Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!emory!europa.asd.contel.com!gary
From: gary@matterhorn.ctc.contel.com (Gary Bisaga  x4219)
Newsgroups: comp.lang.c++
Subject: Re: Inline Isn't
Message-ID: <1991Jun5.082021@matterhorn.ctc.contel.com>
Date: 5 Jun 91 12:20:21 GMT
References: <1666@msa3b.UUCP>
Sender: news@europa.asd.contel.com (News)
Reply-To: gary@ctc.contel.com
Organization: Contel Technology Center (until the GTE debacle)
Lines: 13
Nntp-Posting-Host: matterhorn.ctc.contel.com


In article <1666@msa3b.UUCP>, kevin@msa3b.UUCP (Kevin P. Kleinfelter) writes:
> When one declares a class in a .h file, and specifies the body of
> a member function as inline in the .cpp file, do most compilers fail
> to inline it?
> 
> i.e. In the following code, is it too much to expect that "f.bar()" gets
> inlined? This would require a smart linker.
> [example of putting an inline function in another file]
Yes, but not because it's just the linker.  The ARM (don't have it handy,
so I can't remember the section reference) says that inline functions must
have internal linkage, i.e., be defined in the same source file where they're
used.