On Fri, Aug 31, 2001 at 05:50:28PM +0200, Tels wrote: > Did I explain it better now? Yes. Actually in retrospect you covered it well enough before, I was just being dense. Maths never was my strong suit. > No problem with the name ;) (Could have an uncovered() alias, though ;) Done. It gave me a solid incentive to make your form of documentation work too :) > I think just disabling /^(/ by default might work with overload, but could > up users of legitimate /^(/ sub names - although I can not imagine why > someone would do this. My concern is more that overload changes in the > future and then you relied on an undocumented feature and your module > breaks. So, finding a real solution like detecting these overload > funcction-stubs (or whatever they are) is better. > > But the hack would work, if you document it as such. Well it's documented now, let me know if the docs make sense. Though I may sometimes be a documentation nazi, it doesn't seem follow that I'll be good at writing them. > > Sorry to be deferring on so many things this time round. I'm helluva > > busy at the moment and won't have a real chance to look at anything > > until tomorrow night when I get back home. > > I didn't want to push you. You do great work and it is much appreciated! I > just wanted to let you know of these issues, since you asked ;) Excuse my stressed reply, I was rushed when I wrote it, and my tone was more in reaction to that than your post. I also forgot to thank you for your kind words. I can't really take too much credit though, there were many people there for the conception, and the idea seemed so simple that I was sure it must have already been done. We're really glad that people are finding uses for it. On Tue, Aug 28, 2001 at 08:57:45PM -0400, Michael G Schwern wrote: > Here's another suggestion. You could check to see if the module ISA > Exporter and only consider the exportable functions for POD coverage. > UNIVERSAL::exports will tell you what's exportable. I lifted the code and in 0.06 am unveiling Pod::Coverage::ExportOnly, which does pretty much what it says on the tin. I cribbed the code from UNIVERSAL::exports, since polluting UNIVERSAL just to test for docs seemed a little like overkill. > private_check would be called for anything not explicitly declared > private. I jumped the other way, subliminally doing as Tony had before me[0]. Now if you want to change how private things are examined you can derive your own class from Pod::Coverage, and override the _private_check method. Tony, can you give me feedback on if 0.06 is now more like what you hacked 0.02 into, or does it still need that separate _load_code interface? Everyone else, feedback on the new interface would be cool, I want to finalise it within the next few revisions so people can shoot off and write Pod::Coverage::MyWay safe in the knowledge that I won't stomp on the interface much more. And now. Drumroll please. 0.06 of Pod::Coverage, heading towards CPAN, and available in my little corner of the interweb - http://unixbeard.net/~richardc/lab/Pod-Coverage/Pod-Coverage-0.06.tar.gz I probably won't hack on it much tomorrow, but next week I'm back in the London office so I'll be able to fit some hacking in then, which will help my stress levels no end :) Later [0] I only just remembered when reviewing this thread for things I said were on the todo list. -- Richard Clamp <[EMAIL PROTECTED]> god damn, even superman shot himself
