Forum Discussion
Hi,
Colleagues of mine say that an exe with debug info would not potentially behave like an exe without them.
So if you use debug infos, you will always risk having errors in your non regression tests.
You may have the error of missing something thats occurs or spending time in a false-positive.
I'm scripting with TestComplete non regression tests for two big delphi apps since a few years now, and I never had the opportunity to include debug info. This is not negociable at my work, for the reasons I explained before.
- Colin_McCrae8 years agoCommunity Hero
I can only offer my experience.
As I say, the debug info is exposed by the compile options (primarily, including the BPL libraries for me). So the code base is unchanged. So functionally, it should remain the same. That makes sense to me.
m_essaid .... "Colleagues of mine say that an exe with debug info would not potentially behave like an exe without them."
I'm curious, why do they think this? Have they explained why they think it would behave differently? I'd be interested to know so it's something I can watch for myself. (As behaviour is identical for me ...) The build engineer I worked with to set up the compile options on our Delphi project didn't see any problem with compiling it this way.
Also, as I've mentioned, some of the methods required to properly interact with, and extract details from, certain controls is simply not possible with having the native methods available. The certainly applies to a lot of the grids we use. The basic published methods and properties simply aren't enough. And these controls make up a sizeable part of what goes on on most panels. Without being able to use them properly, I wouldn't be able to test much.
Related Content
- 5 years agoCooly12
- 6 years agogenki
- 5 years agoCooly12
- 8 years agorasool_saheb
Recent Discussions
- 22 hours agoMW_Didata