Re: [lmi] MinGW 3.4.5, stdio formatting functions and multiple definitions

[Greg Chicares]

On 2014-12-07 02:08, Vadim Zeitlin wrote:
> On Sun, 07 Dec 2014 01:14:11 +0000 Greg Chicares <address@hidden> wrote:
> 
> GC> >  So far I've just quickly tested that doing this indeed works, i.e. we 
> can
> GC> > omit -Wl,--allow-multiple-definition and everything still links (and 
> runs)
> GC> > successfully if we patch boost.regex to use ::sprintf(). Is it worth it 
> to
> GC> > prepare a real patch to lmi doing this
> GC> 
> GC> Yes. I suggest
> [...suggestion snipped...]
> 
>  Yes, this was exactly what I planned to do and I just did it quickly and
> it seems to work fine, it's nice how reusable install_miscellanea.make is.
> Please find the patch against the latest svn trunk, i.e. r6052, attached.

Applied 20150308T1823Z, revision 6122. The former workaround smelled
nasty, and had been spreading of late, so it's really good to get this
problem solved in a better way.

I used the following procedure to ensure everything gets updated:

/lmi/src/lmi[0]$rm -rf ../build
/lmi/src/lmi[0]$make $coefficiency -f install_miscellanea.make clobber
/lmi/src/lmi[0]$make -f install_miscellanea.make >../log 2>&1

BTW, to make the output linear for clarity, I didn't use $coefficiency
in the last command, where this 'make' output is normal and expected...

[ -e xmlwrapp-0.7.1.patch ] && patch --directory=scratch --strip=1 < 
xmlwrapp-0.7.1.patch
install_miscellanea.make:224: recipe for target 'xmlwrapp' failed
make: [xmlwrapp] Error 1 (ignored)

...because no 'xmlwrapp-0.7.1.patch' exists. It might be nice to
make that not fail, but I don't have time for that right now.
(I think adding ' || true' at the end of the command might do it,
but I'm not sure that's guaranteed to work portably.)

Then I had to reinstall a locally modified /opt/lmi/data/qx_cso.??? .

Probably it would have been just as good, and less trouble, to do:

/lmi/src/lmi[0]$rm -rf ../build
/lmi/src/lmi[0]$make $coefficiency -f install_miscellanea.make boost >../log 
2>&1