Thread: Moderation
Emails from my confirmed email address have been held up for moderation today. Is there a new policy about all posts to lists going through moderation or is there a bug that is doing this? -- Adrian Klaver adrian.klaver@aklaver.com
Adrian Klaver <adrian.klaver@aklaver.com> writes:
> Emails from my confirmed email address have been held up for moderation
> today. Is there a new policy about all posts to lists going through
> moderation or is there a bug that is doing this?
Most likely the thread is tripping across one of the strings that
get filtered ... words like "s*bscribe" tend to do that, especially
at the start of a line.
regards, tom lane
On 8/13/21 9:04 AM, Tom Lane wrote: > Adrian Klaver <adrian.klaver@aklaver.com> writes: >> Emails from my confirmed email address have been held up for moderation >> today. Is there a new policy about all posts to lists going through >> moderation or is there a bug that is doing this? > > Most likely the thread is tripping across one of the strings that > get filtered ... words like "s*bscribe" tend to do that, especially > at the start of a line. Well my reply on --general to this: https://www.postgresql.org/message-id/AS8PR04MB77191F86C4BD4759762E1B90A7FA9%40AS8PR04MB7719.eurprd04.prod.outlook.com is also hung in the queue. The common factor is that the subjects both start with 'RE(e):'. > > regards, tom lane > -- Adrian Klaver adrian.klaver@aklaver.com
On Fri, Aug 13, 2021, 18:51 Adrian Klaver <adrian.klaver@aklaver.com> wrote:
On 8/13/21 9:04 AM, Tom Lane wrote:
> Adrian Klaver <adrian.klaver@aklaver.com> writes:
>> Emails from my confirmed email address have been held up for moderation
>> today. Is there a new policy about all posts to lists going through
>> moderation or is there a bug that is doing this?
>
> Most likely the thread is tripping across one of the strings that
> get filtered ... words like "s*bscribe" tend to do that, especially
> at the start of a line.
Well my reply on --general to this:
https://www.postgresql.org/message-id/AS8PR04MB77191F86C4BD4759762E1B90A7FA9%40AS8PR04MB7719.eurprd04.prod.outlook.com
is also hung in the queue.
The common factor is that the subjects both start with 'RE(e):'.
Nope, I think the common factor is they're cced to multiple lists. We discourage that pretty strongly and for that reason all such emails are moderated.
/Magnus
Adrian Klaver <adrian.klaver@aklaver.com> writes: > Well my reply on --general to this: > https://www.postgresql.org/message-id/AS8PR04MB77191F86C4BD4759762E1B90A7FA9%40AS8PR04MB7719.eurprd04.prod.outlook.com > is also hung in the queue. That one likely is stuck because of the not-that-new-anymore policy that cross-posting to multiple lists gets moderated. If you want to answer somebody who's done that, I'd suggest picking the most appropriate list and cc'ing only that one. regards, tom lane
On 8/13/21 9:53 AM, Magnus Hagander wrote: > Nope, I think the common factor is they're cced to multiple lists. We > discourage that pretty strongly and for that reason all such emails are > moderated. I don't suppose it is possible to include that information in the moderation message? If nothing else as a hint. Muscle memory says hit Reply All. If I got a reminder of the policy I could cancel the message and resend to a single list. > > /Magnus > -- Adrian Klaver adrian.klaver@aklaver.com
Greetings, * Adrian Klaver (adrian.klaver@aklaver.com) wrote: > On 8/13/21 9:04 AM, Tom Lane wrote: > >Adrian Klaver <adrian.klaver@aklaver.com> writes: > >>Emails from my confirmed email address have been held up for moderation > >>today. Is there a new policy about all posts to lists going through > >>moderation or is there a bug that is doing this? > > > >Most likely the thread is tripping across one of the strings that > >get filtered ... words like "s*bscribe" tend to do that, especially > >at the start of a line. > > Well my reply on --general to this: > > https://www.postgresql.org/message-id/AS8PR04MB77191F86C4BD4759762E1B90A7FA9%40AS8PR04MB7719.eurprd04.prod.outlook.com > > is also hung in the queue. > > The common factor is that the subjects both start with 'RE(e):'. The common factor is that they're cross-posted across multiple lists. We've