On Mon, Oct 05, 2020 at 07:53:05AM +0100, Julian Gilbey wrote: > On Mon, Oct 05, 2020 at 12:57:59AM +0100, Julian Gilbey wrote: > > 4) The output of "greylist list" is now somewhat mangled: the "Data" > > field shows something like: > > [ ' 1 . 2 . 3 . 4 / 2 4 ' , ' i n f o @ e x a m p l e . c o m ' , ' j d > > g @ d e b i a n . o r g ' ] > > rather than > > 1.2.3.4/24 i...@example.com j...@debian.org > > I guess this should be easy to fix? > > Oh, it turns out that this is not so straightforward: the command > greylist add 1.2.3.4/24 i...@example.com j...@debian.org > yields the following line in /var/lib/greylistd/triplets, which will > probably break things: > 4921632547390726349 = 1.2.3.4/24 i...@example.com j...@debian.org > > So it seems the triplets need turning into simple strings before they > are stored in the triplets file / triplets hash, rather than being > stored as Python literals with lots of excess spaces.
Hi Benedikt and Thorsten, I've found the cause of this and made a pull request on your GitHub repository. If you're happy with it or a week or so passes and I don't hear from you or the maintainer (Thorsten), I'll upload an NMU to the delayed queue. Thorsten: there are many more changes here than would usually be permitted in an NMU. But the package is currently completely broken and won't get into testing. This set of patches fixes the critical issues and several other lintian warnings (as well as Benedikt's other cleaning-up). The big remaining lintian warning is that there is no systemd service file. This will take some effort to do, as systemd will happily look after the socket creation, but then the Python code will need adapting to accept the systemd socket rather than creating one itself. I've never done this and don't really have time right now to do so. Best wishes, Julian
