[mp2] foo_set() returning previous value

I'm reviewing parts of the API and I have second thoughts about this return-the-prev-value-on-set feature. While it's nice and perlish, it's not always efficient. If the user doesn't care what was the previous value and just wants to set a new value, we shouldn't enforce on her the overhead of

Re: win32 finfo->fname

Geoffrey Young wrote: At the moment the only idea I have is to have 2 different sets of APIs which underneath will call the same core functions. e.g.: skip_have_... have_... the only difference is that skip will also push the reason, and non-skip won't. But may be there are better ideas. I don't

Re: win32 finfo->fname

> At the moment the only idea I have is to have 2 different sets of APIs > which underneath will call the same core functions. e.g.: > > skip_have_... > have_... > > the only difference is that skip will also push the reason, and non-skip > won't. But may be there are better ideas. I don't thin