had issues with people essentially spraying across a bunch of lists like -admin, -general and -hackers with one email and so we put in a policy to moderate cross-posted emails. There's a thread from back in January 2019 about this and subsequent discussion in May of 2020, probably the most relevant being: https://www.postgresql.org/message-id/20200522174838.GT3418%40tamriel.snowman.net I've tried to keep up with the moderation but I'd be just as happy for people to stop doing quite so much cross-posting across multiple lists. We've talked about having some kind of 'exclusion' list of posters who are allowed to cross-post but that capability hasn't been implemented in pglister at this point. Thanks, Stephen
Attachment
Greetings, * Adrian Klaver (adrian.klaver@aklaver.com) wrote: > On 8/13/21 9:53 AM, Magnus Hagander wrote: > > >Nope, I think the common factor is they're cced to multiple lists. We > >discourage that pretty strongly and for that reason all such emails are > >moderated. > > I don't suppose it is possible to include that information in the moderation > message? If nothing else as a hint. Muscle memory says hit Reply All. If I > got a reminder of the policy I could cancel the message and resend to a > single list. Isn't implemented today ... but +1 on figuring out a way to do that. If anyone wants to hack on pglister, the project lives here: https://gitlab.com/pglister/pglister Thanks, Stephen
Attachment
On 8/13/21 10:03 AM, Stephen Frost wrote: > Greetings, > > * Adrian Klaver (adrian.klaver@aklaver.com) wrote: >> On 8/13/21 9:53 AM, Magnus Hagander wrote: >> >>> Nope, I think the common factor is they're cced to multiple lists. We >>> discourage that pretty strongly and for that reason all such emails are >>> moderated. >> >> I don't suppose it is possible to include that information in the moderation >> message? If nothing else as a hint. Muscle memory says hit Reply All. If I >> got a reminder of the policy I could cancel the message and resend to a >> single list. > > Isn't implemented today ... but +1 on figuring out a way to do that. Could it not be added as: Hint: You may be getting this message if your post was cross posted. Either wait for moderator approval or cancel the post and resend to a single list. to here?: https://gitlab.com/pglister/pglister/-/blob/master/mailtemplates/moderation_response.txt > > If anyone wants to hack on pglister, the project lives here: > > https://gitlab.com/pglister/pglister > > Thanks, > > Stephen > -- Adrian Klaver adrian.klaver@aklaver.com
Greetings, * Adrian Klaver (adrian.klaver@aklaver.com) wrote: > On 8/13/21 10:03 AM, Stephen Frost wrote: > >* Adrian Klaver (adrian.klaver@aklaver.com) wrote: > >>On 8/13/21 9:53 AM, Magnus Hagander wrote: > >> > >>>Nope, I think the common factor is they're cced to multiple lists. We > >>>discourage that pretty strongly and for that reason all such emails are > >>>moderated. > >> > >>I don't suppose it is possible to include that information in the moderation > >>message? If nothing else as a hint. Muscle memory says hit Reply All. If I > >>got a reminder of the policy I could cancel the message and resend to a > >>single list. > > > >Isn't implemented today ... but +1 on figuring out a way to do that. > > Could it not be added as: > > Hint: You may be getting this message if your post was cross posted. Either > wait for moderator approval or cancel the post and resend to a single list. > > to here?: > > https://gitlab.com/pglister/pglister/-/blob/master/mailtemplates/moderation_response.txt No, I don't think just adding it to every moderation notice that is sent out would be very good. If we're going to do this then we should make it so that the notice actually says why it was moderated, if the issue was cross-posting and then have a custom notice for that. Thanks, Stephen
Attachment
On 8/13/21 10:18 AM, Stephen Frost wrote: > Greetings, > > * >> https://gitlab.com/pglister/pglister/-/blob/master/mailtemplates/moderation_response.txt > > No, I don't think just adding it to every moderation notice that is sent > out would be very good. If we're going to do this then we should make > it so that the notice actually says why it was moderated, if the issue > was cross-posting and then have a custom notice for that. I don't suppose there is a cheat sheet somewhere that gives an outline of how information is processed in pglister? > > Thanks, > > Stephen > -- Adrian Klaver adrian.klaver@aklaver.com
