Hi,
sorry for reviving this old thread (it wasn't long though). The topic came up again in IRC today.
On 03.04.18 18:47, ron minnich wrote:
So I would actually be in favor of what paul is advocating, but not the inconistency. To keep it consistent, I don't see that just using 1U everywhere is that huge a deal. C is not that pretty any more, so this little bit of ugliness doesn't strike me as a big deal. Bugs from stuff like 3<<31 scare me more.
What I just noticed, and what Julius already told us (and we'd have probably paid more attention if we weren't topic posting). The warning in question `-Wshift-overflow=2` explicitly only checks the sign bit problem, e.g. 1 << 31. The bug variant 3 << 31 is already covered by the sane `-Wshift-overflow`.
So there is absolutely nothing to gain from the =2 version by defi- nition.
Should we enable more warnings? yes! Should we enable warnings that are discouraged by their authors? probably not.
Nico
PS. This is actually not as bad as what I have seen people doing to scripts because of shellcheck. But that thought reminds me that such tools, and excessive compiler warnings, can harm software quality.
-
Nico Huber