If make the ?Text? mode better

Hi all,
I try to use eggplant to test our application for two weeks,Thanks for your help in your Forums.

Now Eggplant uses two kind of methods to identify UI elements: ?Image? mode and ?Text? mode. Currently ?Text? mode doesn?t work very well.So Eggplant is useless if UI changes are dramatic.

I hope to make the ?Text? mode better in the future versions.I think its more userful than ?Image? mode,beacuse its dramatic.

As you pointed out, the ability to dynamically generate images of text can be very valuable for a variety of reasons. Text images have proven to be useful and reliable both in our tests and for a number of our customers. We know that identifying the correct text characteristics to use can be a challenge sometimes but once those are identified the text image capability should be usable in many (but not all) situations.

At the same time, Eggplant’s Text Image Generation capability is fairly new, so if you are having specific difficulties with generated text images we’d like to know about it. Sending reports directly from the Help menu in Eggplant is generally the best way to let us know of any problems. Thanks!

Please enhance the text image capturing feature so that it can detect the font properties by itself.

Currently, appln i’m working on is Java based and Eggplant can’t capture 50% of the text. I have to spend a lot of time trying to find the right properties although i have a small script for that.


Thanks for the input. We will see if we can do something to make this easier in the future. In the meantime, your approach of writing a small script to loop through different font sizes and/or other text characteristics, looking for a particular bit of text that is visible on the SUT is probably the best way to do this right now.

We have recently discovered at least one application written in Java in which the spacing between characters was different than the text images Eggplant was generating. We found that in this case, specifying fractional point sizes (such as 13.1 instead of 13) for the generated text images produced better results. Since your application is written in Java, you may want to give this a try and let us know if it helps.

Is it possible for the script to be posted under Examples. It would help everyone alot until its been improve in the future builds. I know it’ll help me out. Its a pain trying to find the correct font and size.


Related to this, it would be cool (and I may write a script for this anyway–probably wouldn’t be terribly difficult) if Eggplant’s Text Prefs pane would allow you to import a CSS file and generate a text style directly from CSS code.

Granted, this has it’s problems, we discovered today that despite a CSS file suggesting that the font used on a web page was Arial 10, it was actually Arial 9.1 (Thanks Firefox :lol:)

If your app isn’t using default fonts, try asking the developers what fonts/sizes are being used. I’d suggest that you request that this information be part of the deliverable when the application passes from development to testing. It’s just surprising to me that so many QA people are having to guess what fonts are being used by the apps that they’re testing.

Note: If you’re running against a Windows machine, try using ScreenOCR (www.screenocr.com) to determine what fonts are being used. The application can be kind of flaky sometimes, but it can often tell you what the font characteristics are.