[Debian-science-sagemath] plans to package primecount (the C++ library)

[Dima Pasechnik]

Hi Jerome,

it looks as if on debian 11 there is a bug in the cmake-related
configuration of libprimesieve-dev.
Namely, it does not provide cmake config files.
Other similar packages, such as qhull and boost do this: one can see
there is nothing there:

$ ls /usr/lib/x86_64-linux-gnu/cmake/
Boost-1.74.0  BoostDetectToolset-1.74.0.cmake  boost_headers-1.74.0  norm  Qhull

As a result, packages using cmake discovery tools (find_package, etc)
are unable to locate it, e.g. primecount:
(as observed while working on https://trac.sagemath.org/ticket/25009)

[primecount-7.1] Configuring primecount.
[primecount-7.1] Configuring primecount-7.1 with cmake
[primecount-7.1] -- The CXX compiler identification is GNU 10.2.1
[...]
[primecount-7.1] -- Performing Test mpopcnt - Success
[primecount-7.1] CMake Error at CMakeLists.txt:258 (find_package):
[primecount-7.1]   By not providing "Findprimesieve.cmake" in
CMAKE_MODULE_PATH this project
[primecount-7.1]   has asked CMake to find a package configuration
file provided by
[primecount-7.1]   "primesieve", but CMake did not find one.
[primecount-7.1]
[primecount-7.1]   Could not find a package configuration file
provided by "primesieve"
[primecount-7.1]   (requested version 7.6) with any of the following names:
[primecount-7.1]
[primecount-7.1]     primesieveConfig.cmake
[primecount-7.1]     primesieve-config.cmake
[primecount-7.1]
[primecount-7.1]   Add the installation prefix of "primesieve" to
CMAKE_PREFIX_PATH or set
[primecount-7.1]   "primesieve_DIR" to a directory containing one of
the above files.  If
[primecount-7.1]   "primesieve" provides a separate development
package or SDK, be sure it has
[primecount-7.1]   been installed.
[primecount-7.1]
[primecount-7.1]
[primecount-7.1] -- Configuring incomplete, errors occurred!

I think one needs to configure primesieve with
-DCMAKE_INSTALL_PREFIX=XXX
where XXX is the "general"  installation prefix.

---
Thanks
Dima

On Fri, Nov 19, 2021 at 8:50 PM Jerome BENOIT <calculus at rezozer.net> wrote:
>
> Hi, thanks for the clarification.
> I got that upstream makes things a bit complicated here.
>
>
> On 19/11/2021 20:25, Julien Puydt wrote:
> > Le vendredi 19 novembre 2021 à 19:22 +0100, Jerome BENOIT a écrit :
> >>   I am getting confused.
> >
> > If I understood Dima's explanations:
> >
> > - upstream vendors its own libprimesieve in primecount ;
> >
> > - but you can define a switch to when compiling primecount, it doesn't
> > use the vendored one but the system one.
> >
> > I would say a d/copyright eliminating the vendored version and a small
> > patch to make the vendor-switch off by default will do the trick.
>
> I agree.
>
> >
> > I can start on it monday if you don't want to.
>
> I guess it would make a better sense if I package it.
> I am familiar with the upstream maintainer.
> And at one point, the packages might merge.
>
>
> Best wishes,
> Jerome
> >
> > Cheers,
> >
> > J.Puydt
> >
> > _______________________________________________
> > Debian-science-sagemath mailing list
> > Debian-science-sagemath at alioth-lists.debian.net
> > https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-sagemath
> >
>
> --
> Jerome BENOIT | calculus+at-rezozer^dot*net
> https://qa.debian.org/developer.php?login=calculus@rezozer.net
> AE28 AE15 710D FF1D 87E5  A762 3F92 19A6 7F36 C68B
>
> _______________________________________________
> Debian-science-sagemath mailing list
> Debian-science-sagemath at alioth-lists.debian.net
> https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-sagemath