Re: Enhancements to the include system
- Posted by cchris005 at fastmail.fm Dec 17, 2006
- 727 views
> Subject: Re: Enhancements to the include system > > > posted by: Pete Lomax <petelomax at blueyonder.co.uk> > > CChris wrote: > > I'm afraid that it makes maintainance/evolution of packages needlessly > > tricky. > Sorry, far too vague for my tiny mind. What is difficult to maintain? > What do you think is going to go wrong with it? > Changing the internal organisation of the library, and more generally of the whole include tree of an application, is something you expect do be able to do freely, and actually the user shouldn't have to bother about it. Using closeness rules - ie rules that depend on the geometry of the include tree - to resolve ambiguous unquaified global symbols may lead to surreptitious changes in program semantics when the include tree changes. The bugs arising from this will be fun to pinpoint... I didn't study the EuGTK package in detail, but, at that time, Irv mentioned that the library might need some rewriting to cope with such an otherwise seemingly practical change of rules. The aim - I think - is to be able to run pre-3.x code without need to edit it. This is why I think it is safer, and hardly complicated, to have a couple more explicit directives in the language in order to manage global symbols from outside the place they were defined, since the problems arise outside the file where the globals are defined. Regards CChris > > Why not chime in in the Wiki? > I'm not /that/ bored > > Pete > > -- cchris005 at fastmail.fm -- http://www.fastmail.fm - Access your email from home and the web