You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@pdfbox.apache.org by "John Hewson (JIRA)" <ji...@apache.org> on 2014/10/11 03:07:34 UTC

[jira] [Comment Edited] (PDFBOX-1396) The Length of stream should be written directly, not by reference

    [ https://issues.apache.org/jira/browse/PDFBOX-1396?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14167897#comment-14167897 ] 

John Hewson edited comment on PDFBOX-1396 at 10/11/14 1:07 AM:
---------------------------------------------------------------

> Currently parser gets that information after stream was read

No, it doesn't because PDFs are not sequential files, they are random-access. Some parsers (such as the old PDFBox parser) try to read the file sequentially - but this is their problem.


was (Author: jahewson):
No, it doesn't because PDFs are not sequential files, they are random-access.

> The Length of stream should be written directly, not by reference
> -----------------------------------------------------------------
>
>                 Key: PDFBOX-1396
>                 URL: https://issues.apache.org/jira/browse/PDFBOX-1396
>             Project: PDFBox
>          Issue Type: Improvement
>          Components: Writing
>    Affects Versions: 1.7.1
>            Reporter: Juraj Lonc
>            Priority: Minor
>         Attachments: PDFBOX-1396_COSStream.patch, PDFBOX-1396_COSWriter.patch
>
>
> The Length value of stream is written by reference:
> "2 0 obj
> <<
> /Length 8 0 R
> /Subtype /XML
> /Type /Metadata
> >>
> stream"
> but it is always better to tell parser the length of stream BEFORE the stream comes to parser. Currently parser gets that information after stream was read, which is sort of useless...
> Proper structure should look like this:
> "2 0 obj
> <<
> /Length 3576
> /Subtype /XML
> /Type /Metadata
> >>
> stream"



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)