Better handling of broken applet codebase

Better handling of broken applet codebase

A broken codebase would result in using an invalid url as base url in
the KURL constructor and that never results in anything useful (could
even crash). Since HTMLAppletElement is deprecated and badly documented
the goal of this patch is mostly to avoid the known badness and not to
make HTMLAppletElement perfect.

R=fs@opera.com

BUG=374110

[Daniel Bratell, 2015-04-08 13:30:52]

bratell@opera.com changed reviewers:
+ fs@opera.com, mkwst@chromium.org

[Daniel Bratell, 2015-04-08 13:30:53]

I had really planned on doing the big fix (fix kurl) but that was beyond the
grasp of anyone as long as chromium and blink are in different repositories.

But this will fix the main problems and it is the right thing to do as long as
applets are not dead.

[fs, 2015-04-08 13:59:17]

lgtm

Test?

[Daniel Bratell, 2015-04-08 15:56:28]

The CQ bit was checked by bratell@opera.com

[Daniel Bratell, 2015-04-08 15:56:28]

The patchset sent to the CQ was uploaded after l-g-t-m from fs@opera.com
Link to the patchset: https://codereview.chromium.org/315683002/#ps40001 (title:
"Now with a test")

[Mike West, 2015-04-10 11:24:31]

'<applet>'. :(

This patch and the test LGTM, thanks. Do we have similar issues for properties
on <object>? `data`, for instance?

[Daniel Bratell, 2015-04-10 11:34:28]

On 2015/04/10 11:24:31, Mike West wrote:
> '<applet>'. :(
> 
> This patch and the test LGTM, thanks. Do we have similar issues for properties
> on <object>? `data`, for instance?

This can only happen if you use an unchecked url ("codebase" in this case) as
base url when resolving an url. That seems rare and I remember not finding any
other cases when looking at this. Most methods just use document.completeURL()
and everything will be fine.

[Mike West, 2015-04-10 11:36:58]

Wunderbar. Thanks for fixing this instance of badness.

[Daniel Bratell, 2015-04-13 09:26:46]

The CQ bit was unchecked by bratell@opera.com

[Daniel Bratell, 2015-04-13 09:26:51]

The CQ bit was checked by bratell@opera.com

[Daniel Bratell, 2015-04-13 09:27:07]

The CQ bit was unchecked by bratell@opera.com

[Daniel Bratell, 2015-04-13 09:27:10]

The CQ bit was checked by bratell@opera.com

[Daniel Bratell, 2015-04-13 09:27:29]

Looks like CQ refuses to pick this one up.

[fs, 2015-04-13 09:30:02]

On 2015/04/13 09:27:29, Daniel Bratell wrote:
> Looks like CQ refuses to pick this one up.

Maybe try un-ticking and re-ticking the commit checkbox.

[Daniel Bratell, 2015-04-13 09:42:26]

On 2015/04/13 09:30:02, fs wrote:
> On 2015/04/13 09:27:29, Daniel Bratell wrote:
> > Looks like CQ refuses to pick this one up.
> 
> Maybe try un-ticking and re-ticking the commit checkbox.

I've clicked it some. No effect. I guess it's the age of the issue that does it.

[Daniel Bratell, 2015-04-13 09:46:10]

On 2015/04/13 09:42:26, Daniel Bratell wrote:
> On 2015/04/13 09:30:02, fs wrote:
> > On 2015/04/13 09:27:29, Daniel Bratell wrote:
> > > Looks like CQ refuses to pick this one up.
> > 
> > Maybe try un-ticking and re-ticking the commit checkbox.
> 
> I've clicked it some. No effect. I guess it's the age of the issue that does
it.

So it is. From the infra people: "New CLs have "Project" field assigned, which
is currently used by CQ to find CLs for a given project.".

I will reupload it and ask for a new lgtm.

[Daniel Bratell, 2015-04-13 09:48:22]

On 2015/04/13 09:46:10, Daniel Bratell wrote:
> On 2015/04/13 09:42:26, Daniel Bratell wrote:
> > On 2015/04/13 09:30:02, fs wrote:
> > > On 2015/04/13 09:27:29, Daniel Bratell wrote:
> > > > Looks like CQ refuses to pick this one up.
> > > 
> > > Maybe try un-ticking and re-ticking the commit checkbox.
> > 
> > I've clicked it some. No effect. I guess it's the age of the issue that does
> it.
> 
> So it is. From the infra people: "New CLs have "Project" field assigned, which
> is currently used by CQ to find CLs for a given project.".
> 
> I will reupload it and ask for a new lgtm.

https://codereview.chromium.org/1079953002