Hacker News Reader

But would you be sure you've actually fixed the bug? Perhaps your edit changes the state that would be reached by the continue (e.g. bug is on first iteration of a loop, and your edit inadvertantly changes the initialisation).

I think it's useful for fleshing out a bug and could save time if there's another issue downstream, but I wouldn't count it as solved until it has been tested on a fresh run.

View on HN

Again there is no need for black and white. Take each case as it comes. Often edit and continue is enough sometimes you do need to stop and restart to be sure. Sometimes you are 99% sure and that is enough when you have 4 weeks of bugs and 1 week to ship a product. Heck sometimes you are trying to repro a complex but and half way through hit and unrelated issue, I make a sound of pain and someone else asks what's up, I tell them, they say it is fixed in latest, but I can edit and continue a fix now to save having to spend ages reproing. There are loads of times the feature is useful. I would suggest the people finding it amateur and not useful aren't using it correctly or at sensible times.

That's easy to do. After having the fix, go from the beginning and see if it still happens. Of course, this flow does not apply to all kinds of bugs, but I'm willing to bet that at least 90% of them can be fixed like this.

Reply on news.ycombinator.com