On Sun, Jan 08 2023, Daniel Dickman <didick...@gmail.com> wrote: >> On Jan 8, 2023, at 11:59 AM, Omar Polo <o...@omarpolo.com> wrote: >> >> On 2023/01/08 16:25:16 +0100, Jérémie Courrèges-Anglas <j...@wxcvbn.org> >> wrote: >>> https://wxcvbn.org/~jca/build-failures/amd64-clang/2023-01-03/devel/fnc.txt >> >> fails building the bundled sqlite3 copy because of: >> >> : lib/sqlite3.c:57910:7: error: variable 'nUri' set but not used >> [-Werror,-Wunused-but-set-variable] >> : int nUri = 0; /* Number of URI parameters */ >> : ^ >> : 1 error generated. >> >> don't have the newer clang here yet but this should fix the build. >> >> ok? > > > Instead of removing error, would turning off the specific check work instead? > ie. Add > > -Wno-unused-but-set-variable
I'm not sure how you are asking: op@ the OP, sdk@ (maintainer) or me, but I'll reply anyway. As for me, I'm perfectly happy with dropping -Werror in this port, and more generally in any port. My opinion is that -Werror in ports mostly written on other operating systems wastes porters time. If a maintainer wants to keep -Werror, then he gets to keep the pieces when the port breaks. For devel/fnc, Stefan is the maintainer, so he gets the last word. But if a port has no maintainer, and we strive to keep on using -Werror, who's going to fix the port the next time it breaks? Should we add a MAINTAINER-Werror variable? Don't get me wrong, compiler warnings are usually useful, and -Werror *sometimes* has value. But the ports tree is a deep jungle with lots of weird animals in it. -- jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE
