> >> no, this is good news. I've had to fix so much broken software on my > >> alpha becuase of sloppy typecasting it isn't even funny. > > > > I've run into similar headaches with my amd64 machine. If you forget > > to prototype a function that returns a pointer, you crash. > > sizeof(int) is not the same as sizeof(long) or sizeof(void*). Etc. > > This definitely IS good news.
I'm glad to hear that everybody thinks that all this breakage is good news since in some abstract sense this will force open-source programmers to maintain better type discipline. Unfortunately, most programming errors I am aware of are malloc/free errors which gcc still permits, but never mind that detail. . . . Meanwhile, in the real world we have a large problem: Lots of broken open-source EDA software. Who will fix the problems? Lessee. . . . who owns what?? * GTKWave is essentially abandoned. Somebody needs to adopt it & fix it. * Gerbv is still owned, so it stands a chance of getting fixed, but when? * Icarus Verilog is actively maintained by Steve Williams, so I am confident that he will be on it when his schedule allows it. * GSpiceUI is actively maintained, so it will probably get fixed, but when? So of the four non-compiling programs, three will likely be fixed eventually (hint hint, developers), and one may be dead forever if it is not adopted. Any volunteers? DJ: Since you wear a Red Hat these days, are there people in your organization who might be interested in adopting some of these programs and submitting patches to get them to work with gcc-4.x? I am most worried about GTKWave, which has been recently abandoned. I am happy to look at some of them, but, frankly, my free time is in very short supply nowadays. I will look at some of the issues, but can't promise to do so soon [1]. . . . . Meanwhile, this is an opportunity for new folks to make a real impact by submitting patches which fix the type problems! Stuart [1] Remember what happened to my promise to work on the PCB documentation. . . .