On 8/13/21 10:49 AM, Adrian Klaver wrote: > On 8/13/21 10:18 AM, Stephen Frost wrote: >> Greetings, >> >> * >>> https://gitlab.com/pglister/pglister/-/blob/master/mailtemplates/moderation_response.txt >>> >> >> No, I don't think just adding it to every moderation notice that is sent >> out would be very good. If we're going to do this then we should make >> it so that the notice actually says why it was moderated, if the issue >> was cross-posting and then have a custom notice for that. > > I don't suppose there is a cheat sheet somewhere that gives an outline > of how information is processed in pglister? Alright pulled out the shovel and I believe the place to intercept this is: https://gitlab.com/pglister/pglister/-/blob/master/lib/handlers/mailhandler.py MailHandler.moderate() Am I getting close? > >> >> Thanks, >> >> Stephen >> > > -- Adrian Klaver adrian.klaver@aklaver.com
Greetings, * Adrian Klaver (adrian.klaver@aklaver.com) wrote: > On 8/13/21 10:49 AM, Adrian Klaver wrote: > >On 8/13/21 10:18 AM, Stephen Frost wrote: > >>>https://gitlab.com/pglister/pglister/-/blob/master/mailtemplates/moderation_response.txt > >>> > >> > >>No, I don't think just adding it to every moderation notice that is sent > >>out would be very good. If we're going to do this then we should make > >>it so that the notice actually says why it was moderated, if the issue > >>was cross-posting and then have a custom notice for that. > > > >I don't suppose there is a cheat sheet somewhere that gives an outline of > >how information is processed in pglister? > > Alright pulled out the shovel and I believe the place to intercept this is: > > https://gitlab.com/pglister/pglister/-/blob/master/lib/handlers/mailhandler.py > > MailHandler.moderate() > > Am I getting close? Yes, offhand that looks like the right area. Of course, installing it and setting it up and testing it by making changes would be the way to go ... and writing docs along the way would be certainly welcome. :) Thanks! Stephen
Attachment
On 8/13/21 11:35 AM, Stephen Frost wrote: > Greetings, > > * Adrian Klaver (adrian.klaver@aklaver.com) wrote: >> On 8/13/21 10:49 AM, Adrian Klaver wrote: >>> On 8/13/21 10:18 AM, Stephen Frost wrote: >>>>> https://gitlab.com/pglister/pglister/-/blob/master/mailtemplates/moderation_response.txt >>>>> >>>> >>>> No, I don't think just adding it to every moderation notice that is sent >>>> out would be very good. If we're going to do this then we should make >>>> it so that the notice actually says why it was moderated, if the issue >>>> was cross-posting and then have a custom notice for that. >>> >>> I don't suppose there is a cheat sheet somewhere that gives an outline of >>> how information is processed in pglister? >> >> Alright pulled out the shovel and I believe the place to intercept this is: >> >> https://gitlab.com/pglister/pglister/-/blob/master/lib/handlers/mailhandler.py >> >> MailHandler.moderate() >> >> Am I getting close? > > Yes, offhand that looks like the right area. > > Of course, installing it and setting it up and testing it by making > changes would be the way to go ... and writing docs along the way would > be certainly welcome. :) Am I right that there is no documentation currently? If there is where is it hidden? Any pointers as to where to look for setup information e.g requirements.txt. While I'm at it is there a reason inline comments are used instead of docstrings? > > Thanks! > > Stephen > -- Adrian Klaver adrian.klaver@aklaver.com
