On Mon, 2008-06-23 at 07:17 -0400, Daniel J Walsh wrote: > Stephen Smalley wrote: > > On Thu, 2008-06-05 at 11:11 -0400, Daniel J Walsh wrote: > >> The problem I have is the compiler is too stupid to understand the > >> differences between a gen_requires block defining the required types and > >> the actual type definition. > >> > >> So I end up in a catch 22 where the compiler tells me I need to require > >> $1_rootwindow_t, but if I gen_require type $1_rootwindow_t, it tells me > >> I have a duplicate definition. > >> > >> So if you have a derived type in a gen_requires block the compiler can > >> not handle it. > > > > I'm a little unclear as to why this is required (why do you need to > > require and declare the same symbol again?). However, is there some > > reason we can't just automatically promote a require to a declaration > > upon encountering the latter? Seems like we've talked about this > > before. Not sure whether that should happen within libsepol > > symtab_insert() or in the callers, e.g. declare_type(). > > > I don't know, All I know is the compiler complains if it is there and > if it is not there. Catch 22. I end up going to great lengths to hack > around compiler errors... We add requires to templates, so that if they're used outside xserver, the caller gets the appropriate require. But then we also use the template inside xserver for code reuse, which is where the problem creeps up. There are a couple other examples of this in refpolicy, but I was able to work around them by reordering statements. It sounds like Dan's situation may not be something that can be easily worked around without some restructuring. -- Chris PeBenito Tresys Technology, LLC (410) 290-1411 x150 -- This message was distributed to subscribers of the selinux mailing list. If you no longer wish to subscribe, send mail to majordomo@xxxxxxxxxxxxx with the words "unsubscribe selinux" without quotes as the message.
