Hi --

I've merged (as much as I am comfortable) the several patches
since popt-1.13. There are some widely divergent styles of fixing
essentially a simple problem, that strlen() cannot be used to calculate
display alignment with multibyte characters.

Attached (for convenience) are the changes I've merged already.

Note that I've rewritten some of the patches:
    e.g. POPT_dgetext() instead of _D_().

I have _NOT_ added this odd define (which should be added to
POPT_next_char() imho directly instead):

     #define POPT_next_char(p) (char *)((p) + utf8_skip_data[*(const  
unsigned char *)(p)])

And while I've isolated the changes necessary for removing  
POPT_WCHAR_HACK
into a stringDisplayWidth() routine, without a functional  
POPT_next_char() I cannot
remove POPT_WCHAR_HACK entirely.

I've checked Robert Scheck's LANG=de_DE@euro against rpm --help. The  
columns
align, but are perhaps indented more than I would expect.

Takao: Would you mind resending what remains from this
     http://rpm5.org/community/popt-devel/0058.html
patch against popt cvs HEAD? Thanks.

Except for documenting POPT_ARG_ARGV and writing a test case, popt-1.14
is done except for merging in the UTF-8 and POPT_fprintf() changes.

I won't release popt-1.1.4 before getting explicit WORKSFORME from  
both Robert and Takao however.

73 de Jeff