Restructure the IC / Handler compilers

Restructure the IC / Handler compilers

BUG=
R=ishell@chromium.org, mvstanton@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=22622

[Toon Verwaest, 2014-07-23 19:40:56]

PTAL

[mvstanton, 2014-07-24 16:33:47]

Just a quick comment for now, I don't like the formatting on the functions with
many (many) arguments where the arguments are bunched into as few lines as
possible.

Also is there any chance of breaking this up into a couple of CLs? :p

https://codereview.chromium.org/411973002/diff/1/src/ia32/stub-cache-ia32.cc
File src/ia32/stub-cache-ia32.cc (right):

https://codereview.chromium.org/411973002/diff/1/src/ia32/stub-cache-ia32.cc#...
src/ia32/stub-cache-ia32.cc:302: MacroAssembler* masm, const CallOptimization&
optimization,
nit: I like the old style better, what about starting the arguments on a new
line, like 

void PropertyHandlerCompiler::GenerateFastApiCall(
    MacroAssembler* masm,
    ....

https://codereview.chromium.org/411973002/diff/1/src/ia32/stub-cache-ia32.cc#...
src/ia32/stub-cache-ia32.cc:420: void
NamedStoreHandlerCompiler::GenerateStoreTransition(
Same here, it just hurts my eyes...

[Sven Panne, 2014-07-25 06:25:11]

DBC

https://codereview.chromium.org/411973002/diff/1/src/ia32/stub-cache-ia32.cc
File src/ia32/stub-cache-ia32.cc (right):

https://codereview.chromium.org/411973002/diff/1/src/ia32/stub-cache-ia32.cc#...
src/ia32/stub-cache-ia32.cc:302: MacroAssembler* masm, const CallOptimization&
optimization,
On 2014/07/24 16:33:47, mvstanton wrote:
> nit: I like the old style better, what about starting the arguments on a new
> line, like 
> 
> void PropertyHandlerCompiler::GenerateFastApiCall(
>     MacroAssembler* masm,
>     ....

This kind of formatting bikeshedding is gone with "git cl format"... ;-) Toon's
CL is correctly formatted according to that tool, so we should keep it like it
is.

Extended comment: The *real* problem is that functions with 8 arguments are
ugly, unreadable and completely untestable. "Uncle Bob" et. al. recommend 2-3
arguments at most.

[Toon Verwaest, 2014-07-25 08:43:25]

Ha, that's exactly what I said in person yesterday :-)

[mvstanton, 2014-07-25 12:28:42]

Yep, I'm absorbing the facts-on-the-ground of clang-format with some degree of
ill-humor. :p

[Igor Sheludko, 2014-07-25 13:43:31]

lgtm

Nice cleanup! lgtm

[mvstanton, 2014-07-25 14:40:47]

Is okay for me too.
lgtm.

[Toon Verwaest, 2014-07-25 17:51:05]

Committed patchset #1 manually as r22622 (presubmit successful).