|From:||Darius Blasband <email@example.com>|
|Date:||22 Feb 1997 23:03:30 -0500|
|Organization:||Phidani Software, Brussels|
> Is is easier to have two separate lexical specifications -- one for
> using states? What do people usually do?
Out of our experience, using a single analyzer is probably not such a
good idea. When analyzing HTML, we even use different (and scaffolded)
analyzers for the main body, and the INPUT tags. It is also true that
our analysis was unorthodox, because we were optimizing it for
performance for our specific case.
Return to the
Search the comp.compilers archives again.