"You have found a bug or have suggestions for improvement? Let us know!" points to the bugzilla. So I write it here: On http://www.portagefilelist.de/ there is a bad upload: /home/student/01/b01902062/gentoo/usr/share/texmf-dist/tex/latex/notoccite/notoccite.sty Perhaps it makes sense to grep for all uploads containing "/home/student/" and remove them.
I don't think this is a bad entry per se. It is just were tex-live is installed on this system. Probably under a prefix. There is not only /home/student/01/b01902062/gentoo but also /extra/prefix for dev-texlive/texlive-latex-2017. @Daniel: Just noticed you were not on CC. What do you think?
Oh, must have overseen the last notification e-Mail. Anyway: I know there are more bad uploads. I already only upload packages which are located in the base gentoo repository. Is the prefix available within the package meta data after emerge? May I can look into it and filter out every package which has a different prefix than "/"? regards Daniel
I don't know how to reliable filter this out. I am not using Gentoo prefix. Maybe the value of EPREFIX is stored in the zipped environment file. @prefix team: I suspect the files uploaded to http://www.portagefilelist.de/ come from a prefix installation. Is it possible to detect this from the ebuild metadata?
I have no idea how this works, but the VDB (thus also the binpkg) contain EPREFIX property.
@Daniel: How can these entries enter the database of portagefilelist? I guess somebody must be running pfl from within a portage prefix? So if do not want these, isn't it enough to reject these entries if EPREFIX is defined (portage.const.EPREFIX)?
Created attachment 527306 [details] PFL collect script / python 3 / only ROOT-tree Hi, I changed the way that the collect script is using the portage API. It should now only collect those packages which where build with ROOT=/. I tested this by emerging a single package with ROOT=/test prefix (e.g. ROOT=/test emerge -vta htop). The emerged package was not present in the collection. Please be aware that this script is also python3 enabled. See https://bugs.gentoo.org/626116 for more information. regards Daniel
Fixed in Git with pfl-3.0. Thanks for the report!