Examples > E-mail address > Detecting a conflict
So far no problems occurred and the project seems to be ready. But if you select the start rule mailbox and parse it two errors are shown:
mailbox: LL(1) Error: "\"" is the start of several alternatives
mailbox: LL(1) Error: "ATOM" is the start of several alternatives
If such a LL(1)-conflict exists, the actual text don't suffices for the parser to decide, which rule should be applied next.
The TextTransformer - in contrast to some other related tools - helps to discover such errors. But to remove them, you need some understanding of the grammar, you are developing and some power of deduction.
In the mailbox production:
addr_spez | phrase route_addr
the alternatives are beginning with addr_spez and phrase. If you select addr_spez in the navigation tree at the right side of the screen and click the right mouse button, the first set of this production can be shown:
If you do the same with the phrase production, you will see the same First set.
If you look at the definitions of these productions, the cause of the conflict still isn't clear:
addr_spez ::= local_part "@" domain
phrase ::= word+
But if you go down another step, the conflict becomes obvious:
local_part ::= word ("." word)*
word ::= ATOM | quoted_string
Indirectly as well addr_spez as phrase are beginning with word.