We use the linux kernel implementation at neovim. But a libc (and possibly asm optimized) implementation would be nice too (it could avoid two runs over the data). Not that it's likely to be a big improvement, strlen and memcpy are usually some of the most optimized functions there are.
Only if the implementation details (aside from what is dictated by the spec) of strlcpy meets the originality threshold of a copyrightable work, no?
edit: Also depends on the jurisdiction, http://en.wikipedia.org/wiki/Threshold_of_originality#Exampl... - and you don't necessarily have to change your own license, a valid resolution is also to stop distributing versions of your program that incorporate GPL'd code.
If you would care to look at the code, it is obvious that this is one of the simplest ways to build strlcpy (basically a call to strlen and memcpy and some branching), perhaps I should've said linux kernel-style such that readers familiar with it would know it without needing to look. It is trivial to come up with such an implementation.
IANAL, but at what point do you need to change the license of the entire project? 1 line of code, 3 lines, 5 lines, what constitutes using code from another project? If you call memcpy and your parameters are named the same as the parameters in my project for a memcpy call, who gets to declare his license over that line and thus the project, if it so happens that you once looked at my project? The first one who "invented" it?
If they include work derivative of a GPLed work they don't have to change the license. The license just has to be compatible and redistribution has to conform to the requirements of the GPL.
This distinction is important because someone could take the non-GPL part and distribute it separately under the original license.
