Re: [Nano-devel] the compiled size of nano

[Mike Frysinger]

On 08 Feb 2016 21:52, Benno Schulenberg wrote:
> On Mon, Feb 8, 2016, at 04:39, Mike Frysinger wrote:
> > On 07 Feb 2016 19:51, Benno Schulenberg wrote:
> > > On Sun, Feb 7, 2016, at 01:11, Mike Frysinger wrote:
> > > > my system:
> > > >   linux-4.4.0
> > > >   glibc-2.22
> > > >   gcc-5.3.0
> > > 
> > > Mine:
> > >     linux-2.6.32
> > >     glibc-2.14
> > >     gcc-4.4.3
> > 
> > well, those versions are old and have known bugs.  you'd have to look
> > at the configure output to see what's being detected as broken.
> 
> Nothing seems to be broken.  A full diff of a ./configure on
> an old and a new system is pasted at the end.
> 
> > you
> > can start with the objects actually created under lib/ and the symbols
> > each contain.
> 
> As far as I can tell it's just getopt.o and mbrtowc.o.

the version of glibc you're using has a known bug:
https://www.gnu.org/software/gnulib/manual/gnulib.html#getopt
* This function crashes if the option string includes W; on some platforms: 
glibc 2.14.

in this particular case, there's no way for gnulib to know if we're
using getopt in that way, so it just always includes it.  it doesn't
have a cache value either for us to signal that we aren't using "W;"
so we can't disable it.  considering how old glibc-2.14 is, i think
taking the hit is fine.  how many people are running a system that is
5+ years out of date and building latest nano from source ?

> > all in all, i'd say it's WAI.
> 
> What's "WAI"?  Way after ...?

Working As Intended
-mike

signature.asc
Description: Digital signature