1380015 - PDF Viewer shows blank PDF when the document was scanned as jpeg

Status: VERIFIED FIXED

(Firefox :: PDF Viewer, defect)

Description

[Zilmar]

User Agent: Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:54.0) Gecko/20100101 Firefox/54.0
Build ID: 20170628075643

Steps to reproduce:

Open a PDF scanned as jpeg
Example: http://www.pucsp.br/sites/default/files/download/calendario_2017_r5.pdf


Actual results:

PDF document show only blank pages


Expected results:

PDF should be rendered properly

Comment 1

[Tooru Fujisawa [:arai]]

reproduced on 54.0.1, but not reproduced on Nightly 56.0a1 (2017-07-11) (64-bit)
maybe already fixed?

Comment 2

[Zilmar]

On Nightly it works like a charm. But I couldn't find any references to this issue. 
I heard about this bug in SUMO (https://support.mozilla.org/pt-BR/questions/1167256), it seems that the user has migrated to Nightly.

Comment 3

[YF (Yang)]

Web Console:
Error: JPEG error: unexpected marker ffd9  pdf.worker.js:345:3
......

Bug 1350416:
#8170 Try to recover when encountering JPEG markers with too short marker lengths (issue 8169)

Comment 4

[Ryan VanderMeulen [:RyanVM]]

This was caused by bug 1306522, but I can't set it as a blocker due to circular dependencies.

The fix in question doesn't look too bad, but I'm not sure this issue is severe enough to warrant backporting a fix to ESR52 for anyway. Any thoughts, Jonas?

Comment 5

[Jonas Jenwald [:Snuffleupagus]]

(In reply to YF (Yang) from comment #3)
> Web Console:
> Error: JPEG error: unexpected marker ffd9  pdf.worker.js:345:3
> ......
> 
> Bug 1350416:
> #8170 Try to recover when encountering JPEG markers with too short marker
> lengths (issue 8169)

I'm not convinced that PR 8170 is actually what fixed this particular bug. I've not bisected this upstream, but PR 8164 seems like a more probable candidate to me.

(In reply to Ryan VanderMeulen [:RyanVM] from comment #4)
> This was caused by bug 1306522, but I can't set it as a blocker due to
> circular dependencies.
> 
> The fix in question doesn't look too bad, but I'm not sure this issue is
> severe enough to warrant backporting a fix to ESR52 for anyway. Any
> thoughts, Jonas?

First of all, if we decide to backport PDF.js JPEG fixes to ESR52, then there's actually a few upstream patches that we'd want to include. Since there's always a risk that those patches could cause other regressions, and given then nature of ESR releases, I'm just not convinced that it's worth the risk.
So, in my opinion, I don't think that we need to backport this to ESR52.

Comment 6

[Ryan VanderMeulen [:RyanVM]]

Sounds good, thanks for the reply.

Comment 7

[Oana Botisan, Desktop QA]

* Last good: Nightly 52.0a1 from 2016-09-30 (20160930030315, 59db3b0781908d9e2145477ff8dc5a85729e25bf)
* First bad: Nightly 52.0a1 from 2016-10-01 (20161002030227, 87cd291d2db621da6b3eb1057027cc0725b6eb1d)
* Pushlog: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=59db3b0781908d9e2145477ff8dc5a85729e25bf&tochange=87cd291d2db621da6b3eb1057027cc0725b6eb1d

This bug persisted for a time. It was fixed again only in the Nightly from 2017-03-25. But now it doesn't seem to reproduce anymore and I tested it both on the latest Nightly and Beta 55.0b13.

Comment 8

[Bogdan Maris, Desktop QA]

Thanks for verifying, marking accordingly.