Re: [Debian-science-sagemath] Sagemath Build-Depends backported - unfortunately build error (Was: Help needed to backport ipywidgets ... (Was: Backporting sagemath))
Hi Ximin,
On Sun, Apr 16, 2017 at 12:57:00PM +0000, Ximin Luo wrote:
> [..]
> libnettle4{a} libnspr4{a} libnss3{a} libntl-dev{a} libntl27{a} libntl5{a}
> [..]
> [dochtml] /usr/lib/libntl.so.27(_ZNK3NTL11ZZ_pContext7restoreEv+0x33)[0x7f20158baf83]
> [dochtml] /usr/lib/libntl.so.5(_ZN3NTL4ZZ_p4initERKNS_2ZZE+0x2c)[0x7f1dc3b1b6cc]
>
> Somehow sage is still linked to both ntl 27 and 5.
I have no idea why the packages from backports (new packages are
remaining in backports-policy) are prefered over those in my local
repository sometimes and sometimes not - I managed to solve this by
versioned Build-Depends now (hopefully).
> Perhaps it would be good to add a Build-Conflicts: field then the build would fail earlier.
>
> Could you also take a look at the log yourself first, before emailing this list?
Fair enough. Sorry for beeing too lazy here - I somehow seemed to be
blind here. I now successfully built once when I tried to build for
final upload. Unfortunately the build does not seem to be reliable
since I was running into another failure. Since the log is quite a
large file now I have uploaded it here
https://people.debian.org/~tille/tmp/sagemath/sagemath_7.4-9~bpo8+1_amd64.build.gz
Kind regards and thanks again for your patience
Andreas.
--
http://fam-tille.de
Reply to: