Em sex., 18 de jul. de 2025 às 08:16, Dave Chinner <david@xxxxxxxxxxxxx> escreveu: > > On Thu, Jul 17, 2025 at 02:34:25PM -0300, Marcelo Moreira wrote: > > Given that the original `strncpy()` is safe and correctly implemented > > for this context, and understanding that `memcpy()` would be the > > correct replacement if a change were deemed necessary for > > non-NUL-terminated raw data, I have a question: > > > > Do you still think a replacement is necessary here, or would you > > prefer to keep the existing `strncpy()` given its correct and safe > > usage in this context? If a replacement is preferred, I will resubmit > > a V2 using `memcpy()` instead. > > IMO: if it isn't broken, don't try to fix it. Hence I -personally- > wouldn't change it. > > However, modernisation and cleaning up of the code base to be > consistent is a useful endeavour. So from this perspective replacing > strncpy with memcpy() would be something I'd consider an acceptible > change. > > Largely it is up to the maintainer to decide..... Hmm ok, thanks for the teaching :) So, I'll prepare v2 replacing `strncpy` with `memcpy` aiming for that modernization and cleanup you mentioned, but it's clear that the intention is to focus on changes that cause real bugs. Thanks! -- Cheers, Marcelo Moreira
