Date: Wed, 13 Aug 2025 11:43:44 -0600 From: Warner Losh <imp@bsdimp.com> To: sgk@troutmask.apl.washington.edu Cc: freebsd-current@freebsd.org Subject: Re: a question about style(9) and inline Message-ID: <CANCZdfqM22c1h-wvGJAGmUWi8V0ZR1=b0aO_H_i%2BZkBoF60cnw@mail.gmail.com> In-Reply-To: <aJzL8JUop1vDFPNJ@troutmask.apl.washington.edu>
index | next in thread | previous in thread | raw e-mail
[-- Attachment #1 --] On Wed, Aug 13, 2025 at 11:32 AM Steve Kargl < sgk@troutmask.apl.washington.edu> wrote: > In looking at lib/msun/math_private, one finds > > static __inline float complex > static __inline double complex > static __inline long double complex > static inline double > static inline float > static inline long double > static __inline int > static __inline int > static __inline int > static inline int32_t > static inline int32_t > > style(9) seems to not contain any preference with respect > to __inline versus inline. As a matter of consistency, > I would like to use whatever is the preferred keyword. > So, which should be used. > We generally have static __inline, though the reasons for that are historical. We originally did it to support building FreeBSD with a K&R compiler. Now, we've narrowed the scope of K&R support so we only really require it for public files since we support K&R compilers that are like how gcc implemented this (which basically is to have ansi keywords in the identifier space). In that environment, __inline is an extension. In C code, this is just a compiler extension meaning the same thing as inline. For C++ mode, we redefine __inline to inline. And we have some vestigial support for doing the same for the C compiler that doesn't support __inline. However, I did a bit of a survey just now, and more recently we've given up on that and just use a raw inline by and large. With __inline being a legacy item. In this context, though, math_private.h isn't public, so I'd just use inline. It's ancient enough that the old-school considerations mandated __inline (not least because bde favored building with such compilers). Now, I don't think it matters anymore, and we should just use the standard way of doing it. Warner [-- Attachment #2 --] <div dir="ltr"><div dir="ltr"><br></div><br><div class="gmail_quote gmail_quote_container"><div dir="ltr" class="gmail_attr">On Wed, Aug 13, 2025 at 11:32 AM Steve Kargl <<a href="mailto:sgk@troutmask.apl.washington.edu">sgk@troutmask.apl.washington.edu</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">In looking at lib/msun/math_private, one finds<br> <br> static __inline float complex<br> static __inline double complex<br> static __inline long double complex<br> static inline double<br> static inline float<br> static inline long double<br> static __inline int<br> static __inline int<br> static __inline int<br> static inline int32_t<br> static inline int32_t<br> <br> style(9) seems to not contain any preference with respect<br> to __inline versus inline. As a matter of consistency,<br> I would like to use whatever is the preferred keyword.<br> So, which should be used.<br></blockquote><div><br></div><div>We generally have static __inline, though the reasons for that are historical. We originally did it to support building FreeBSD with a K&R compiler. Now, we've narrowed the scope of K&R support so we only really require it for public files since we support K&R compilers that are like how gcc implemented this (which basically is to have ansi keywords in the identifier space). In that environment, __inline is an extension. In C code, this is just a compiler extension meaning the same thing as inline. For C++ mode, we redefine __inline to inline. And we have some vestigial support for doing the same for the C compiler that doesn't support __inline.</div><div><br></div><div>However, I did a bit of a survey just now, and more recently we've given up on that and just use a raw inline by and large. With __inline being a legacy item.</div><div><br></div><div>In this context, though, math_private.h isn't public, so I'd just use inline. It's ancient enough that the old-school considerations mandated __inline (not least because bde favored building with such compilers). Now, I don't think it matters anymore, and we should just use the standard way of doing it.</div><div><br></div><div>Warner</div></div></div>home | help
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CANCZdfqM22c1h-wvGJAGmUWi8V0ZR1=b0aO_H_i%2BZkBoF60cnw>