El dilluns, 30 de maig de 2016, a les 15:53:51 CEST, Carlos Garcia Campos va escriure: > Jakub Kucharski <[email protected]> writes: > > On Sun, 2016-05-29 at 22:50 +0200, Albert Astals Cid wrote: > >> Hi guys, what do you think of us adding "some" of the C++11 features > >> in > >> poppler? > >> > >> One that would be useful for example is std::unique_ptr, that is a > >> class that > >> will take ownership of a pointer and delete it when it goes out of > >> scope. > > > > I think *some* C++11 features (e.g. std::unique_ptr) would make things > > easier. However there are some which could make things harder, e.g. > > type inference (auto) - it sometimes makes it hard to reason about the > > code when you don't know what type a function returns. So yes, some > > C++11 features would be OK, but perhaps we could make guidelines for > > it. > > Well, we could just document the cases where it could be used.
Do we want to make that list now or go over it on a patch-by-patch, case-by- case basis? Cheers, Albert > We don't > use templates in poppler, so we don't have huge types where auto is > quite useful, but still there are cases where things look much better, > like in modern for loops: > > for (auto foo : fooList) > > also in cases where the type is twice like: > > Foo* f = reinterpret_cast<Foo*>(b); -> auto f = reinterpret_cast<Foo*>(b); > or > Foo* f = std::make_unique<Foo>(); -> auto f = std::make_unique<Foo>(); > > > Jakub > > > > _______________________________________________ > > poppler mailing list > > [email protected] > > https://lists.freedesktop.org/mailman/listinfo/poppler _______________________________________________ poppler mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/poppler
