Closed Bug 855011 Opened 11 years ago Closed 11 years ago

Plugin CTP request: Match Firefox CTP for ESR 17.0.4-17.0.*

Categories

(Toolkit :: Blocklist Policy Requests, defect)

defect
Not set
normal

Tracking

()

RESOLVED FIXED
2013-04-25
Tracking Status
firefox-esr17 21+ fixed

People

(Reporter: akeybl, Assigned: jorgev)

Details

(Whiteboard: [plugin])

Flash:
p176
p178
p260
p290
p330
p332 

Java (already blocked on 17 and up)

Silverlight:
p328

Adobe Reader (already blocked)

Once staged, we can test internally and also vend to the enterprise list. Once they've had the opportunity to test and react, we can roll these blocks out.
(In reply to Jorge Villalobos [:jorgev] from comment #1)
> All of the corresponding block updates have been staged:
> 
> https://addons-dev.allizom.org/firefox/blocked/p176
> https://addons-dev.allizom.org/firefox/blocked/p178
> https://addons-dev.allizom.org/firefox/blocked/p267
> https://addons-dev.allizom.org/firefox/blocked/p281
> https://addons-dev.allizom.org/firefox/blocked/p305
> https://addons-dev.allizom.org/firefox/blocked/p307

All of these blocks are working as expected on staging with Firefox 17.0.4esr.

> https://addons-dev.allizom.org/firefox/blocked/p303

Matt, since I don't currently have access to my Mac can you please help me test the final block? Please use Silverlight 5.1.10411.0, Firefox 17.0.4esr, and Netflix to test whether click-to-play blocking is working. I sent you instructions via email just in case.

Thank you
Matt notified me that he was having problems getting the Silverlight block to work.

Paul, can you give it a try tonight? If you can't use Netflix do to geolocation please try some other Silverlight pages. Again, we're looking to see that Silverlight 5.1 and 5.1.20124.* is CTP blocked on Mac.

Thanks
Everything looks good on staging:
Silverlight 5.1.10411.0 is CTP blocked.
Silverlight 5.1.20125.0 is not blocked.

Tested with some silverlight games (http://www.vectorlight.net/silverlight/games.aspx), Netflix is N/A outside of the US.
Keywords: qawanted
Thank you Paul. I'm fairly confident we can push this live as soon as we want.
(In reply to Anthony Hughes, Mozilla QA (:ashughes) from comment #5)
> Thank you Paul. I'm fairly confident we can push this live as soon as we
> want.

Of course we should wait for Enterprises to test this internally before doing so.
(In reply to Anthony Hughes, Mozilla QA (:ashughes) from comment #6)
> (In reply to Anthony Hughes, Mozilla QA (:ashughes) from comment #5)
> > Thank you Paul. I'm fairly confident we can push this live as soon as we
> > want.
> 
> Of course we should wait for Enterprises to test this internally before
> doing so.

Doing this now, let's target pushing out these blocks the week of 4/8 to give enterprise IT time to test.
I'll push the blocks live tomorrow morning if there are no objections.
(In reply to Jorge Villalobos [:jorgev] from comment #8)
> I'll push the blocks live tomorrow morning if there are no objections.

Tomorrow we will be doing a quick 21b2 turnaround so I won't have time to test any live blocklist pushes. I'll put on the table three options (in order of preference):

1. Delay push until Wednesday morning and I will test
2. Push late tomorrow afternoon and Paul at Softvision can test overnight
3. Push tomorrow morning and get Enterprise testers to confirm without QA sign-off
I'm okay with waiting till Wednesday.
The blocks are now live.
Status: NEW → RESOLVED
Closed: 11 years ago
Resolution: --- → FIXED
Sorry for the delay, this took me most of the day to test. All of the blocks are working in production except for two (both on Mac OSX 10.8 with Firefox 17.0.5esr):

* Flash 10.1.85.3 is showing an info-bar (should be CTP via P267)
* Silverlight 5.1.10411.0 is not blocked at all (should be CTP via P303)

In both of these cases I'm not seeing the respective entries in my blocklist.xml file.

Reopening for further investigation.
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
(In reply to Anthony Hughes, Mozilla QA (:ashughes) from comment #12)
> * Silverlight 5.1.10411.0 is not blocked at all (should be CTP via P303)

This one I just forgot about. I just pushed it live.

