PDF is tagged, Reads with "Read Out Loud" but not with NVDA or Jaws #7021

raebened · 2017-03-27T16:34:43Z

I have a series of PDFs made with Indesign that are tagged. I'm cleaning them up and completed one that passes all checks (Acrobat, PAC and CommonLook) and yet will not pass a simple screen reader test. It never loads and reads as "blank."

This document came to me already tagged, but needed significant cleanup. However, it won't read at all even before I work on it, so I don't believe it's anything I'm doing.

I've tried various techniques none of which work:

removing all the form fields
removing graphics
removing tags and retagging automatically
saving as Word and remaking PDF with tags (even this did not work!)
using Acrobat to create PDF from file -- this results in a "screen-readable" document but the text is scrambled and meaningless as if the document were encoded.

This file is using embedded Truetype (CID) fonts. The file was originally created in 2013 on a Mac. The other files in this set that do not have this problem were originally created in 2014.

I'm quite puzzled and cannot resolve this problem.

Brian1Gaff · 2017-03-27T17:00:38Z

I've seen pdfs like this, strangely its got worse since a recent adobe reader update. No idea what is going on unless some protected mode is in use not obvious to the user. I think pdfs are getting so messy and mis used of late it makes me simple steer clear and insist on word or text copies. Brian bglists@blueyonder.co.uk Sent via blueyonder. Please address personal email to:- briang1@blueyonder.co.uk, putting 'Brian Gaff' in the display name field. ----- Original Message ----- From: "raebened" <notifications@github.com> To: "nvaccess/nvda" <nvda@noreply.github.com> Cc: "Subscribed" <subscribed@noreply.github.com> Sent: Monday, March 27, 2017 5:34 PM Subject: [nvaccess/nvda] PDF is tagged, Reads with "Read Out Loud" but not with NVDA or Jaws (#7021)

…

I have a series of PDFs made with Indesign that are tagged. I'm cleaning them up and completed one that passes all checks (Acrobat, PAC and CommonLook) and yet will not pass a simple screen reader test. It never loads and reads as "blank." This document came to me already tagged, but needed significant cleanup. However, it won't read at all even before I work on it, so I don't believe it's anything I'm doing. I've tried various techniques none of which work: - removing all the form fields - removing graphics - removing tags and retagging automatically - saving as Word and remaking PDF with tags (even this did not work!) - using Acrobat to create PDF from file -- this results in a "screen-readable" document but the text is scrambled and meaningless as if the document were encoded. This file is using embedded Truetype (CID) fonts. The file was originally created in 2013 on a Mac. The other files in this set that do not have this problem were originally created in 2014. I'm quite puzzled and cannot resolve this problem. -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: #7021

josephsl · 2017-03-27T17:26:37Z

Hi, have you attempted putting through OCR? Thanks. CC @jcsteh From: raebened [mailto:notifications@github.com] Sent: Monday, March 27, 2017 9:35 AM To: nvaccess/nvda <nvda@noreply.github.com> Cc: Subscribed <subscribed@noreply.github.com> Subject: [nvaccess/nvda] PDF is tagged, Reads with "Read Out Loud" but not with NVDA or Jaws (#7021) I have a series of PDFs made with Indesign that are tagged. I'm cleaning them up and completed one that passes all checks (Acrobat, PAC and CommonLook) and yet will not pass a simple screen reader test. It never loads and reads as "blank." This document came to me already tagged, but needed significant cleanup. However, it won't read at all even before I work on it, so I don't believe it's anything I'm doing. I've tried various techniques none of which work: * removing all the form fields * removing graphics * removing tags and retagging automatically * saving as Word and remaking PDF with tags (even this did not work!) * using Acrobat to create PDF from file -- this results in a "screen-readable" document but the text is scrambled and meaningless as if the document were encoded. This file is using embedded Truetype (CID) fonts. The file was originally created in 2013 on a Mac. The other files in this set that do not have this problem were originally created in 2014. I'm quite puzzled and cannot resolve this problem. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#7021> , or mute the thread <https://github.com/notifications/unsubscribe-auth/AHgLkOGgMpYeLO8ZW_k2B3ipwl2OHrRhks5rp-UkgaJpZM4MqhDT> .

raebened · 2017-03-27T19:38:29Z

Thanks josephsl but there is text in the document. I know "blank" is the kind of error you get when a document is just a scanned file but this file has text, I've tagged it. I can select it in the PDF, I can save it as a Word file and the text is all there.

