On Wed, Jul 22, 2009 at 02:30:55PM -0400, Laine Stump wrote:
On 07/22/2009 11:36 AM, Daniel P. Berrange wrote: It's actually because I like doing this that I'd like to know the preferred method of eliminating the warnings I mentioned. There are a bunch of them pre-existing in the code that I want to get rid of so I can turn on warnings=error (without turning off these warnings in CFLAGS), and I want to do it the "accepted" way. For example, from domain_conf.c:2137:
virDomainReportError(conn, VIR_ERR_XML_ERROR, _("invalid security type"));
suppose one of the translators made a mistake and updated with a wrong string from somewhere else replacing it with "Erreur de securite %s" for example due to a cut and paste mistake, it's better to see the %s out than have the application crash, right ;-) ? Of couse this can happen with formats embedding '%s' but that's just another argument to use "%s" in my book.
spits out the warning. We all know that it really *is* literal, but the macro is changing the class so the compile thinks it isn't. It would be simple to just change it to:
virDomainReportError(conn, VIR_ERR_XML_ERROR, "%s", _("invalid security type"));
Feels safer to me, really.
(and there are plenty of those too), but that's inefficient, and doesn't do the _() around the "%s" (is that correct or not?).
If someone wants to tell me the preferred way of doing these, I'll handle the grunt work of making the changes.
Thanks :-) Daniel -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel@veillard.com | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/