is merged, it would be good to revisit this one. It's one of the very oldest, and seemed to be trending towards completion when last we left it. I will start working on the test stability in advance of that. |
…requirements, and to document the test problems with npm 3.10.10
@amb26, after painstaking troubleshooting, I have confirmed that the test failures only occur when running the tests using |
I get the following failure when running the tests by any means:
|
OK, I fixed the problem that was preventing the launch of istanbul, but that only cleared the way to hit the next problem. I'm currently encountering "address in use" errors not seen on linux or OS X, but only when running the browser tests (they use the same harness as the node tests). I will continue investigating and comment here when I have more information. |
The "address in use" errors were a red herring (although they did encourage me to do a bit of review and refactoring of the test environments, which isn't a bad thing). It turns out that my problem was that chromedriver wasn't in the path on my windows VM. @amb26, as far as I can see, this is ready for review again. |
I've installed chromedriver using npm install -g chromedriver, and verified that it is on the PATH by running "chromedriver" from the command line. Unfortunately for some reason it is still not being picked up by the tests. I get the following error:
|
Looks like selenium-webdriver's detection of chromedriver is bullshit - it looks explicitly for "chromedriver.exe" in the path - https://github.com/SeleniumHQ/selenium/blob/master/javascript/node/selenium-webdriver/chrome.js#L189 - whereas npm install -g chromedriver only places "chromedriver" and "chromedriver.cmd" on the PATH |
I've filed giggio/node-chromedriver#90 against node-chromedriver. I suggest we just add a brief note in the README under the "chromedriver" heading that this npm module is not to be used and that the user must manually follow the install instructions at https://sites.google.com/a/chromium.org/chromedriver/getting-started to download this to their PATH |
I have updated the README with caveats about the |
…r install instructions.
I also removed the dev dependency on the chromedriver npm package. |
See GPII-1336 for details.