Hi everyone,
Long time lurker, don't think I've ever posted, but I do have a lot of
github experience. Is the expectation that everyone who cares about the
github issues/etc also have github accounts? Would it be too much to ask
for those people to "Watch" the cf-conventions repository? They would
then get issue and PR notifications (and all comments I think).
Dave
On 11/7/18 12:33 PM, Painter, Jeff wrote:
I’ll tell you why things are set up this way and what we can do about it.
For many years, proposed changes to the CF Conventions were described
and discussed in a Trac issue tracker hosted at LLNL (Lawrence Livermore
National Laboratory in California). This traffic was rather light, and
all was copied to a mailing list, cf-metadata, hosted at LLNL. The
mailing list membership is identical to another cf-metadata list hosted
at NCAR. The reason for having two mailing lists was that if anyone
didn’t want the Trac messages, he could ask to be deleted from
cf-metadata at LLNL, and still get the other cf-metadata messages. In
practice, that never happened.
More recently, a consensus developed in the community that we should
move from Trac to the Github issue tracker, and move implementation of
document changes from one person (me) to the whole community, again
through Github. We did so, in such a way as to keep the same mailing
list connections as before. Some people knew at the time that there was
a danger that these changes would increase traffic. Not only are
proposed changes discussed as before, but with the new system we get
implementation messages (such as “merged into master”).
So now that high-traffic issue has cropped up!
What can we do? The simplest is to stop the transmission from Github to
the mailing list. If only a few people are bothered, I can take them
off the LLNL list. I can make either of those changes in a few minutes,
once we have a community consensus about what to do. More complicated
would be to send some messages but not others to the mailing list. I
might be able to do that, depending what criteria we need for choosing
messages.
- Jeff Painter
*From: *<[email protected]> on behalf of Chris Barker
<[email protected]>
*Date: *Wednesday, November 7, 2018 at 9:26 AM
*To: *Janine Aquino <[email protected]>
*Cc: *Chris Barker - NOAA Federal
<[email protected]>, Martin Juckes - UKRI
STFC <[email protected]>, Erik Quaeghebeur
<[email protected]>, Jim Biard <[email protected]>,
"[email protected]" <[email protected]>
*Subject: *Re: Please stop sending Github messages to the ML
On Wed, Nov 7, 2018 at 8:50 AM, Janine Aquino <[email protected]
<mailto:[email protected]>> wrote:
I am not convinced that is the source of the problem. -- I have
never personally commented on this issue.
no one thinks you did -- but somehow "cf-metadata-list" did comment on
the issue. And there does appear to be such a user:
https://github.com/cf-metadata-list
(Joined gitHub Oct 7, 2017)
I suspect that the cf-metadata-list user was created so that we could
have gitHub issues mirrored to the list.
I have been receiving GitHub messages via the list for months
This particular issue has been very active, but it is NOT months old.
And the gitHub account has not been all that active before that. So I
think that the list is purposely notified when new issues are created, etc.
So I _think_ that what happened was:
The CF gitHub account is set up to email the list (via the gitHub
cf-metadata-list user) in certain circumstances -- I hope that's issue
creation (and maybe PR creation), but not all comments.
But the way the gitHub email linking works, when gitHub sends an email
about an issue (usually to a regular gitHub user), when they respond to
that email, their response is posted as a comment on that issue.
This is a nice feature, as it allows users to participate in discussion
completely through email.
But it also means that if someone responds to an email that was sent to
the list by gitHub, then gitHub thinks the list itself responded, so it
posts the mesage as a comment (good), but then that user is now added to
the "wants to be notified of new comments" list, and presto -- a lo tof
discussion spamming the list.
To solve the problem at hand, someone needs to log in to gitHub as the
cf-metadata-list user and take themselves off the I want to follow this
issue list. _maybe_ the issue creator (Jim in this case) can do that,
but I suspect not.
So who knows the password of the cf-metadata-list user ??
In the future, I have no idea how to maintain a bridge to the list
without risking this happening again. But maybe there is a way to set
that user up to receive emails, but not be able to post via email -- I
have no idea.
I am currently trying to unsubscribe from the cf-metadata list, but
so far no luck.
have you tried this page?
http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
-CHB
-----------
Emergency Response Division
NOAA/NOS/OR&R (206) 526-6959 voice
7600 Sand Point Way NE (206) 526-6329 fax
Seattle, WA 98115 (206) 526-6317 main reception
[email protected] <mailto:[email protected]>
