Add a template-specialized constructor for der::Input for constants

Add a template-specialized constructor for der::Input for constants

Make it easier to create der::Input's from constants by the magical
power of Templates.

Old and busted: der::Input foo(kFoo, arraysize(kFoo));
New hotness: der::Input foo(kFoo);

BUG=none
R=mattm@chromium.org

Committed: https://crrev.com/c415dfed788a32ad110d30f97a434b4cac99637b
Cr-Commit-Position: refs/heads/master@{#330882}

[Ryan Sleevi, 2015-05-21 00:53:06]

Matt: Easy one for ya :)

[mattm, 2015-05-21 01:31:10]

https://codereview.chromium.org/1146283004/diff/1/net/der/input.h
File net/der/input.h (right):

https://codereview.chromium.org/1146283004/diff/1/net/der/input.h#newcode53
net/der/input.h:53: explicit Input(const std::string& s);
It occurs to me that this constructor is kinda dangerous due to the std::string
implicit conversion constructors.

At first I was thinking about:
der::Input foo("te\0st");

which would do the wrong thing with std::string conversion parsing it as a null
terminated string. 

But then I realized it's even worse since the temporary std::string gets
destroyed immediately and the Input is pointing at freed memory.

Maybe we should also create an explicit char array constructor to avoid that?

[Ryan Sleevi, 2015-05-21 01:38:13]

https://codereview.chromium.org/1146283004/diff/1/net/der/input.h
File net/der/input.h (right):

https://codereview.chromium.org/1146283004/diff/1/net/der/input.h#newcode53
net/der/input.h:53: explicit Input(const std::string& s);
On 2015/05/21 01:31:10, mattm wrote:
> It occurs to me that this constructor is kinda dangerous due to the
std::string
> implicit conversion constructors.
> 
> At first I was thinking about:
> der::Input foo("te\0st");
> 
> which would do the wrong thing with std::string conversion parsing it as a
null
> terminated string. 
> 
> But then I realized it's even worse since the temporary std::string gets
> destroyed immediately and the Input is pointing at freed memory.
> 
> Maybe we should also create an explicit char array constructor to avoid that?

Holy crap! You're totally right, and in fact all the usages that I can find are
explicitly wrong (from a memory-safety POV) because of this, because they're all
implicitly casting chars to strings.

Luckily, it's all unit tests, so we're not in danger, but yeah...

[mattm, 2015-05-21 01:53:51]

I see you filed a separate bug for that, so this one lgtm.

[Ryan Sleevi, 2015-05-21 01:54:28]

The CQ bit was checked by rsleevi@chromium.org

[commit-bot: I haz the power, 2015-05-21 01:55:47]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1146283004/1

[commit-bot: I haz the power, 2015-05-21 04:32:11]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-05-21 04:34:02]

Patchset 1 (id:??) landed as
https://crrev.com/c415dfed788a32ad110d30f97a434b4cac99637b
Cr-Commit-Position: refs/heads/master@{#330882}