Prevent a NULL deref

Prevent a NULL deref

CID=16704

Committed: http://code.google.com/p/v8/source/detail?r=8384

[groby-ooo-7-16, 2011-06-22 00:13:42]

Would you mind reviewing this? (Part of Coverity Fixlet)

[mnaganov (inactive), 2011-06-22 05:02:47]

LGTM with a comment.

http://codereview.chromium.org/7218036/diff/1/src/log.cc
File src/log.cc (right):

http://codereview.chromium.org/7218036/diff/1/src/log.cc#newcode404
src/log.cc:404: DeleteArray(static_cast<const char*>(entry->value));
We allocate an array of char, why do we cast to const char* here?

[groby-ooo-7-16, 2011-06-22 17:17:26]

http://codereview.chromium.org/7218036/diff/1/src/log.cc
File src/log.cc (right):

http://codereview.chromium.org/7218036/diff/1/src/log.cc#newcode404
src/log.cc:404: DeleteArray(static_cast<const char*>(entry->value));
On 2011/06/22 05:02:47, Mikhail Naganov (Chromium) wrote:
> We allocate an array of char, why do we cast to const char* here?

Presumably because DeleteArray takes a pointer to the first array element. That
seems to be that way since the beginning of time (or specifically, SVN r9 :)

[mnaganov (inactive), 2011-06-22 18:54:23]

http://codereview.chromium.org/7218036/diff/1/src/log.cc
File src/log.cc (right):

http://codereview.chromium.org/7218036/diff/1/src/log.cc#newcode404
src/log.cc:404: DeleteArray(static_cast<const char*>(entry->value));
On 2011/06/22 17:17:26, groby wrote:
> On 2011/06/22 05:02:47, Mikhail Naganov (Chromium) wrote:
> > We allocate an array of char, why do we cast to const char* here?
> 
> Presumably because DeleteArray takes a pointer to the first array element.
That
> seems to be that way since the beginning of time (or specifically, SVN r9 :)

I mean that a name string is allocated like this: NewArray<char>, that is, char*
And then it is casted to <const char*> here. Not a big deal, but still.

Do you have rights to land this?

[groby-ooo-7-16, 2011-06-22 19:14:39]

Preliminary committer, so yes. (It's working its way through the try bots right
now)

Didn't touch the const char* because I'm not that familiar with the code base
yet, but it seems that you are right...

[mnaganov (inactive), 2011-06-22 19:19:06]

On 2011/06/22 19:14:39, groby wrote:
> Preliminary committer, so yes. (It's working its way through the try bots
right
> now)
> 

This is V8, in a separate repository from Chromium. I don't actually see you in
V8 project members, so I'll land it.

> Didn't touch the const char* because I'm not that familiar with the code base
> yet, but it seems that you are right...