Arthur Schwarz wrote: > You're reading too much into what needs to be done. > [...] but I don't think it will take considerable code.
I know you're not a troll, but honestly, you're starting to sound like one. Akim, Kaz and I have given several specific arguments. You do not address them at all and just keep saying "I think it should be easy". > There are difficulties (fortunately - that keeps our minds occupied and > our pockets full). If you follow the list, you'll see that lack of things to do is not our main problem, quite the contrary, and we're not getting paid for working on Bison. > As to a heuristic, why? That would be a mystery. Unless you read some of the points we raised ... > My thought is that dependencies are established via function arguments > (fun(bad_boy)), and within data structures (struct name { <type> thing = > bad-boy; };), and through simple assignment. You're severely underestimating the complexity of languages such as C++. I gave some examples in my previous mail; I don't think repeating them will be of any use. > I haven't thought this thing through > and if you think that it's a time waster, don't do it. > I haven't given it much thought and if its looks like a big job, don't > do it. You guys have enough on your plate. Try to see this discussion from the POV of a neutral observer: There are people who have worked on the code pointing out specific problems, and someone who admits he hasn't thought it through or given it much though who doesn't think it should be hard. Whom would you tend to believe? But of course, Bison is free software. Perhaps we're all overlooking the easy solution. You can just make a patch (which would be short, in your words) and send it here, then we have something to go on.