Avoid using Pickle::WriteSize(), which writes an architecture-dependent amount

Avoid using Pickle::WriteSize(), which writes an architecture-dependent amount
of data, in histogram pickles. (The goal is to remove that method entirely.
Uses that never persist or send pickles over the network are [probably] safe,
but having the method around is waiting for accidental misuses.)
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=126514

[Mike Mammarella, 2012-03-12 18:29:02]

jar: this one's for you. (Pointing it out since I've cc'd you on all the others
as well.)

[jar (doing other things), 2012-03-12 18:33:00]

LGTM  I don't think the perf difference here is that critical.

I thought the real problem was for folks that *persisted* the pickled data. 
This example is used for IPC.

Is that problematic? Do folks connect 64 bit apps to 32 bit apps, where the
difference could be problematic?

[Mike Mammarella, 2012-03-12 18:37:46]

On 2012/03/12 18:33:00, jar wrote:
> LGTM  I don't think the perf difference here is that critical.
> 
> I thought the real problem was for folks that *persisted* the pickled data. 
> This example is used for IPC.
> 
> Is that problematic? Do folks connect 64 bit apps to 32 bit apps, where the
> difference could be problematic?

Yes, the problem is for people who persist the data. This is not such a use, nor
are any of the other uses that I've sent out so far (though there is one use
that persists it in the HTTP cache).

You weren't excited about the comment warning against using the method though,
so I'm trying to remove the method entirely instead. There aren't very many
uses.

Did I misunderstand your intent? What would you prefer?

[commit-bot: I haz the power, 2012-03-12 20:19:53]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/mdm@chromium.org/9696001/1

[commit-bot: I haz the power, 2012-03-12 23:30:16]

Try job failure for 9696001-1 (retry) on win_rel for step "browser_tests".
It's a second try, previously, steps "ui_tests, browser_tests" failed.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win_rel&nu...

[Mike Mammarella, 2012-03-14 00:07:39]

I give up on win_rel. It's failing with unrelated stuff every time. I'm just
going to commit and watch carefully.