On Sun, Mar 1, 2026 at 12:07 PM Dag-Erling Smørgrav <[email protected]> wrote:
> Enji Cooper <[email protected]> writes: > > I wasn’t aware that their version didn’t really work; the RCS Id in > > the file was ancient, hence the question about reconciling with > > OpenBSD (I thought there was a chance that we could maybe leverage > > some of their code/fixes in the future to reduce maintenance effort). > > I think the opposite would be more likely. The only functional changes > to OpenBSD's diff3 since the version recorded in our copy was to fix > issues with inputs that don't end in a newline. > > > About the sysexits change: is there a reason why it was made? I’m > > asking because another recent change made by an external contributor > > introduced sysexits use, whereas this change removed it. > > EXIT_SUCCESS and EXIT_FAILURE are not sysexits, they are ISO C aliases > for 0 and 1. They were only used in three places, while the rest of the > code used 0, 1, or 2. I'm not opposed to switching to sysexits, but I > don't really see any benefit. > Nor I. Aren't the current valutes mandated by POSIX or by what POSIX mandates diff return though? Switching to sysexits would make that harder since they don't align well with POSIX requirements. Warner
