pdfbox-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andreas Lehmkühler (JIRA) <j...@apache.org>
Subject [jira] Commented: (PDFBOX-136) sloooooooooooooooooowwwww.....
Date Thu, 02 Jul 2009 07:48:47 GMT

    [ https://issues.apache.org/jira/browse/PDFBOX-136?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12726371#action_12726371
] 

Andreas Lehmkühler commented on PDFBOX-136:
-------------------------------------------

As fas as I understand mig-o made the experience that the writing performance of pdfbox depends
on the used outputstream-class. And I guess that everyone agrees to that fact. But the suggested
solution isn't appropriate. The class COSStandardOutputStream and at least the class COSWriter
are using the abstract class java.io.OutputStream as type for the stream the data will be
written to give the user the chance to choose his favorite kind of OutputStream. instead of
changing the pdfbox code,  the user has to create the preferred kind of outputstream and has
to pass it to pdfbox.

> sloooooooooooooooooowwwww.....
> ------------------------------
>
>                 Key: PDFBOX-136
>                 URL: https://issues.apache.org/jira/browse/PDFBOX-136
>             Project: PDFBox
>          Issue Type: Bug
>          Components: Writing
>
> [imported from SourceForge]
> http://sourceforge.net/tracker/index.php?group_id=78314&atid=552832&aid=1432253
> Originally submitted by mig-o on 2006-02-15 07:53.
> And you just had to replace 
>     public COSStandardOutputStream(OutputStream out)
>     {
>         super( out );
>     }
> by 
>     public COSStandardOutputStream(OutputStream out)
>     {
>         super( new BufferedOutputStream(out,16384) );
>     }
> . Here are some timings for my local drive (i do some
> other stuff for the first 4 seconds...).
> old way: 5103 ms
> new way: 4403 ms 
> On my remote drive:
> old way: 136792 ms
> new way:   4756 ms
> The reason should be clear... Network-packages for
> nearly each and every written byte in the pdf.
> [comment on SourceForge]
> Originally sent by nobody.
> Logged In: NO 
> @previous post: anyone=everyone
> [comment on SourceForge]
> Originally sent by nobody.
> Logged In: NO 
> The bug is, that only the old way exists in the repository,
> so i wrote this, to have someone change this in cvs. 
> Or does anyone has cvs write access here? Then the bug is my
> stupidity, since i could have it done myself :)
> [comment on SourceForge]
> Originally sent by benlitchfield.
> Logged In: YES 
> user_id=601708
> I don't think I understand what the bug is?  If old was was 
> 136 seconds and new way is 4 seconds then what is the 
> problem?

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message