Just for the heck I did run OCR and it says it can't do it because there is text on the page. I continued anyway to see if it would change something but it did not.

josephsl · 2017-03-27T19:44:26Z

Hi, when the PDF is opened and in foreground with NVDA running, can you press NvDA+SPACE to see if you can use browse mode and focus mode? Thanks. From: raebened [mailto:notifications@github.com] Sent: Monday, March 27, 2017 12:39 PM To: nvaccess/nvda <nvda@noreply.github.com> Cc: Joseph Lee <joseph.lee22590@gmail.com>; Comment <comment@noreply.github.com> Subject: Re: [nvaccess/nvda] PDF is tagged, Reads with "Read Out Loud" but not with NVDA or Jaws (#7021) Thanks josephsl but there is text in the document. I know "blank" is the kind of error you get when a document is just a scanned file but this file has text, I've tagged it. I can select it in the PDF, I can save it as a Word file and the text is all there. Just for the heck I did run OCR and it says it can't do it because there is text on the page. I continued anyway to see if it would change something but it did not. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#7021 (comment)> , or mute the thread <https://github.com/notifications/unsubscribe-auth/AHgLkDhJFVaaRiaFxAiwHu-Mte7K1GVnks5rqBA3gaJpZM4MqhDT> .

jcsteh · 2017-03-27T20:17:29Z

We cannot diagnose this without the PDF in question. I understand you may not be able to share it, but if that is the case, there is nothing we can do here.

raebened · 2017-03-28T15:52:19Z

SAFER_Contingency_Planning.pdf

Here is the original file before I added read only form fields for the body text on form pages. It's the simplest version I have that should help eliminate the source of the problem. Thank.

jcsteh · 2017-03-29T23:37:28Z

I get a crash when I try to open this PDF in Adobe Reader DC with NVDA running. A colleague sees the same issue. I'll investigate.

raebened · 2017-03-30T00:51:47Z

Thanks.

…

On Mar 29, 2017 7:37 PM, "James Teh" ***@***.***> wrote: I get a crash when I try to open this PDF in Adobe Reader DC with NVDA running. A colleague sees the same issue. I'll investigate. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#7021 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ASt_48-pQAflAk43QnoJEnvojLTMMXRCks5rqus6gaJpZM4MqhDT> .

PratikP1 · 2017-03-30T04:24:54Z

FYI, I am also getting a crash with NVDA. In addition, I get a crash with JAWS 18 as well.

jcsteh · 2017-03-30T05:25:10Z

Yeah, I think I'm going to need to work with Adobe on this one. Technical: Getting the IPDDomNode for one of the nodes in this document fails with E_FAIL for some reason and returns null. The initial crash is because NVDA doesn't check for this and tries to dereference it. It probably shouldn't be failing anyway, but that doesn't excuse the lack of a check. Even once I guard against this, though, Reader then crashes when we try to query accessibility states (IAccessible::accState). This is a bug in Reader code and I don't see how we can work around it. I'm going to make a crash dump and ask Adobe to look into it.

jcsteh · 2017-03-30T05:31:28Z

The null pointer check is in the branch i7021ReaderCrash.

jcsteh · 2017-03-31T05:49:40Z

I've fixed the first crash in #7035, as #7034 reported the same crash (and it does fix that one).

Regarding the second crash in Reader itself that occurs with the PDF you provided, I've reported this to Adobe with a crash dump. Hopefully, they'll be able to track it down.

raebened · 2017-03-31T12:10:23Z

Thanks so much for looking at this file. I really appreciate it. Have tried everything I could . Rae Benedetto

…

On Fri, Mar 31, 2017 at 1:49 AM, James Teh ***@***.***> wrote: I've fixed the first crash in #7035 <#7035>, as #7034 <#7034> reported the same crash (and it does fix that one). Regarding the second crash in Reader itself that occurs with the PDF you provided, I've reported this to Adobe with a crash dump. Hopefully, they'll be able to track it down. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#7021 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ASt_4_8h15CJ-8WDLJhxq3U-ab7TVuD-ks5rrJP2gaJpZM4MqhDT> .

-- Rae Benedetto Accessibility and Remediation Specialist 9907 Georgetown Pike Great Falls, VA 22066 703 431 8206 raebenedetto@gmail.com

…butes. Incubates #7035 (issue #7034, #7021).

jcsteh · 2017-04-12T06:48:35Z

@jcsteh commented on 31 Mar. 2017, 3:49 pm AEST:

Regarding the second crash in Reader itself that occurs with the PDF you provided, I've reported this to Adobe with a crash dump. Hopefully, they'll be able to track it down.

