7.4.3 fails too with python tar<br><br><div class="gmail_quote">On 10 November 2011 09:38, Andre Heinecke <span dir="ltr"><<a href="mailto:aheinecke@intevation.de">aheinecke@intevation.de</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
Hi,<br>
<div class="im"><br>
On Tuesday 18 October <a href="tel:2011%2017" value="+49201117">2011 17</a>:30:49 Ralf Habacker wrote:<br>
> Am 18.10.2011 16:05, schrieb Patrick von Reth:<br>
> > Hi<br>
> ><br>
> > If I try to unpack kdelibs 4.7.2 on emerge branch kde-4.7, python<br>
> > untar only extracts a nepumuk direcetory.<br>
> > But the tar is complete, with 7z I can extract it without problems.<br>
> > Does it work for you?<br>
><br>
> no, same issue here<br>
><br>
> did it work in the past ?<br>
><br>
> > As a quick fix I patched my utils.py to untar with 7z, I know its not<br>
> > a solution but its working ;)<br>
> ><br>
> > From some tests with EMERGE_MEASURE_TIME=True in the past I know that<br>
> > unpacking tars with 7z is nearly as fast as with python untar.<br>
><br>
> so 7z should be added as recommended package for emerge ?<br>
</div>We had sevenzip as default since Oct 30 now, this lead to strange errors with<br>
openssl, after investigation this turns out to be caused by the symbolic link<br>
handling of 7z on Windows, some symbolic links were extracted as copys, some<br>
were corrupted. A google search on "sevenzip windows symlinks" turns up<br>
several bug reports about this, even some mentioning openssl as an example.<br>
I've reverted untaring with 7zip with commit<br>
<a href="http://commits.kde.org/emerge/22b9634cda6d526140487f4c7cf0d4ace1d31be8" target="_blank">http://commits.kde.org/emerge/22b9634cda6d526140487f4c7cf0d4ace1d31be8</a><br>
<br>
Python extracts symlinks as a straightforward copy, this is defined behavior<br>
and it works on all windows platforms.<br>
I would rather have the untar fail early and loudly as it did with kdelibs<br>
then to have it "succeed" and corrupting my data :)<br>
<br>
Now with regards to the kdelibs tarball, i think we should check if that is<br>
really a valid tarfile and if so report a bug to python and if not get in<br>
contact with packagers of kdelibs. Since we had this problem only once with<br>
all the hundreds (maybe thousands) different tarballs that were unTared by<br>
emerge (and are untared by python in general) i am inclined to think that<br>
something was wrong with that tarball.<br>
<br>
Regards,<br>
Andre<br>
<span class="HOEnZb"><font color="#888888"><br>
<br>
--<br>
Andre Heinecke | <a href="tel:%2B%2B49-541-335083-262" value="+49541335083262">++49-541-335083-262</a> | <a href="http://www.intevation.de/" target="_blank">http://www.intevation.de/</a><br>
Intevation GmbH, Neuer Graben 17, 49074 Osnabrück | AG Osnabrück, HR B 18998<br>
Geschäftsführer: Frank Koormann, Bernhard Reiter, Dr. Jan-Oliver Wagner<br>
</font></span><div class="HOEnZb"><div class="h5">_______________________________________________<br>
Kde-windows mailing list<br>
<a href="mailto:Kde-windows@kde.org">Kde-windows@kde.org</a><br>
<a href="https://mail.kde.org/mailman/listinfo/kde-windows" target="_blank">https://mail.kde.org/mailman/listinfo/kde-windows</a><br>
</div></div></blockquote></div><br>