As a maintainer at SBo, a thing that frequently makes me want to pull my hair off is providing a proper download link for sources hosted at Github. I really should not list all the issues that I have had, because I will get angry. So, straight to the point.
I am preparing SlackBuilds for the Bio++ Suite and the libraries it depends on. As it says on their web-page, Bio++ is a set of C++ libraries for bioinformatics, including sequence analysis, phylogenetics, molecular evolution and population genetics. The project has been hosted at Github for some time and I want my scripts to build the suite from there, since the latest release was in 2014.
Let’s take as an example the Bio++ Core Library, which is the first library that should be built. There are these components of the archive download link:
- BioPP is the project name
- bpp-core is the name of the archive I want to download
- The first 7 symbols of the commit, which at the moment are 669fa51
- Clicking there leads to 669fa51fe288a66b9857ba192d37520aae3f4dd9, which is the commit itself
Therefore, in the
bpp-core.info file, I am going to use this:
Whether the source is downloaded by wget or from within the web browser, it will be named bpp-core–669fa51fe288a66b9857ba192d37520aae3f4dd9.tar.gz (not 669fa51fe288a66b9857ba192d37520aae3f4dd9.tar.gz or something more nuts).
Pfew. Many thanks to David Spencer and the SlackBuilds core team for fixing some of my scripts, pointing to that solution.