Adobe have logged this internally as bug ADC-4210565. I don't have any further info at this stage.

raebened · 2017-04-12T16:05:24Z

Thanks, I put it aside for a bit and came back to it today. Appears that the problem is in the first 6 or so pages. When I delete those pages (*after* clean up using Axes Quick Fix and removal of all FormXObjects) the remaining pages appear to read ok. I'm working on that now. This would save me a lot of time if it, in fact, works. I had extracted all the pages as separate files and found they read individually, except for page 5. Unfortunately extracting the pages using the extract feature deletes the tags and messes up the forms fields considerably, while deleting pages leaves tags and fields intact for the remaining pages. Thanks for your help with this. Rae

…

On Wed, Apr 12, 2017 at 2:48 AM, James Teh ***@***.***> wrote: ***@***.**** <https://github.com/jcsteh> commented on 31 Mar. 2017, 3:49 pm AEST <#7021 (comment)> : Regarding the second crash in Reader itself that occurs with the PDF you provided, I've reported this to Adobe with a crash dump. Hopefully, they'll be able to track it down. Adobe have logged this internally as bug ADC-4210565. I don't have any further info at this stage. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#7021 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ASt_46wEX3kbmabpDSf6ISBx1cWHAIHAks5rvHPFgaJpZM4MqhDT> .

-- Rae Benedetto Accessibility and Remediation Specialist 9907 Georgetown Pike Great Falls, VA 22066 703 431 8206 raebenedetto@gmail.com

…ifically, those containing empty ActualText attributes). (PR #7035; issues #7021, #7034) * Adobe Acrobat Reader no longer crashes in certain PDF documents (specifically, those containing empty ActualText attributes). adobeAcrobat vbuf backend: get_PDDomNode can fail and return null, so check for a null pointer before trying to use it. * Address review comments.

michaelDCurran · 2017-10-02T04:51:33Z

Testing with Adobe Reader 2017 Release | Version 2017.012.20098 I can no longer reproduce the crash with that pdf.
@raebened: can you still reproduce it?

feerrenrut · 2017-10-17T02:11:16Z

Given that we have had no response, we will now close this issue. Please comment or re-open if this can still be reproduced.

jcsteh added AdobeWork bug/app-crash p2 https://github.com/nvaccess/nvda/blob/master/projectDocs/issues/triage.md#priority labels Mar 29, 2017

jcsteh self-assigned this Mar 29, 2017

This was referenced Mar 31, 2017

NVDA PDF Actual Text bug #7034

Closed

Fix Acrobat Reader crash in PDF documents with empty ActualText attributes #7035

Merged

jcsteh added blocked blocked/needs-external-fix labels Mar 31, 2017

jcsteh added a commit that referenced this issue Apr 4, 2017

Fix Acrobat Reader crash in PDF documents with empty ActualText attri…

2c9709e

…butes. Incubates #7035 (issue #7034, #7021).

jcsteh removed their assignment Sep 5, 2017

feerrenrut closed this as completed Oct 17, 2017

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

PDF is tagged, Reads with "Read Out Loud" but not with NVDA or Jaws #7021

PDF is tagged, Reads with "Read Out Loud" but not with NVDA or Jaws #7021

raebened commented Mar 27, 2017

Brian1Gaff commented Mar 27, 2017 via email

josephsl commented Mar 27, 2017 via email

raebened commented Mar 27, 2017

josephsl commented Mar 27, 2017 via email

jcsteh commented Mar 27, 2017 via email

raebened commented Mar 28, 2017

jcsteh commented Mar 29, 2017

raebened commented Mar 30, 2017 via email

PratikP1 commented Mar 30, 2017 via email

jcsteh commented Mar 30, 2017 via email

jcsteh commented Mar 30, 2017

jcsteh commented Mar 31, 2017

raebened commented Mar 31, 2017 via email

jcsteh commented Apr 12, 2017

raebened commented Apr 12, 2017 via email

michaelDCurran commented Oct 2, 2017

feerrenrut commented Oct 17, 2017

PDF is tagged, Reads with "Read Out Loud" but not with NVDA or Jaws #7021

PDF is tagged, Reads with "Read Out Loud" but not with NVDA or Jaws #7021

Comments

raebened commented Mar 27, 2017

Brian1Gaff commented Mar 27, 2017 via email

josephsl commented Mar 27, 2017 via email

raebened commented Mar 27, 2017

josephsl commented Mar 27, 2017 via email

jcsteh commented Mar 27, 2017 via email

raebened commented Mar 28, 2017

jcsteh commented Mar 29, 2017

raebened commented Mar 30, 2017 via email

PratikP1 commented Mar 30, 2017 via email

jcsteh commented Mar 30, 2017 via email

jcsteh commented Mar 30, 2017

jcsteh commented Mar 31, 2017

raebened commented Mar 31, 2017 via email

jcsteh commented Apr 12, 2017

raebened commented Apr 12, 2017 via email

michaelDCurran commented Oct 2, 2017

feerrenrut commented Oct 17, 2017