966132 - [b2g][pdfjs] gaia master fails to open any pdf with TypeError

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[Ben Kelly [:bkelly, not reviewing]]

Trying to open a pdf with gaia rev:

  aedd5c9636f305d4433491056a0ca984dfb859b1

I get the following and the progress bar never moves:

  E/GeckoConsole( 2256): [JavaScript Error: "setting a property that has only a getter" {file: "app://pdfjs.gaiamobile.org/content/build/pdf.worker.js" line: 34495}]

This does not occur on v1.3.

Comment 1

[Jason Smith [:jsmith]]

So this isn't going to be a Gaia regression - we haven't changed version of pdf.js since Oct 21st. The line of code failing is:

globalScope.console = workerConsole;

My guess is that this is a DOM regression, so let's start there.

Will work on getting a regression window to help diagnose this though.

Comment 2

[Kyle Huey (Exited; not receiving bugmail, old account, do not use)]

I'm going to put money on bug 620935 being the cause, although I suspect the fix is in gaia.

Comment 3

[Boris Zbarsky [:bzbarsky]]

Jason, thank you for looking up the relevant code bit!

On main thread, window.console is a replaceable property.  I would say we should do that in workers as well.  Then the code in question will just work (and override the built-in console, but that's the whole point of making it replaceable).

Comment 4

[Andrea Marchesini [:baku]]

Attachment: worker.patch

Comment 5

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 8368535 [details] [diff] [review]
worker.patch

>+    if (!event.data.status || event.data.last)

Why do you need that first condition?  If it tests true, won't we double-finish the test?  I think we should finish() here only if event.data.last.

r=me with that.

Comment 6

[Andrea Marchesini [:baku]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/c108d4419d9a

Comment 7

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/c108d4419d9a