CPUProfiler: Simplify logging part of CreateCodeEvent functions.

CPUProfiler: Simplify logging part of CreateCodeEvent functions.

We have 5 overloaded functions with name CreateCodeEvent.
All these functions have many common parts. I'd like to eliminate the difference between them.

TEST=existing tests
R=yangguo@chromium.org, yurys@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=15287

[yurys, 2013-06-21 09:32:50]

lgtm

Please check that --prof and tick-processors are not affected by this change.

https://codereview.chromium.org/16901014/diff/1/src/log.cc
File src/log.cc (right):

https://codereview.chromium.org/16901014/diff/1/src/log.cc#newcode922
src/log.cc:922: ASSERT(symbol);
No need in this assert.

https://codereview.chromium.org/16901014/diff/1/src/log.cc#newcode1199
src/log.cc:1199: AppendCodeCreateHeader(&msg, REG_EXP_TAG, code, -2);
I wonder if we could have a constant with descriptive name for -2.

[loislo, 2013-06-21 10:05:13]

On 2013/06/21 09:32:50, Yury Semikhatsky wrote:
> lgtm
> 
> Please check that --prof and tick-processors are not affected by this change.

Looks like it works fine.

> 
> https://codereview.chromium.org/16901014/diff/1/src/log.cc
> File src/log.cc (right):
> 
> https://codereview.chromium.org/16901014/diff/1/src/log.cc#newcode922
> src/log.cc:922: ASSERT(symbol);
> No need in this assert.

done


> 
> https://codereview.chromium.org/16901014/diff/1/src/log.cc#newcode1199
> src/log.cc:1199: AppendCodeCreateHeader(&msg, REG_EXP_TAG, code, -2);
> I wonder if we could have a constant with descriptive name for -2.

I have no idea why we use -2 here and what is the meaning of this constant.

[Sven Panne, 2013-06-21 11:15:33]

On 2013/06/21 10:05:13, loislo wrote:
> [...] I have no idea why we use -2 here and what is the meaning of this
constant.

Hmmm, I have no clue, either. tickprocessor.js doesn't seem to use it if I see
things correctly, so the question is: Do we need this "kind" column at all?

[loislo, 2013-06-21 14:01:15]

On 2013/06/21 11:15:33, Sven Panne wrote:
> On 2013/06/21 10:05:13, loislo wrote:
> > [...] I have no idea why we use -2 here and what is the meaning of this
> constant.
> 
> Hmmm, I have no clue, either. tickprocessor.js doesn't seem to use it if I see
> things correctly, so the question is: Do we need this "kind" column at all?

First of all we use it in plot-timer-events.js
Also if we remove it then we break backward compatibility with the old v8 logs.
As far as I see we use -2 as a constant for RegExp in the logging code but in
the 
c++ code we map regexp to builtins. see
https://code.google.com/p/chromium/codesearch#chromium/src/v8/src/objects.h&s...

[loislo, 2013-06-24 12:10:02]

+yangguo

I found that REGEXP kind is used for plot-timer-events.js.

[Yang, 2013-06-24 12:49:15]

Got some comments. LGTM once those are fixed.

https://codereview.chromium.org/16901014/diff/7001/src/log.cc
File src/log.cc (right):

https://codereview.chromium.org/16901014/diff/7001/src/log.cc#newcode865
src/log.cc:865: msg.Append("%s,%s,-3,",
We also got "-3" here to indicate external call. You changed this to -2 in
plot-timer-events.js, so you need to apply the change here as well.

Maybe you can unify this part as well, and use an enum or a constant instead of
an inline "-3".

https://codereview.chromium.org/16901014/diff/7001/src/log.cc#newcode1514
src/log.cc:1514: case Code::REGEXP:
while we are at this, could we fix this here?
description = "Regular expression code";
tag = Logger::REG_EXP_TAG;
break;

[loislo, 2013-06-24 12:54:31]

https://codereview.chromium.org/16901014/diff/7001/src/log.cc
File src/log.cc (right):

https://codereview.chromium.org/16901014/diff/7001/src/log.cc#newcode865
src/log.cc:865: msg.Append("%s,%s,-3,",
On 2013/06/24 12:49:15, Yang wrote:
> We also got "-3" here to indicate external call. You changed this to -2 in
> plot-timer-events.js, so you need to apply the change here as well.
> 
> Maybe you can unify this part as well, and use an enum or a constant instead
of
> an inline "-3".

Done.

https://codereview.chromium.org/16901014/diff/7001/src/log.cc#newcode1514
src/log.cc:1514: case Code::REGEXP:
On 2013/06/24 12:49:15, Yang wrote:
> while we are at this, could we fix this here?
> description = "Regular expression code";
> tag = Logger::REG_EXP_TAG;
> break;

Done.

[loislo, 2013-06-24 12:55:35]

Committed patchset #3 manually as r15287 (presubmit successful).