Re: Current trunk is broken for MinGW, patches included

[Eli Zaretskii]

> Date: Mon, 29 Dec 2014 05:19:28 +0100 (CET)
> Cc: address@hidden, address@hidden
> From: Werner LEMBERG <address@hidden>
> 
> >> This appears to be an issue with varying implementations/standards.
> >> Later standards declare the parameter as a 'char **', presumably
> >> because char ** doesn't convert to const char **. We could always
> >> stick a cast in there if the warning is a problem.
> > 
> > It's not a catastrophe, but with the latest versions of GCC becoming
> > more and more noisy, my personal preference is to shut up as many
> > warnings as possible, so that real problems are clearly visible.
> 
> Hmm, a simiar issue should be fixed already in gnulib with a recent patch
> from Daiki (commit facebc6bcfd01af65097f5ae96c75c94268354ff)...  Or is
> this indeed a different problem?

It's a different, but related problem.  ICONV_CONST is already
computed correctly (see config.h), but it is not used in info-utils.c
where 'iconv' is being called.