Is there an easy way to test some web page looking for some errors. I mean usually while doing web testing we want to check the display which could be very difficult to maintain.
So I wanted to know if as an alternative strategy there are some practice to massively test the URLs of a website and just look for any kind of error including JS error in the console.
I think it is possible using a framework like Selenium but it might be a bit overkill no ?
Also the idea will be to do that on a production server, in addition to test server.
The website have some authentication so just hitting the URLs will not work.
Related
I am new to Golang. And I am looking for automating signup, login processes in a web app. Please suggest a good tool like Selenium and how can I implement it in the go language.
I want to do the following process automatically using Golang:
Start a browser. Currently, I'm using https://github.com/skratchdot/open-golang
Auto entry on the signup page and auto-submit a form.
Login check for the registered user. Everything needs to be done automatically for more users.
You can also use Playwright for Go, which is a wrapper for the Playwright project. Playwright provides a single API to automate Chromium, Firefox, and WebKit to automate browsers which was created by Microsoft. With it you interact with the sites, record videos, make screenshots, and emulate other browser specific behaviour.
If you are going to use GO for web automation testing - Selenium is a good option. Still it's nothing more than a library that allows you to interact with browsers. So you are going to need to develop your own framework or reuse someone already implemented.
My advice is to consider Agouti, since it supports Ginkgo BDD and xUnit Gomega. Everything else is pretty much the same from architectural perspective. You can design it like any other language binding. There are common patterns that appear over and over again in browser automation frameworks, like
PageObjects: A simple abstraction of the UI of your web app.
LoadableComponent: Modeling PageObjects as components.
BotStyleTests: command-based approach
Another good resource for building your Test framework is the xunitpatterns guide. It gives a great content overview of the patterns, smells and refactoring strategies you can use. Also look at this test frameworks tutorial. It'll help you choose the most proper solution for your case.
My guess is that you are going to need some CI server support for
everything needs to be done auto for more users.
Here is a good article how-to achieve this with TravisCI.
update:
you can use Selenium for Golang
I'm hosting an mvc4 web api and would like to know what are some good load/perf testing tools?
I just want to send an http get and make sure it isn't being cached by the server (my app) each time.
(i.e) -> http://mysrv.com/api/dev/loadtest?data=1
Thx
You should look into wcat:
http://www.iis.net/downloads/community/2007/05/wcat-63-(x86)
http://www.iis.net/downloads/community/2007/05/wcat-63-(x64)
It does exactly the kind of load testing you want.
I am about to start building a mobile site which is dynamic, working from a lot of dynamic content which must come from the database.
I have already written a REST API for the site which the IOS and Android applications are using to interact with the information.
My question is what would be the absolute best practise for building this site, would it be:
1- Make the mobile classes an extension of the existing site functions
(The downside I see here is that the mobile site would be dependant on the main site library meaning that any bad heat on the main site would also affect the mobile site)
2- Make the mobile site a completely stand alone site running from itself
(The downside I see here is that any change to the main site library will need to be reflected here so in essence we would almost be writing code twice)
3- Make the mobile site run from the REST API and standalone
(The downside i see here is just increased number of HTTP requests for the information rather than communicating with the server directly)
Each one would function normally and there wouldn't really be any problem there, coding is really not too difficult, though if I make it standalone I would need to recreate a lot of the functions from the main site and adapt them for the mobile site which isn't ideal.
Look forward to your comments! Thanks
I would go with 3rd point, but that needs to be architect well.
We will prioritize standalone application after that API, also we can have 2 way communication, any content changed on server it will coordinate with clients to get that updated.
Also I would also suggest go with Bootstrap framework, its an awesome framework and have responsive and adaptive design
I was asked to build a control-system for a Ebay-like Finnish auction-site huuto.net.
The system would reopen closed auctions by a specific rules. It would be completely external from the main site, running at an external website.
The site is however unwilling to release its API and Schema. I know no way to build such a system without knowing its API.
How do you build an internet site without its API and Schema?
You could try some form of automatic browsing: mechanize
Edit:
Examples here.
I think you're asking about building a site that interacts with another site without using a well-defined API. Is that right?
You can interact with an external site without using an official API - in order to do so, you need to imitate a normal site visitor and send your requests to the site frontend (in much the same way as a web crawler does). Tools like hpricot, mechanize and curl can help you parse the content of pages and send requests, but in doing so your system may be quite brittle. Any change to the target site might mean you have to rewrite portions of your system.
It might be possible to get the data you need by screen scraping the site. You could perform the operations you want to do by POSTing data into their forms or using a WebClient type API to make your program act like a web browser but that's likely to be an extremely brittle solution.
Honestly though, without an API, there really is no good solution.
you either need access to the database or an API, otherwise no point in even trying.
Is there a way to test logging into a site with open id using Selenium?
In Selenium all the tests live in the server, so once filled the open id URL in the appropriate field in the web page I am taken to the 3rd party web page for entering the credentials and my test can't run anymore.
Is there a way around this?
Yes - use Selenium RC. It gets around the cross-domain problem of basic Selenium Core and allows you to script against multiple sites.
I guess, technically speaking, you could include a really dumb OpenID server on your testing domain, but Patrick's suggestion of a testing framework that supports cross-domain operations sounds like a much better idea.
Although, I guess that depends on what you're trying to test. It could be that using a third party OpenID server is bad for your tests, because a change to the UI of that server could cause your tests to break. Or maybe you want to make sure that your code is interoperating correctly with that server, in which case using the 3rd party is exactly what you want to test.