A
A
AstonMartin2019-06-07 15:04:49
Software testing
AstonMartin, 2019-06-07 15:04:49

How to organize testing of your site in a small company?

Good afternoon!
We have a small company. Sawing our site-shop and CRM for it.
It happens from time to time that it turns out that some of the functionality on the site has not been working for some time, for example, on iPhones with safari.
Question: how to properly and inexpensively organize the process of testing / site performance from different devices? To always be sure that all the functionality of the site works exactly, for example, with 90% of client devices.
It is not possible to hire a separate tester.
Maybe there are some services that, in manual or automatic mode, can test our resource at certain intervals according to given scenarios with a given list of devices?
Or is it better and cheaper to organize the process yourself?

Answer the question

In order to leave comments, you need to log in

4 answer(s)
S
sim3x, 2019-06-07
@sim3x

You need to switch to TDD with functional and unit tests.
The longer you delay with the implementation, the more difficult and expensive it will be to start.
Selenium already has the functionality to track such errors
https://stackoverflow.com/questions/32562561/selen...

4
4tlen, 2019-06-08
@4tlen

Or write full-fledged tests or outsource. If releases are rare, then it is cheaper to outsource testing.

G
grinat, 2019-06-08
@grinat

If it’s free, then the whole zoo will have to be deployed, some you won’t deploy (in the sense that you integrate horseradish into ci), write tests manually, there are utilities that simplify writing them:
https://github.com/checkly/puppeteer-recorder
https://github.com/SeleniumHQ/selenium-ide
but write complex scripts by hand. And those scenarios where something breaks on one of the browsers are usually always difficult. If only layout breaks, then this is a separate story: galenframework.com
If for money, then https://www.browserstack.com/ and https://saucelabs.com/
If you do it according to deshman, then it's easier to just write out a list of critical pages, and manually check them before release. Those. deploy in a virtual box, a bunch of virtual machines with browsers, go there and check everything. So even better. In fact, all e2e tests are complex, capricious, tend to break down periodically, for various inexplicable reasons.

S
Sergey Nizhny Novgorod, 2019-06-07
@Terras

1) TestRail or its analogue avikot - start test cases. When we roll some changes, we check all the pens according to the list.
2) Cover your code with integration tests.
3) As you grow up to an adequate size and large volumes, hire QA-automation, which will transfer tests from point 1 to automation (I will say right away that this is long, expensive, but then it saves you time and money).
4) Or hire a decoy tester who will check everything from point 1 during releases (if you are not from the capital, then there is a salary of 35-45 max)

Didn't find what you were looking for?

Ask your question

Ask a Question

731 491 924 answers to any question