Imho, it is better to break the use of __at <num> in favor of __at(<num>).
пт, 17 июн. 2022 г. в 20:13, Philipp Klaus Krause <p...@spth.de>: > > While working on C2X support in the parser (and trying to fix some > existing bugs in the parser already affecting earlier standards), it > looks to me like SDCC painted itself into a corner with __using, > __interrupt and __at. > > I want to break compability (already did in the parser branch, but not > in trunk). > > Currently for each of these, a constant expression can follow them, and > for __at, it must. This creates problems in the parser, and my proposed > solution is to restrict these: > > * For __using and __interrupt, the constant expression will be required > to be parenthesized. > * For __at, the constant expression will be required to be parenthesized > or be just a plain constant. > > I.e. > > int __at 3+3 i; // No longer allowed from SDCC 4.2.3 > int __at 6 i; // Still allowed > int __at (3+3) i; // Still allowed > > Obviously, this will break some existing code. > > Philipp > > > _______________________________________________ > sdcc-devel mailing list > sdcc-de...@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/sdcc-devel _______________________________________________ Sdcc-user mailing list Sdcc-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/sdcc-user
