<div>The following PDF reproduces the bug: <a href="http://midnight-labs.org/testextract.pdf">http://midnight-labs.org/testextract.pdf</a> </div><div><br></div><div>I tried to get the frontpage from a few other PDFs, but when I extract with `pdftk` poppler decides to like the new PDF that's produced.</div>
<div><br></div><div>On 6 August 2010 14:52, Albert Astals Cid <span dir="ltr"><<a href="mailto:aacid@kde.org">aacid@kde.org</a>></span> wrote:</div><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
A Dijous, 5 d'agost de 2010, Harry Roberts va escriure:<br>
<div><div></div><div class="h5">> Just to give you all a heads up I've had to revert (in my local branch) the<br>
> code which used getRawChars because it's breaking a huge number of PDFs.<br>
><br>
> See Albert's e-mail from 14 Jun:<br>
> <a href="http://www.mail-archive.com/poppler@lists.freedesktop.org/msg03784.html" target="_blank">http://www.mail-archive.com/poppler@lists.freedesktop.org/msg03784.html</a><br>
><br>
> Has anybody else experienced this? I normally pull changes in from the main<br>
> git repo every few weeks when I do test builds but only got round to<br>
> debugging it today.<br>
><br>
> e.g. the error:<br>
><br>
> Internal: called getRawChars() on non-predictor stream<br>
><br>
> It looks quite trivial to implement `getRawChars` on regular `Streams`, but<br>
> I've not tried it locally yet.<br>
<br>
</div></div>Make sure you rebuild your programs using libpoppler as my change breaks<br>
binary compatibility and if you don't rebuild your programs the incorrect<br>
symbols will be used resulting in weird behaviour.<br>
<br>
If that fails provide a pdf that gets broken.<br>
</blockquote></div><br>