I wrote: > snprintf is required by POSIX going back to SUSv2, so it's pretty darn > hard to imagine any currently-used platform that hasn't got it. Even > my now-extinct dinosaur gaur had it (per digging in backup files). > I think we could certainly assume its presence in the branches that > require C99.
After further thought, I think the best compromise is just that: (1) apply s/sprintf/snprintf/ patch in branches back to v12, where we began to require C99. (2) in v11 and back to 9.2, enable -Wno-deprecated if available. One thing motivating this choice is that we're just a couple weeks away from the final release of v10. So I'm hesitant to do anything that might turn out to be moving the portability goalposts in v10. But we're already assuming we can detect -Wno-foo options correctly in v10 and older (e.g. 4c5a29c0e), so point (2) seems pretty low-risk. regards, tom lane