Hi Pablo, On Mon, Sep 30, 2019 at 04:17:53PM +0200, Pablo Neira Ayuso wrote: > On Wed, Sep 25, 2019 at 03:14:19PM +0200, Fernando Fernandez Mancera wrote: > > Currently clang requires EXPORT_SYMBOL() to be above the function > > implementation. At the same time doxygen is not generating the proper > > documentation because of that. > > > > This patch solves that problem but EXPORT_SYMBOL looks less like the Linux > > kernel way exporting symbols. > > Applied, thanks. I missed this earlier - take a look at the man pages / html doc with this patch. E.g. man attr: > attr(3) libmnl attr(3) > > > > NAME > attr - Netlink attribute helpers > > Functions > EXPORT_SYMBOL uint16_t mnl_attr_get_type (const struct nlattr *attr) > EXPORT_SYMBOL uint16_t mnl_attr_get_len (const struct nlattr *attr) > EXPORT_SYMBOL uint16_t mnl_attr_get_payload_len (const struct nlattr > *attr) > EXPORT_SYMBOL void * mnl_attr_get_payload (const struct nlattr *attr) > EXPORT_SYMBOL bool mnl_attr_ok (const struct nlattr *attr, int len) The web pages are the same. Shunting all the EXPORT_SYMBOL lines to the start of the file as in my rejected patch might have been ugly, but at least it left the documentation looking as it should. I just finished making a patch for libnetfilter_queue using the redefined EXPORT_SYMBOL as above but taking care to avoid generating lines over 80 chars and preserving (or fixing!) alignment of subsequent parameter lines. But I won't submit it, because it results in the same horrible documentation. I think it should be not too onerous to move the EXPORT_SYMBOL lines to before the start of documentation, which should satisfy both doxygen adn clang. Would you like me to go ahead with that? Cheers ... Duncan.
