DO NOT REPLY [Bug 46705] New: Enhancement: PDF Accessibility

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705


Jeremias Maerki <jeremias@...> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #23262|0                           |1
           is patch|                            |
  Attachment #23262|0                           |1
        is obsolete|                            |




--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705





--- Comment #16 from Jeremias Maerki <jeremias@...>  2009-02-19 11:58:34 PST ---
I've applied the patch (2009-02-19 02:23 PST) to a new branch as discussed
(with modifications). The code is now under:
https://svn.apache.org/repos/asf/xmlgraphics/fop/branches/Temp_Accessibility
http://svn.apache.org/viewvc?rev=745949&view=rev

I'll continue working on the code when my batteries are filled again.

Jost, thanks a lot for your work here. This is great new functionality. It will
still take a bit to polish a few areas. I hope to see you become a regular
contributor now that the hardest part is over. Please don't hesitate to publish
follow-up patches against the branch (or trunk if you have other ideas). I'll
keep the issue here open until we've adressed all the feedback, after which
we'll see to merging this into trunk.

--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705





--- Comment #17 from Jeremias Maerki <jeremias@...>  2009-04-17 07:42:17 PST ---
Most of the feedback has now been incorporated into the development branch. One
notable exception is the discussion around replacing foi:ptr with the "id"
property but I'm running out of time to work on this. Another point that I'm
not entirely happy with is the fact that the enriched FO is fully buffered in a
byte array before FOP starts processing. An idea here would be to build a
reduced structure tree per page-sequence and attach it as such to the area tree
instead of passing it to PDFDocumentHandler through the user agent. Anyway,
when accessibility is disabled (the default), FOP does not suffer any
performance drawbacks. So, IMO, the development branch could be merged with the
Trunk. At any rate, the end-user documentation is now also in place. I've added
a hopefully comprehensive list of limitations in addition to what Jost has
written up.

--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705


Andreas L. Delmelle <adelmelle@...> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |vinayak.patil1@...




--- Comment #18 from Andreas L. Delmelle <adelmelle@...>  2009-06-05 12:02:37 PST ---
*** Bug 47130 has been marked as a duplicate of this bug. ***

--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705





--- Comment #19 from Vincent Hennebert <vhennebert@...>  2009-06-26 03:48:13 PST ---
I've been (slowly) reviewing the patch and I noticed that classes in the
layoutmgr package are affected by the changes. There's no reason why it should
be the case, layout has little to do with accessibility.
In fact the layout managers just pass the ptr trait over to the area tree. This
calls for a more generic way of doing this. Indeed there are other properties
that don't affect layout but only areas (color, z-index among others).

I'll try to see how this can be handled. Just thought I'd made a note for
later.

Vincent

--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705

--- Comment #20 from Vincent Hennebert <vhennebert@...> 2009-09-17 09:07:20 PDT ---
Note: I've been asked to look at integrating this patch into Trunk, but also to
implement tagged PDF in the legacy PDFRenderer. More later.

Vincent

--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705

--- Comment #21 from Vincent Hennebert <vhennebert@...> 2009-09-22 04:59:24 PDT ---
I've just noticed that the structure tree stored in the IF XML is not
re-parsed. Running FOP with the IF as input (-afin) and the -a switch even
leads to a NPE. So tagged PDF is currently not supported when processing the FO
document in two steps, first to the intermediate format then to PDF.

--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705

--- Comment #22 from Vincent Hennebert <vhennebert@...> 2009-10-01 04:02:45 PDT ---
(In reply to comment #21)
> I've just noticed that the structure tree stored in the IF XML is not
> re-parsed. Running FOP with the IF as input (-afin) and the -a switch even
> leads to a NPE. So tagged PDF is currently not supported when processing the FO
> document in two steps, first to the intermediate format then to PDF.

This has now been addressed:
http://svn.apache.org/viewvc?rev=819585&view=rev

--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705

--- Comment #23 from Vincent Hennebert <vhennebert@...> 2009-10-01 04:05:43 PDT ---
The number tree corresponding to the ParentTree entry in the StructTreeRoot
object is invalid, as the values are directly stored in the array instead of
being reference. See discussion on fop-dev:
http://markmail.org/thread/oijggvxzfgjyewwk

--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705

--- Comment #24 from Vincent Hennebert <vhennebert@...> 2009-10-01 04:10:58 PDT ---
Are images supposed to work? I've tested a simple document that contains an
fo:external-graphic referring to a PNG image. The alt-text isn't read aloud by
Acrobat, and the image doesn't appear in the list of tags, nor the Order tab.

