<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On 16 July 2013 11:37, Andreas Beckmann <span dir="ltr"><<a href="mailto:anbe@debian.org" target="_blank">anbe@debian.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Yes:<br>
<br>
dh_shlibdeps -a<br>
dpkg-shlibdeps: error: couldn't find library libcuda.so.1 needed by<br>
debian/nvidia-profiler/usr/bin/nvprof (ELF format: 'elf64-x86-64';<br>
RPATH: '')<br></blockquote><div><br></div><div>Thanks. I wonder if we can override this. My thinking is there's no compiling or linking happening when nvidia-cuda-toolkit is built, files are just unpacked and put in the right places, unless I missed something.<br>
</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="im"><br>
> i.e. does the build system really need an Nvidia driver installed?<br>
<br>
</div>Unless libcuda1 is split out to a separate package :-)<br></blockquote><div><br></div><div>I thought about this for Ubuntu, and in the end I don't think we'd gain much. Because we have multiple nvidia driver versions per distribution, we'd still end up with libcuda-304, libcuda-319, etc. and still need to employ all the .shlibs tricks to work around that.<br>
</div><div><br></div></div></div></div>