Ensure pdf plugin sends DidStartLoading() consistently

Ensure pdf plugin sends DidStartLoading() consistently

There are situations where the pdf plugin will initialize itself in such a way
that it never notifies content/renderer/ that it has started. In so doing, it
also fails to send a stop message, since it tries to ensure it doesn't send
mismatched start/stop messages.

Print preview decides when to display the preview based on when loading
stops. Since this only considers the page's content and not the pdf plugin's
initialization, there is a race condition where we may try to preview before
the plugin is ready. In this cases, we show "Loading preivew..." and never load
the actual preview.

There is a hack in place in print_preview_web_helper.cc to delay the process of
showing the preview to maximize the chance that the pdf plugin will be ready.
This change should allow us to rip that out.

BUG=376969
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=283837

[Nate Chapin, 2014-07-16 18:13:40]

vitalybuka: Please review the revert of the nasty hack in
print_web_view_helper.cc :)

jam: Please review the src/pdf change. This adds a DidStartLoading() in a case
we were missing before.

[Vitaly Buka (NO REVIEWS), 2014-07-16 18:32:45]

lgtm

https://codereview.chromium.org/397713005/diff/20001/pdf/instance.cc
File pdf/instance.cc (right):

https://codereview.chromium.org/397713005/diff/20001/pdf/instance.cc#newcode353
pdf/instance.cc:353: DCHECK(!did_call_start_loading_);
Maybe CHECK() to make crash in release, so we could see reports

[jam, 2014-07-16 21:25:43]

https://codereview.chromium.org/397713005/diff/20001/pdf/instance.cc
File pdf/instance.cc (right):

https://codereview.chromium.org/397713005/diff/20001/pdf/instance.cc#newcode355
pdf/instance.cc:355: did_call_start_loading_ = true;
I'm a bit confused, I tried doing window.print() (per the code being edited in
the other file). I hit line 348 twice in the renderer for print preview, but
both times the full_ was false. how is this else statement being reached?

[Nate Chapin, 2014-07-16 21:39:06]

On 2014/07/16 21:25:43, jam wrote:
> https://codereview.chromium.org/397713005/diff/20001/pdf/instance.cc
> File pdf/instance.cc (right):
> 
>
https://codereview.chromium.org/397713005/diff/20001/pdf/instance.cc#newcode355
> pdf/instance.cc:355: did_call_start_loading_ = true;
> I'm a bit confused, I tried doing window.print() (per the code being edited in
> the other file). I hit line 348 twice in the renderer for print preview, but
> both times the full_ was false. how is this else statement being reached?

Try the repro steps in
https://code.google.com/p/chromium/issues/detail?id=376969

[jam, 2014-07-17 17:23:54]

thanks, i wasn't trying those steps. lgtm

[jam, 2014-07-17 17:23:59]

The CQ bit was checked by jam@chromium.org

[commit-bot: I haz the power, 2014-07-17 17:25:13]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/japhet@chromium.org/397713005/20001

[commit-bot: I haz the power, 2014-07-17 19:19:31]

Change committed as 283837