Forum Discussion
hlalumiere
12 years agoRegular Contributor
Just an opinion, but wouldn't time be better served writing a proper API that ANY build/test management tool can interface with? The main problem I have with automating TC is the oudated, rigid, and oversimplified API. It's a pain in the butt to get working properly, and when you do get it to work properly you have missing features at every corner. Stuff like accessing project or project suite variables, or pausing the running test, or running a specific test group in different configs, etc, are all impossible to do through COM right now, today. Write a PROPER managed library to interface with TestComplete, and all those problems go away. Then you can use your own library to write plugins to all those other softwares without hassle.
I just don't see how a single plugin helps the situation at all. Now you have yet another non-unified piece of code to support and maintain in the already huge mess that must be TestComplete's source code. If you had a proper API, most of the other vendors (JetBrains, Jenkins, etc) would already have written their own plugin for TestComplete, that you wouldn't even have to support.
My rant is related to the fact that I tried (and utterly failed) to integrate TC into TeamCity months ago, and the programming model for automating TestComplete left me with a very sour taste.
TestComplete does its job well, but honestly when you dig past the skin layer it reeks of bad programming practices and archaic software design. You should fix THAT before trying to add new features only a small percentage of your users will ever have a need for.
I just don't see how a single plugin helps the situation at all. Now you have yet another non-unified piece of code to support and maintain in the already huge mess that must be TestComplete's source code. If you had a proper API, most of the other vendors (JetBrains, Jenkins, etc) would already have written their own plugin for TestComplete, that you wouldn't even have to support.
My rant is related to the fact that I tried (and utterly failed) to integrate TC into TeamCity months ago, and the programming model for automating TestComplete left me with a very sour taste.
TestComplete does its job well, but honestly when you dig past the skin layer it reeks of bad programming practices and archaic software design. You should fix THAT before trying to add new features only a small percentage of your users will ever have a need for.
mgroen2
9 years agoSuper Contributor
hlalumiere I agree partially with you, regarding the hard work to get TestComplete integrated with other tools (in my case Cucumber, Lettuce). You end up keeping the framework alive while as a tester you should focus on the AUT. See my posting here