Question

This may have been asked somewhere else – unfortunately it’s quite a difficult thing to google for.

Regularly when programming I find myself with constructs of the form (I’m looking particularly for a Java answer, but I’d be fascinated by a general solution)

String a = getStringFromPlace();
a=processStringInSomeWay(a);
sendStringToSomePlace(a); 

My problem is that processStringInSomeWay(a) breaks if given some particular character, “£”, say, and so I end up writing this…

String a = getStringFromPlace();
a=a.replace("£","replacevalue");
a=processStringInSomeWay(a);
a=a.replace("replacevalue","£");
sendStringToSomePlace(a); 

but this, of course, breaks if ‘replacevalue’ happens to be in the input. I could just choose a ridiculous value of ‘replacevalue’ but that’s clearly not good practice. What is the best practice in this situation?

EDIT – this is in the particualar case where one does not have control over the ‘processStringInSomeWay()’ method. And I’m also interested in the situation were a has to be processed as one unit, it can’t be split.

Answer 1

Your question is quite generic in the sense that without knowing what you need to do with these strings is almost impossible to give one precise answer.

Solutions I see are:

• remove directly the characters that the method shouldn’t process (eg replaceAll("£","") )
• change the processing method to skip character with are not handled in current version
• split strings whenever an unwanted character is found and process them separately
• just place an escape sequence that is always the same for every symbol that is unmanaged (eg replaceAll("£",escape+"£") )

This problem is common in many situations and from what I saw you usually end up by just deciding that you have a sequence which can’t be used naturally because it’s used as an escape sequence. This is true in compression protocols, network protocols and in many other cases.