On 29 Apr 2026 16:05, Alexander Grund via Boost wrote:
Am 29.04.26 um 13:34 schrieb Andrey Semashev via Boost:
On 29 Apr 2026 12:16, Dennis Luehring via Boost wrote:
Am 29.04.2026 um 11:02 schrieb Andrey Semashev via Boost:
Personally, I never use -Werror and don't consider this a good practice. rational for this? does it forces too much that people try to prevent warnings at any cost/without thinking about? This topic has been beaten many times, but in short, compiler warnings are not portable and are often bogus and/or subjective. Even within the same compiler line, one version may not generate a warning and the next one will, for the same code. Requesting the compiler to fail in these conditions is counter-productive and hostile towards users.
Obviously Werror should be the users choice, not be forced upon them.
Well, it started from Boost.int128 enforcing a -Werror-like behavior on users and claiming this was uncontroversial.
I fully disagree with that they "are often bogus and/or subjective", or compiler implementers wouldn't have implemented them.
That's just my experience with warnings. I think, most of the warnings I get are bogus; I can't remember the last time I got a useful one, apart from deprecation warnings after upgrading a library. Quite a few of the bogus ones are coming from the standard library and system headers, actually. And no, headers residing in a system directory doesn't silence these warnings, at least not in my case.