[jira] [Updated] (PDFBOX-1155) setSuppressDuplicateOverlappingText sometimes removes characters that it shouldn't

["Michael McCandless (Updated) (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/PDFBOX-1155?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Michael McCandless updated PDFBOX-1155:
---------------------------------------

    Attachment: dedup.diffs.txt
                000527.pdf

Example PDF document showing the problem.  I suspect (not confirmed)
what's happening is the dedup logic sees the same character in the
line just above it and incorrectly removes it; if so, maybe the y
tolerance just needs to be tightened up?

Here are some example diffs (I also attached the full diffs); first
line is with setSuppressDuplicateOverlappingText disabled and second
line is with it enabled:

{noformat}
253c253
< Linux routine can safely be called from real-time tasks. However, some communication 
---
> Linux routine can safely be called from real-ti e t sks. However, some communication 
326c326
< used chained jumps instead of subroutine calls because the latter would not fully emulate 
---
> used chained jumps instead of subroutine calls because the latter would not fully em ate 
334c334
< level: interrupt emulation and real-time task construction, scheduling, timing and 
---
> level: interrupt emulation and real-tim  task construction, scheduling, timing and 
337c337
< scheduling helps us to run all kinds of real-time tasks in the right schedule and thus meet 
---
> scheduling helps us to run all kinds of real-time t sks in the right schedule and thus meet 
{noformat}

                
> setSuppressDuplicateOverlappingText sometimes removes characters that it shouldn't
> ----------------------------------------------------------------------------------
>
>                 Key: PDFBOX-1155
>                 URL: https://issues.apache.org/jira/browse/PDFBOX-1155
>             Project: PDFBox
>          Issue Type: Bug
>          Components: Text extraction
>            Reporter: Michael McCandless
>            Priority: Minor
>         Attachments: 000527.pdf, dedup.diffs.txt
>
>
> The duplicate detection (in PDFTextStripper.java) checks whether the
> same character was placed "nearish" to where we are about to place
> another and de-dups it if so; this is to catch documents that rewind
> and overwrite in order to bold word(s).
> But in some cases I see it removing valid characters (that were not
> dups).

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira