Hacker News Reader

Do you have a strategy for doing this? Since C++ stays relatively backwards compatible I would imagine there is lots of legacy code that gets untouched but newer features would go into the new stuff - with legacy code getting replaced only for big refactorings?

View on HN

Our code tends to be shipped running on top of a relatively up-to-date version of linux. So we tend to go through the pain of testing/refactoring needed to get to the next version of C++ when we upgrade to the next stable version of linux that we're going to support.

Reply on news.ycombinator.com