Why do we want it? On Wed 26 Jul 2017 at 16:46, Daniel Ferreira (theiostream) <[email protected]> wrote:
> The macro, sadly, would not export the new declaration as a symbol of its > own, which I believe we want. > > On Jul 26, 2017 12:14, "Ivan Vučica" <[email protected]> wrote: > >> Fwiw original version defined two symbols pointing to constant strings >> with the same value. Fred and I asked for changes. I asked Daniel to >> investigate using NSString const*, and when that worked we went ahead with >> submitting the change. >> >> Sorry about this. We did not consider compiler differences, but that's >> why we have people trying it out \o/ >> >> I like Josh's workaround. >> >> On Wed, Jul 26, 2017, 13:15 Riccardo Mottola <[email protected]> >> wrote: >> >>> Hi, >>> >>> Daniel Ferreira (theiostream) wrote: >>> > Also, the reason it just does not assign the same const string to the >>> > different constants is because the two consts should be the same >>> > pointer, and doing it explicitly seemed like a good way to make that >>> > intent clear and guarantee that would happen. >>> >>> I will commit the "fix" to unbreak build instead of "reverting" so your >>> additions get preserved, however I do like and understand the idea of >>> using and ensuring the samge string with the same poinger gets used. >>> >>> I thus ask compiler experts whats wrong of or if this is a quirk with >>> NSStrings in GCC ? Or somethig known that clang is smarter to accept >>> resolving the assignments to the same pointer? >>> >>> Riccardo >>> >>> _______________________________________________ >>> Gnustep-dev mailing list >>> [email protected] >>> https://lists.gnu.org/mailman/listinfo/gnustep-dev >>> >> -- Sent from Gmail Mobile on iPad
_______________________________________________ Gnustep-dev mailing list [email protected] https://lists.gnu.org/mailman/listinfo/gnustep-dev
