byte-compiling a few font-lock related functions

[Tobias C. Rittweiler]

At the moment, ELP reports the following functions as the most expensive
while scrolling through a large .lisp file (almost 10k lines):

jit-lock-function                                          703         4.114131 
     0.0058522489
jit-lock-fontify-now                                       703         4.107315 
     0.0058425533
font-lock-fontify-region                                   705         
4.0757660000  0.0057812283
font-lock-default-fontify-region                           705         
4.0608700000  0.0057600992
font-lock-fontify-keywords-region                          705         
1.8490479999  0.0026227631

Byte-compiling these files results in the following timings:

jit-lock-function                                          703         
3.4189879999  0.0048634253
jit-lock-fontify-now                                       703         
3.4116139999  0.0048529359
font-lock-fontify-region                                   703         
3.3229810000  0.0047268577
font-lock-default-fontify-region                           703         
3.3077700000  0.0047052204
font-lock-fontify-keywords-region                          703         
1.6209109999  0.0023057055

(CVS version from late April.)

Difference is over half a second, so perhaps it'd make sense to
ship these functions byte-compiled?

  -T.

PS. This is the first scrolling through the file. Later scrolling is
    much snappier due to jit-lock.