On Tue, 25 Jan 2005, Julian Foad wrote:
> Peter N. Lundblad wrote:
> > On Mon, 24 Jan 2005 julianfoad@tigris.org wrote:
> [...]
> >>+ svn_xml_make_open_tag (&sb, pool, svn_xml_protect_pcdata, "commit",
> >>+ "revision",
> >>+ apr_psprintf (pool, "%" SVN_REVNUM_T_FMT,
> >
> > We use %ld everywhere else, so why not here?
>
> Hmm. You're pretty much right - tho only other places where we use it are in:
>
> contrib/client-side/svn-push/svn-push.c
> subversion/bindings/java/javahl/native/*.cpp (3 places)
> subversion/svnadmin/main.c
>
Then I missed those few places in when I did this last spring. They should
be changed as well.
> I believe the reason for not using it is because the localisation/translation
> utility can't parse the macro substitution and string concatenation that
> results from usage like:
>
> _("English text %" SVN_REVNUM_T_FMT " string continues").
>
That's exactly the reason we did it.
> However, in this instance, the string is not one that needs to be translated.
> Do we wish to deprecate SVN_REVNUM_T_FMT? If not, I think we should continue
> to use it where we can, though I will admit that the benefit of using it is
> vastly diminished if it is only used in a few places.
>
We didn't want to deprecate it since others might want to use it and there
is nothing wrong with it. In our own code we should just stick to %ld. We
gain nothing by using the longer form. We can't easily change the type
from long int anyway. So I think using it in a few places is just
confusing.
> Thanks for the review.
>
welcome,
//Peter
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Jan 25 21:01:35 2005