> * Flash 10.1.85.3 is showing an info-bar (should be CTP via P267)

In this case I see the block in blocklist.xml. Note that 267 is the ID on staging, and 260 is the ID in production.

Looking at blocklist.xml, I see something that might be a problem, though. There are 2 entries with <pluginItem  blockID="p260">, one for each version range. I think there should only be one of those, with both filters.

Blair, can you confirm that having these nodes with the same ID cause the blocklist to ignore one of the version ranges?
Flags: needinfo?(bmcbride)
(In reply to Jorge Villalobos [:jorgev] from comment #13)
> Looking at blocklist.xml, I see something that might be a problem, though.
> There are 2 entries with <pluginItem  blockID="p260">, one for each version
> range. I think there should only be one of those, with both filters.
> 
> Blair, can you confirm that having these nodes with the same ID cause the
> blocklist to ignore one of the version ranges?

This shouldn't cause any problem - the blocklist service doesn't use the blockID attribute to identify blocks, and we make no assumptions that the value is unique (it's only use is to construct a URL).
Flags: needinfo?(bmcbride)
(In reply to Anthony Hughes, Mozilla QA (:ashughes) from comment #12)
> * Flash 10.1.85.3 is showing an info-bar (should be CTP via P267)
I confirm this issue on ESR 17.0.1 - 17.0.5 on Mac OS X only
(In reply to Jorge Villalobos [:jorgev] from comment #13)
> (In reply to Anthony Hughes, Mozilla QA (:ashughes) from comment #12)
> > * Silverlight 5.1.10411.0 is not blocked at all (should be CTP via P303)
> 
> This one I just forgot about. I just pushed it live.

Verified fixed, thanks.
(In reply to Paul Silaghi [QA] from comment #15)
> (In reply to Anthony Hughes, Mozilla QA (:ashughes) from comment #12)
> > * Flash 10.1.85.3 is showing an info-bar (should be CTP via P267)
> I confirm this issue on ESR 17.0.1 - 17.0.5 on Mac OS X only

Yes, it looks like this is really a conflict with p94, which is an infobar block that covers 0-17.* and Flash 0 - 10.2.159.1.

Do we care? One possible solution would be to change it from 0-17.* to 0-17.0.3, so there is no overlap.
Flags: needinfo?(akeybl)
(In reply to Jorge Villalobos [:jorgev] from comment #17)
> (In reply to Paul Silaghi [QA] from comment #15)
> > (In reply to Anthony Hughes, Mozilla QA (:ashughes) from comment #12)
> > > * Flash 10.1.85.3 is showing an info-bar (should be CTP via P267)
> > I confirm this issue on ESR 17.0.1 - 17.0.5 on Mac OS X only
> 
> Yes, it looks like this is really a conflict with p94, which is an infobar
> block that covers 0-17.* and Flash 0 - 10.2.159.1.
> 
> Do we care? One possible solution would be to change it from 0-17.* to
> 0-17.0.3, so there is no overlap.

We don't care about removing an pre-existing infobar block - sounds like we're good to go.
Flags: needinfo?(akeybl)
I modified the infobar block so that there's no overlap. Please retest after an hour or so.
Keywords: qawanted
Status: REOPENED → RESOLVED
Closed: 11 years ago11 years ago
Resolution: --- → FIXED
Keywords: qawantedverifyme
I think you made a mistake in pulling back the infobar block. Tested with Flash 10.2.153.1 on Mac OSX 10.8 on Youtube.com:

Firefox 17.0.4esr: click-to-play blocked
Firefox 17.0.3esr: not blocked
Firefox 17.0.2esr: not blocked
Firefox 17.0.1esr: infobar blocked

Looking at the block, p94 it shows:
> <versionRange>minVersion="0.1" maxVersion="17.0.1">

I think this should be fixed to be maxVersion="17.0.3". I'm not sure whether you want to reopen this bug or do that in a new bug since the CTP block is working as expected.
Keywords: verifyme
I moved it back to 17.0.1 because I think that's the only other 17.* non-ESR release we made. That's at least based on looking at our releases FTP site. I doubt it's a big deal either way.
Assignee: nobody → jorge
Target Milestone: --- → 2013-04-25
Product: addons.mozilla.org → Toolkit
You need to log in before you can comment on or make changes to this bug.