[DotGNU]Speeding up cscc -- a report

[Gopal V]

Hi ,
        CSCC as we all know is very fast ... but there are a few
speed freaks around who want to get that extra bit of speed
to help them build huge projects in a blink of an eye :-)

Ok, I've added the -flatin1-charset to all the pnetlib build
files and made CSLatin1Charset in cscc-cs to speed up the 
lexer a bit .. Which has give a slight performance gain in
cscc.

Moving onto the assembler , here's a report on the time usages
of various functions ..

Each sample counts as 0.01 seconds.
  %   cumulative   self              self     total           
 time   seconds   seconds    calls  ms/call  ms/call  name    
 27.06      1.48     1.48        1  1480.00  3290.00  ilasm_parse
 25.59      2.88     1.40  1074049     0.00     0.00  ilasm_lex
 16.27      3.77     0.89                             ILInternString
  4.39      4.01     0.24                             ILInternAppendedString

Showing a high amount of cumulative seconds for ILInternString ...
In general the assembler wastes about 60% of it's time reading the 
source file ie parse/lex and internstring ...

So removing the parse/lex or optimising them should bring a huge
improvement in the CSCC performance ... This is merely a report
of "Scope for optimisations" .

CSCC can therefore be a lot faster than it is now , which is good
news for everyone around.

In other good news, I've been playing around with the "Flat Controls" 
I got from gotdotnet.com and have managed to get WCheckBox in it running
in pnet... see http://symonds.net/~gopalv82/code/flatcontrols.jpg

Gopal
-- 
The difference between insanity and genius is measured by success