从 bison 或 yacc 过滤 y.output？
我正在使用 bison，通过查看 y.output 很难找出冲突。是否有工具可以制作或过滤 y.output 使其更有用？我很想看到进入冲突状态的完整路径，但我想要任何有帮助的东西。
I am using bison and its difficult to figure out the conflicts by looking at y.output. Is there a tool to make or filter y.output so its more useful? i would love to see the full path to the state with the conflict but anything helpful is what i would like.
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
我假设您知道什么是移位/减少和减少/减少冲突以及它们何时发生。鉴于此，我所做的只是使用 vim...
在 y.output 的顶部，它列出了哪些状态有多少冲突。如果您只需输入“/statenumber”并按“n”几次，您应该能够直接进入该状态的转换。
在那里，您将看到用“.”输入的产生式规则。在他们之中。点表示产生式规则中已解析到的位置。如果点位于末尾，则意味着它将尝试“减少”，如果不是，则尝试“移动”。如果下一个标记位于 LHS 非终结符的 FOLLOWSET 中（下一个标记是可以跟随该非终结符的符号），则发生减少产生式，并且对于终结符或 FIRST（非终结符）发生移位在“.”之后。
如果对于下一个可能的标记有两种可能的移动（移位/归约或归约/归约），则会发生冲突。
当您发现冲突时，只需进入该状态，找出哪些产生式规则给您带来冲突，然后： a - 修改语法以消除该冲突，或 b - 使用 %prec 表示法指定产生式规则优先级。
希望有帮助:)
I assume you know what shift/reduce and reduce/reduce conflicts are and when they occur. Given that, what I did was just use vim...
At the top of y.output, it lists which states have how many conflicts. If you just type "/statenumber" and press 'n' a couple of times, you should be able to go straight to the transitions for that state.
There, you'll see the production rules you typed with a '.' in them. The dot indicates the place in the production rule up to which it has parsed. If the dot is at the end, that means that it'll attempt a 'reduce', and if not, a 'shift'. A reduce production occurs if the next token is in the FOLLOWSET of the LHS non-terminal (the next token is a symbol that can follow that non-terminal), and a shift occurs for the terminal or FIRST(non-terminal) that is after the '.'.
A conflict occurs if you have two possible moves (shift/reduce or reduce/reduce) for a possible next token.
When you find conflicts, simply go to that state, find out which production rules are giving you the conflict, and either: a -- modify your grammar to remove that conflict, or b -- specify production rule precedence using the %prec notation.
Hope that helped :)