--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705

--- Comment #25 from benoit.wiart@... 2009-10-01 08:19:20 PDT ---
table-and-caption and table-caption are not supported by FOP but they do not
crash the engine
On the Accessibility branch, a file with a table-and-caption / table-caption
will generate a stacktrace (NPE) because foi:ptr is missing on those elements.

--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705

benoit.wiart@... changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |benoit.wiart@...

--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705

--- Comment #26 from benoit.wiart@... 2009-10-01 09:03:41 PDT ---
the NPE is at the line
structElemType.put(ptr, structElem.get("S").toString());

structElem.get("S") returns null

it seems to be caused by FOToPDFRoleMap.mapFormattingObject(s, parent) :
the STANDARD_MAPPINGS map does not know table-and-caption and table-caption

--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705

--- Comment #27 from Jeremias Maerki <jeremias@...> 2009-10-01 09:13:20 PDT ---
(In reply to comment #24)
> Are images supposed to work? I've tested a simple document that contains an
> fo:external-graphic referring to a PNG image. The alt-text isn't read aloud by
> Acrobat, and the image doesn't appear in the list of tags, nor the Order tab.

Yes, they are supposed to work. One of my accessibility test cases I've just
run through cause the alt-text for the images to be read out loud. I'm running
off a current and clean checkout. If you want to send me some PDF for analysis,
I'm happy to take a look.

--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705

--- Comment #28 from Vincent Hennebert <vhennebert@...> 2009-10-01 09:34:41 PDT ---
(In reply to comment #27)
> (In reply to comment #24)
> > Are images supposed to work? I've tested a simple document that contains an
> > fo:external-graphic referring to a PNG image. The alt-text isn't read aloud by
> > Acrobat, and the image doesn't appear in the list of tags, nor the Order tab.
>
> Yes, they are supposed to work. One of my accessibility test cases I've just
> run through cause the alt-text for the images to be read out loud. I'm running
> off a current and clean checkout. If you want to send me some PDF for analysis,
> I'm happy to take a look.

Must be my local changes then. My local copy is getting a bit messy. I'll have
another look with a fresh checkout.

But why are images handled like text, using marked-content sequences? IIUC they
should be handled as PDF Objects (see "PDF Object as Content Items" in section
9.6.3 of the PDF Reference, Third Edition).

Thanks,
Vincent

--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705

--- Comment #29 from Jeremias Maerki <jeremias@...> 2009-10-02 00:49:27 PDT ---
(In reply to comment #28)
> But why are images handled like text, using marked-content sequences? IIUC they
> should be handled as PDF Objects (see "PDF Object as Content Items" in section
> 9.6.3 of the PDF Reference, Third Edition).

Ideally maybe, but not necessarily so IMO. See the note on page 599 of PDF 1.4:
"If it is important to distinguish between multiple renditions of the same
XObject on the same page, they should be accessed via marked content sequences
enclosing particular invocations of the Do operator, rather than via object
references." So the pattern is not illegal. FOP can reuse the same XObject
multiple times on the same or on different pages.

--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705

--- Comment #30 from Vincent Hennebert <vhennebert@...> 2009-10-09 02:29:20 PDT ---
Jost or Jeremias,

In the PDFPainter.prepareImageMCID method, there is a "fix for Acro Checker"
comment. Do you have a sample FO file triggering that bug in Acrobat? I haven't
been able to reproduce it.

Thanks,
Vincent

--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

https://issues.apache.org/bugzilla/show_bug.cgi?id=46705

--- Comment #31 from Vincent Hennebert <vhennebert@...> 2009-10-13 10:15:35 UTC ---
Hi Benoît,

(In reply to comment #25)
> table-and-caption and table-caption are not supported by FOP but they do not
> crash the engine
> On the Accessibility branch, a file with a table-and-caption / table-caption
> will generate a stacktrace (NPE) because foi:ptr is missing on those elements.

This has been fixed in revision 824845:
http://svn.apache.org/viewvc?rev=824845&view=rev

Thanks for reporting the bug,
Vincent

--
Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.