On Thursday, 12 June 2014 at 17:52:59 UTC, Andrei Alexandrescu wrote:
I very much think the opposite, drawing from many years of hacking into large codebases. I'm completely with Rob here. On a large codebase, even the slightest manual or semi-manual change is painstaking to plan and execute, and almost always suffers of human errors.

I got convinced a dfix tool would be a strategic component of D's offering going forward.

It essentially comes down to persistent compiler-as-a-library issue :( Tools like dscanner can help with some of more simple transition cases but anything more complicated is likely to require full semantic analysis.

Reply via email to