Apply sizeof to variables instead of types.

Apply sizeof to variables instead of types.

Replace hardcoded variable sizes with sizeof.

R=eroman@chromium.org
BUG=none

Committed: https://crrev.com/4e3ae14860d84d92ad89fe5afea84fa2ba77f0d2
Cr-Commit-Position: refs/heads/master@{#333117}

[wtc, 2015-06-04 00:57:29]

Eric: this CL makes the changes you suggested in
https://codereview.chromium.org/1158923005/#msg6

https://codereview.chromium.org/1170493002/diff/1/net/cert/crl_set_storage.cc
File net/cert/crl_set_storage.cc (right):

https://codereview.chromium.org/1170493002/diff/1/net/cert/crl_set_storage.cc...
net/cert/crl_set_storage.cc:269: data->remove_prefix(sizeof(uint8_t));

See similar code on lines 162-166.

[eroman, 2015-06-04 22:25:32]

lgtm

https://codereview.chromium.org/1170493002/diff/1/net/cert/crl_set_storage.cc
File net/cert/crl_set_storage.cc (right):

https://codereview.chromium.org/1170493002/diff/1/net/cert/crl_set_storage.cc...
net/cert/crl_set_storage.cc:154: data->remove_prefix(sizeof(num_serials));
this is a behavioral change from before.
that said, looks fine to me.

https://codereview.chromium.org/1170493002/diff/1/net/cert/crl_set_storage.cc...
net/cert/crl_set_storage.cc:156: if (num_serials > 32 * 1024 * 1024)  // Sanity
check.
I am curious if you know where this number comes from (i.e. is to bound the
total memory usage)
3.1 million is a lot of serial numbers

https://codereview.chromium.org/1170493002/diff/1/net/cert/crl_set_storage.cc...
net/cert/crl_set_storage.cc:266: if (data->size() < sizeof(uint8_t))
Consider changing this line too.

[wtc, 2015-06-05 17:39:33]

https://codereview.chromium.org/1170493002/diff/1/net/cert/crl_set_storage.cc
File net/cert/crl_set_storage.cc (right):

https://codereview.chromium.org/1170493002/diff/1/net/cert/crl_set_storage.cc...
net/cert/crl_set_storage.cc:154: data->remove_prefix(sizeof(num_serials));

On 2015/06/04 22:25:31, eroman wrote:
> this is a behavioral change from before.
> that said, looks fine to me.

Confirmed. I think the new code is better (it immediately removes from |data|
the bytes that have been copied out) and is more consistent with the rest of
the parsing code.

https://codereview.chromium.org/1170493002/diff/1/net/cert/crl_set_storage.cc...
net/cert/crl_set_storage.cc:156: if (num_serials > 32 * 1024 * 1024)  // Sanity
check.

On 2015/06/04 22:25:31, eroman wrote:
> I am curious if you know where this number comes from (i.e. is to bound the
> total memory usage)
> 3.1 million is a lot of serial numbers

I don't know where this number comes from. It is a reasonable maximum CRLSet
size and may have been chosen aribitrarily to bound the total memory usage or
download time.

https://codereview.chromium.org/1170493002/diff/1/net/cert/crl_set_storage.cc...
net/cert/crl_set_storage.cc:266: if (data->size() < sizeof(uint8_t))

On 2015/06/04 22:25:31, eroman wrote:
> Consider changing this line too.

When reading just one byte, I would just use 1. This is why I don't want to
go too far and make the code less readable.

If we ever change this in the future, please also change lines 162-166.

[wtc, 2015-06-05 20:18:23]

The CQ bit was checked by wtc@chromium.org

[commit-bot: I haz the power, 2015-06-05 20:19:04]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1170493002/1

[commit-bot: I haz the power, 2015-06-05 20:24:21]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-06-05 20:25:19]

Patchset 1 (id:??) landed as
https://crrev.com/4e3ae14860d84d92ad89fe5afea84fa2ba77f0d2
Cr-Commit-Position: refs/heads/master@{#333117}