upgrade to jflex 1.9.1 and remove the workaround for lex scanners limits

[tarzanek]

jflex 1.9.0 is out and
jflex-de/jflex#203
got fixed and jflex now introduces:
(jflex-de/jflex#197 jflex-de/jflex#1045 )

%token_size_limit (<identifier>|<number>)

which makes the workaround of manually patching generated lexical scanner obsolete,
though I see @vladak removed it ?!?!?
d051e17
(I don't see 1.8 fixes it ... )
so it needs to be set where applicable so OpenGrok is in sync with lucene limits and their tokenizers

also jflex-de/jflex#542 is fixed so https://github.com/oracle/opengrok/blob/master/pom.xml#L348
can be cleaned up

[sreenath-tm]

Hi I would like to work on this issue can you assign this issue to me

[vladak]

Hi I would like to work on this issue can you assign this issue to me

just submit a PR - the chances that someone else will also work on this are infinitesimaly small.

[tarzanek]

jflex 1.9.1
now
ok, I will try to set some time aside for this, it should help with performance