For such questions, parse.y is usually the place to look. If memory is used, “t” stands for token, while “k” stands for keyword.
Here are the different tokens that denote identifiers (in the sense of names for other things):
%token <id> tIDENTIFIER tFID tGVAR tIVAR tCONSTANT tCVAR tLABEL
The only definition for kEND that I found with a quick search was for k_end :
k_end : keyword_end { token_info_pop("end"); } ;
source share