trying to resurrect discussion about "Cannot signal a process over a channel (rfc 4254, section 6.9)"

[Thierry Parmentelat]

Hi everybody

I’d like to resurrect the discussion on a known issue with the openssh server, regarding signal delivery as described in rfc4254

This has been originally reported about ten years ago in this thread:
https://bugzilla.mindrot.org/show_bug.cgi?id=1424

I am taking he liberty to copy a few people who contributed that thread over time

The discussion does not seem to expose the reasons that lead to the feature being held back for so long
I do share Thomas’s feeling that separating the server and client aspects would be a good way to restart from a clean slate

This is why I would suggest to first address the server side of the matter, which turns out to be our team’s major and single concern here, like I believe all/most of the other requestors

The thread contains, in Comment 44, a link to a patch here - again limited to server side
https://bugzilla.mindrot.org/attachment.cgi?id=3120&action=edit
that we’ve been able to test on a fedora28 box, and that updated version indeed improves the overall behaviour of the openssh server side for our needs, when used against an python/asyncssh client side

Does it make sense for us to submit a PR on the github repo https://github.com/openssh/openssh-portable, or elsewhere ?
or were there deeper concerns about that change that need to be further discussed ?


Many thanks in anticipation — Thierry

PS
I struggled a bit to get this through, so my apologies to people who received multiple copies :)