Conversation

Christian Brauner 🦊🐺

One thing that I end up doing is generating mails/pull requests with fixed message IDs.
This allows me to refer to mails via lore links before they are sent. For example, to mark PR dependencies.

Problem is that they could be grabbed by someone else if I don't send them all at the same time.

Maybe lore could allow reserving message IDs for 24h or sm? @monsieuricon

2
2
0

@brauner
@monsieuricon Maybe it could instead allocate you a namespace that you alone are allowed to use? That way you can use whatever ID you want within that namespace and know it's safe.

1
0
0

@brenns10 @monsieuricon oh yeah, good point. Like <bliblablub@brauner> or sm

2
0
0

@brauner @monsieuricon yeah, you just couldn't let people pick their own namespace, or I guess it would have to be a manual, moderated process, so nobody ends up denying service to a particular common message-id namespace...

0
0
0

@brenns10 @monsieuricon the advantage being that I'm already doing:
<DATE-1stword-2ndword-RANDOM@brauner>

1
0
0
@brauner @monsieuricon the biggest problem is that lore is just an aggregator and has read-only access to some of its sources. As such, it has no ability to prevent collisions. This is the main reason behind my push for signing the patches and using the keyring info for authenticity validation.
1
0
0

@monsieuricon@social.kernel.org @monsieuricon@people.kernel.org honestly I only care about the ability to link between patches in advance kinda

0
0
0