karl@ra.umb.edu ("Karl Berry.") (11/08/88)
What should we do when we find what appears to be a bug, but the source code that produces it is part of quite a large program? In some, perhaps most cases, a small test file can produce the same bug. But in a few, I can't localize it that way. I know that gcc.texinfo addresses this, and says ``send the bug in anyway''. But is an umpteen-thousand line program really feasible to send? How about the binary plus the file where the actual bug occurs? I want to do something; after all, it can't be fixed if it's not reported. My strategy until now has been to hope that someone actually at the FSF will find it, and it will get fixed. This has actually been a pretty successful strategy, but I'd like to do something more active! Karl. karl@umb.edu ...!harvard!umb!karl