4 ms·

I've been working on fixing some old C++ code at my company so that it compiles with the most recent version of the compiler. To my surprise, the errors that th

by TideAd 8y ago

I've been working on fixing some old C++ code at my company so that it compiles with the most recent version of the compiler. To my surprise, the errors that the new compiler gives me are very reasonable. But I always ask myself, "how on earth did this ever compile in the first place?" The crazy part is going into the commit logs and seeing that a lot of the people who wrote this have been very successful. They've become vice presidents, retired rich, and written influential papers.

flukus 8y ago
I've been doing the same, porting (getting to compile and seemingly run) an old c and c++ codebase from 32bit solaris to 64bit linux. For a while I was spending my days alternating between "how did this ever compile" and "how did this ever run" when debugging. It took a lot of modifications to get it to compile on a now 6 year old version of GCC, then I got to debug a lot of issues with type sizes, endianess and posix differences. The team responsible for it is still in favor of keeping it 32bit unfortunately, so we can't even do a desperately needed valgrind run. I wish I could look at the history but I don't have access to the clear case repository I think they're still using. We've got another codebase of similar vintage that is preventing us from deploying spectre patches...