search

ntamas / plfit

Fitting power-law distributions to empirical data, according to the method of Clauset, Shalizi and Newman
GNU General Public License v2.0
48 stars 17 forks source link

enhance cmake material: project entry #29

Closed jgmbenoit closed 3 years ago

jgmbenoit commented 3 years ago

Description: enhance cmake material: project entry Add basic entry fields in cmake project entry in order to allow a better integration of the software. Meant to be submitted to the upstream maintainer. Origin: debian Comment: better integration Author: Jerome Benoit calculus@rezozer.net Last-Update: 2021-07-16

ntamas commented 3 years ago

Thanks! I'm not sure if I like the idea of parsing the version number from plfit.h; I'd rather do it now the way we do it in igraph, i.e. that the version number is declared in CMakeLists.txt explicitly, and we make use of CMake's PACKAGE_VERSION_MAJOR, PACKAGE_VERSION_MINOR, PACKAGE_VERSION_PATCH variables with substitutions. I'll check the other PRs and then get back to this.

jgmbenoit commented 3 years ago

It makes sense to harmonize plfit and igraph.

ntamas commented 3 years ago

See 9400867 for an implementation that generates plfit_version.h from the CMake PROJECT directive instead of the other way round.

jgmbenoit commented 2 years ago

Can you commit a new version so that the changes material get official ?

ntamas commented 2 years ago

Sure, I just need to figure out why the CI tests are failing on Windows first.

jgmbenoit commented 2 years ago

The Debian package has finally reached Debian experimental (to wit, it passed the Copyright test): it builds well except on some exotic architectures. A closer look on the s309x arch lets me think that it is a built issue (cmake or gcc). On this arc, tow shared tests fail, while all the static tests succeed. You can find the details here.

ntamas commented 2 years ago

0.9.0 is now tagged and released. I also managed to fix the Windows bug, which actually turned out to be a genuine one (but affects Windows only; uint32_t was accidentally typedef'd to __int32 instead of unsigned __int32).

jgmbenoit commented 2 years ago

Thanks. On my way to package it.