I did not use the provided "testPhrase" method for the validation of the user's CAPTCHA input string because I thought that this would be exaggerated only to compare two strings. This sometimes led to unsolvable CAPTCHAS, since one of the randomly selected fonts contains only uppercase letters. For example, if the generated phrase is "AbCdE", it becomes "ABCDE" in the CAPTCHA image. If you use the "testPhrase" method to validate the user input, this is not a problem. But if you do the validation yourself, as I did, you can end up with intractable CAPTCHAS.
The documentation states: "You can compare the phrase with user input". I think it would be better to write "must" or "should" here instead of "can" and explain why. A hint that a self-implemented case-sensitive string comparison leads to serious problems due to the upper-case font used would not be bad at this point either.
I did not use the provided "testPhrase" method for the validation of the user's CAPTCHA input string because I thought that this would be exaggerated only to compare two strings. This sometimes led to unsolvable CAPTCHAS, since one of the randomly selected fonts contains only uppercase letters. For example, if the generated phrase is "AbCdE", it becomes "ABCDE" in the CAPTCHA image. If you use the "testPhrase" method to validate the user input, this is not a problem. But if you do the validation yourself, as I did, you can end up with intractable CAPTCHAS. The documentation states: "You can compare the phrase with user input". I think it would be better to write "must" or "should" here instead of "can" and explain why. A hint that a self-implemented case-sensitive string comparison leads to serious problems due to the upper-case font used would not be bad at this point either.