On 2018-05-22, Peter J. Holzer <hjp-pyt...@hjp.at> wrote: > I didn't read on Gmane. I read on my usenet server. But the broken > messages were all coming from Gmane. It is possible that the breakage > only occurs when Gmane passes the message to other Usenet servers, > although I have no idea how that could happen (frankly, I have no idea > why Gmane should replace message-ids at all - it just doesn't make > sense).
I never figured out exactly what the broken scenarios were nor did I try to figure out which gateway was causing them. Ignoring Google Groups, there are 9 possible combinations: Usenet <---[gateway]---> M-List <---[gateway]---> Gmane 1. Usenet followup to M-List posting 2. Usenet followup to Gmane posting 3. Usenet followup to Usenet posting 4. M-List followup to Usenet posting 5. M-List followup to Gmane posting 6. M-List followup to M-List posting 7. Gmane followup to Usenet posting 8. Gmane followup to M-List posting 9. Gmane followup to Gmane posting Most of the combinations seem to work most of the time. It looked like there was at least 1 broken scenario when subscribed either via Gmane or via "real" Usenet, but it's pretty difficult to glean the the signal from the noise created by people with broken MUAs and/or NNTP clients. It's actually pretty impressive it all works as well as it does... In any case, ignoring all postings from Google Groups is recommended. -- Grant Edwards grant.b.edwards Yow! Today, THREE WINOS at from DETROIT sold me a gmail.com framed photo of TAB HUNTER before his MAKEOVER! -- https://mail.python.org/mailman/listinfo/python-list