More progress on ForceHTTPS....

More progress on ForceHTTPS.

Instead of turning on strict HTTPS error processing for every site, we now track which sites have opted in.  Our implementation is still experimental and hidden behing the command line switch --force-https.

R=darin
TEST=No tests yet because this is just an experiment.


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=16464

[darin (slow to review), 2009-05-19 21:25:49]

LGTM

http://codereview.chromium.org/113503/diff/1037/54
File chrome/browser/profile.cc (right):

http://codereview.chromium.org/113503/diff/1037/54#newcode41
Line 41: #include "net/base/force_tls_state.h"
nit: "g" before "n"

http://codereview.chromium.org/113503/diff/1037/54#newcode175
Line 175: virtual net::ForceTLSState* GetForceTLSState() {
just a meta comment:  kind a sad that we do not have an easy way to share
implementations of things like this between the normal profile and the incognito
one.

http://codereview.chromium.org/113503/diff/1037/52
File chrome/browser/profile.h (right):

http://codereview.chromium.org/113503/diff/1037/52#newcode25
Line 25: }  // namepsace net
nit: s/namepsace/namespace/ ... but better yet, just kill the comment.  it is
rather redundant when the namespace spans so few lines.

http://codereview.chromium.org/113503/diff/1037/52#newcode134
Line 134: // Retrieves a pointer to the ForceTLSSate associated with this
profile.
s/ForceTLSSate/ForceTLSState/

http://codereview.chromium.org/113503/diff/1037/52#newcode135
Line 135: // The ForceTLSState is lazily created the first time that htis method
is
s/htis/this/

http://codereview.chromium.org/113503/diff/1037/57
File chrome/browser/ssl/ssl_policy_backend.cc (right):

http://codereview.chromium.org/113503/diff/1037/57#newcode14
Line 14: #include "net/base/force_tls_state.h"
nit: "g" before "n"

http://codereview.chromium.org/113503/diff/1037/58
File chrome/browser/ssl/ssl_policy_backend.h (right):

http://codereview.chromium.org/113503/diff/1037/58#newcode18
Line 18: }  // namespace net
nit: ditto about comment

http://codereview.chromium.org/113503/diff/1037/58#newcode121
Line 121: // ForceTLS state.
also kind of sucks to have code mentioning TLS and SSL.  might be nice if we
unified on only one of those terms.

http://codereview.chromium.org/113503/diff/1037/50
File net/base/force_tls_state.cc (right):

http://codereview.chromium.org/113503/diff/1037/50#newcode16
Line 16: void ForceTLSState::OnReceiveHeader(const GURL& url, const std::string&
value) {
nit: DidReceiveHeader sounds a little better to me.  maybe i'm just too used to
On* corresponding to an event handler.

might want to add a TODO here about parsing the value to extract expiry info,
etc.

http://codereview.chromium.org/113503/diff/1037/50#newcode21
Line 21: // TODO(abarth): Cannonicalize host.
nit: Canonicalize

[abarth-chromium, 2009-05-20 01:23:40]

Thanks Darin.

http://codereview.chromium.org/113503/diff/1037/58
File chrome/browser/ssl/ssl_policy_backend.h (right):

http://codereview.chromium.org/113503/diff/1037/58#newcode121
Line 121: // ForceTLS state.
On 2009/05/19 21:25:49, darin wrote:
> also kind of sucks to have code mentioning TLS and SSL.  might be nice if we
> unified on only one of those terms.

Yeah.  I expect the Force-TLS name to change before its done.