On Wed, Sep 15, 2010 at 02:27:21PM +0100, Julian Foad wrote:
> Also printf formatting type mismatches:
> 
> subversion/libsvn_subr/svn_string.c: In function
> 'svn_cstring_strtoui64':
> subversion/libsvn_subr/svn_string.c:662: format '%lu' expects type 'long
> unsigned int', but argument 5 has type 'apr_uint64_t'
> subversion/libsvn_subr/svn_string.c:662: format '%lu' expects type 'long
> unsigned int', but argument 6 has type 'apr_uint64_t'
> subversion/libsvn_subr/svn_string.c: In function 'svn_cstring_strtoi64':
> subversion/libsvn_subr/svn_string.c:706: format '%ld' expects type 'long
> int', but argument 5 has type 'apr_int64_t'
> subversion/libsvn_subr/svn_string.c:706: format '%ld' expects type 'long
> int', but argument 6 has type 'apr_int64_t'
> 
> (These should use APR_[U]INT64_T_FMT.  Doing so is complicated by the
> fact that the format string localization tools don't like string
> concatenation with macros, and there's no pool easily available to do a
> temporary printf.)

The translation issues is why I didn't use the APR_[U]INT64_T_FMT.
It's either using it and make life hard for translators, or have these
warnings... neither solution is perfect. Or is there a better solution?

We have a similar define for svn_revnum_t BTW, but it's been deprecated
for the same reason.

Stefan

Reply via email to