摘要
现有的前端分析多数是通过编写相应的可能含有二义的LALR(1)文法规范,利用YACC或其变种自动生成的。在这些解析器的Java再工程中,可以用CUP去替代YACC,这样再工程的焦点转移到对文法规范的变换。由于YACC及CUP在二义的解决、解析器的构造等有细微的差别,因此对复杂的文法规范的变换并不是一件容易的事。该文通过剖析YACC和CUP,指出了它们的不同之处,并总结出有二义的YACC文法到CUP文法变换的基本原则和几个变换法则。
Most of the existing parsers are auto-generated by YACC or its variants via writing corresponding and possibly ambiguous LALR（1） syntax specifications. It is simple to reengineer these parsers to Java with CUP instead of YACC, which focuses all intentions on translating the syntax specifications. Since there are some subtle differences between YACC and CUP in recolving conflicts and constructing parsers, etc., it is not easy to translate a complex syntax specification.By means of anatomizing YACC and CUP, this paper indicates the discrepancies among them.Moreover, one basic fundamental and several transform principles are summarized, which can be used to translate any YACC syntax specification into CUP one.
出处
《计算机工程》
EI
CAS
CSCD
北大核心
2005年第15期50-52,共3页
Computer Engineering
基金
国家自然科学基金资助项目(60173049)
Intel中国研究中心(ICRC)基金资助项目
关键词
LALR(1)
解析器
YACC
CUP
二义性
冲突
LALR（1）
Parser
Yet another compiler-compiler （YACC）
Constructor of useful parsers （CUP）
Ambiguity
Conflict
