From halmurray at sonic.net Fri Aug 1 09:14:24 2025 From: halmurray at sonic.net (Hal Murray) Date: Fri, 01 Aug 2025 02:14:24 -0700 Subject: waf: compiler flags, uninitalized variable In-Reply-To: Message from Fred Wright via devel of "Wed, 30 Jul 2025 10:40:05 -0700." <6963e6f9-607d-3b30-4fde-b6e8bfdd8252@fwright.net> Message-ID: <20250801091424.7CFBF620170@107-137-68-211.lightspeed.sntcca.sbcglobal.net> devel at ntpsec.org said: > Usually, you won't get that type of warning unless you enable it. And > usually, you only get that type of warning in an optimized compile, since > the dataflow analysis needed to detect it is part of the optimization > code. I tried O3. That found a different maybe-uninitialized. I tried O6. Nothing new. -- These are my opinions. I hate spam. From fw at fwright.net Fri Aug 1 14:40:27 2025 From: fw at fwright.net (Fred Wright) Date: Fri, 1 Aug 2025 07:40:27 -0700 (PDT) Subject: waf: compiler flags, uninitalized variable In-Reply-To: <20250801091424.7CFBF620170@107-137-68-211.lightspeed.sntcca.sbcglobal.net> References: <20250801091424.7CFBF620170@107-137-68-211.lightspeed.sntcca.sbcglobal.net> Message-ID: On Fri, 1 Aug 2025, Hal Murray wrote: > devel at ntpsec.org said: >> Usually, you won't get that type of warning unless you enable it. And >> usually, you only get that type of warning in an optimized compile, since >> the dataflow analysis needed to detect it is part of the optimization >> code. > > I tried O3. That found a different maybe-uninitialized. > > I tried O6. Nothing new. The other thing I forgot to mention is that sometimes the control-flow analysis isn't as good in older compilers, leading to complaints about something being used uninitialized in a case where the code couldn't actually be reached in the unininitialized case. Unless there's a good reason to avoid it, usually the best fix is just to add an initializer to the declaration, even if it's not logically necessary. Fred Wright From halmurray at sonic.net Fri Aug 1 18:54:32 2025 From: halmurray at sonic.net (Hal Murray) Date: Fri, 01 Aug 2025 11:54:32 -0700 Subject: waf: compiler flags, uninitalized variable In-Reply-To: Message from Fred Wright via devel of "Fri, 01 Aug 2025 07:40:27 -0700." Message-ID: <20250801185432.18146620170@107-137-68-211.lightspeed.sntcca.sbcglobal.net> > The other thing I forgot to mention is that sometimes the control-flow > analysis isn't as good in older compilers, leading to complaints about > something being used uninitialized in a case where the code couldn't > actually be reached in the unininitialized case. Unless there's a good > reason to avoid it, usually the best fix is just to add an initializer to > the declaration, even if it's not logically necessary. I usually add a "keep compiler happy" comment to that sort of code. If I see that sort of code without a comment, I looking for it. The example that started this thread was really simple. -- These are my opinions. I hate spam. From halmurray at sonic.net Fri Aug 1 19:03:09 2025 From: halmurray at sonic.net (Hal Murray) Date: Fri, 01 Aug 2025 12:03:09 -0700 Subject: waf: compiler flags, uninitalized variable In-Reply-To: Message from Hal Murray of "Fri, 01 Aug 2025 11:54:32 -0700." <20250801185432.18146620170@107-137-68-211.lightspeed.sntcca.sbcglobal.net> Message-ID: <20250801190309.95F32620170@107-137-68-211.lightspeed.sntcca.sbcglobal.net> > If I see that sort of code without a comment, I looking for it. Looks like I fatfingered something. That should have been: If I see that sort of code without a comment, I get suspicious wondering if there might be something tricky going on and waste time looking for it. -- These are my opinions. I hate spam.