Greetings, * Adrian Klaver (adrian.klaver@aklaver.com) wrote: > On 8/13/21 11:35 AM, Stephen Frost wrote: > >* Adrian Klaver (adrian.klaver@aklaver.com) wrote: > >>On 8/13/21 10:49 AM, Adrian Klaver wrote: > >>>On 8/13/21 10:18 AM, Stephen Frost wrote: > >>>>>https://gitlab.com/pglister/pglister/-/blob/master/mailtemplates/moderation_response.txt > >>>>> > >>>> > >>>>No, I don't think just adding it to every moderation notice that is sent > >>>>out would be very good. If we're going to do this then we should make > >>>>it so that the notice actually says why it was moderated, if the issue > >>>>was cross-posting and then have a custom notice for that. > >>> > >>>I don't suppose there is a cheat sheet somewhere that gives an outline of > >>>how information is processed in pglister? > >> > >>Alright pulled out the shovel and I believe the place to intercept this is: > >> > >>https://gitlab.com/pglister/pglister/-/blob/master/lib/handlers/mailhandler.py > >> > >>MailHandler.moderate() > >> > >>Am I getting close? > > > >Yes, offhand that looks like the right area. > > > >Of course, installing it and setting it up and testing it by making > >changes would be the way to go ... and writing docs along the way would > >be certainly welcome. :) > > Am I right that there is no documentation currently? Yeah.. I'm afraid that's mostly the case, though maybe I'm forgetting something. Magnus? > Any pointers as to where to look for setup information e.g requirements.txt. Eh, not exactly the same as what you're asking for, but the package we built for it has: Depends: pginfra-wrap-django, pginfra-wrap-eximserver, python-psycopg2, python3-psycopg2, pginfra-wrap-pgbouncer, pginfra-wrap-postgresql-11 | pginfra-wrap-postgresql-12, pginfra-wrap-exim-using-pgsql Django is 2.2.24. Not sure why we don't have a requirements.txt, heh. Would certainly like to improve on all of this, of course. > While I'm at it is there a reason inline comments are used instead of > docstrings? Would have to ask Magnus on that one. :) Thanks! Stephen
Attachment
On 8/13/21 1:51 PM, Stephen Frost wrote: > Greetings, > >> >> Am I right that there is no documentation currently? > > Yeah.. I'm afraid that's mostly the case, though maybe I'm forgetting > something. Magnus? > >> Any pointers as to where to look for setup information e.g requirements.txt. > > Eh, not exactly the same as what you're asking for, but the package we > built for it has: > > Depends: pginfra-wrap-django, pginfra-wrap-eximserver, python-psycopg2, > python3-psycopg2, pginfra-wrap-pgbouncer, > pginfra-wrap-postgresql-11 | pginfra-wrap-postgresql-12, > pginfra-wrap-exim-using-pgsql pginfra lives where? > > Django is 2.2.24. Ooh, ooh, I know that one:) > > Not sure why we don't have a requirements.txt, heh. Would certainly > like to improve on all of this, of course. Yeah, it seems before changes to code can be made there needs to be some sort of README/INSTALL that outlines the process for getting a working install to develop against. Is there a place this lives or could live? > Thanks! > > Stephen > -- Adrian Klaver adrian.klaver@aklaver.com
Greetings, * Adrian Klaver (adrian.klaver@aklaver.com) wrote: > On 8/13/21 1:51 PM, Stephen Frost wrote: > >>Am I right that there is no documentation currently? > > > >Yeah.. I'm afraid that's mostly the case, though maybe I'm forgetting > >something. Magnus? > > > >>Any pointers as to where to look for setup information e.g requirements.txt. > > > >Eh, not exactly the same as what you're asking for, but the package we > >built for it has: > > > >Depends: pginfra-wrap-django, pginfra-wrap-eximserver, python-psycopg2, > > python3-psycopg2, pginfra-wrap-pgbouncer, > > pginfra-wrap-postgresql-11 | pginfra-wrap-postgresql-12, > > pginfra-wrap-exim-using-pgsql > > pginfra lives where? Not sure what you're asking here. > >Django is 2.2.24. > > Ooh, ooh, I know that one:) Good! :) > >Not sure why we don't have a requirements.txt, heh. Would certainly > >like to improve on all of this, of course. > > Yeah, it seems before changes to code can be made there needs to be some > sort of README/INSTALL that outlines the process for getting a working > install to develop against. Is there a place this lives or could live? That'd certainly be helpful. I would think that such a document would live in the git repo.. I'm also trying to pull in some other folks who I know are playing around with pglister who might be able to help. Thanks, Stephen
Attachment
On 8/16/21 7:55 AM, Stephen Frost wrote: > Greetings, > > * Adrian Klaver (adrian.klaver@aklaver.com) wrote: >> On 8/13/21 1:51 PM, Stephen Frost wrote: >>>> Am I right that there is no documentation currently? >>> >>> Yeah.. I'm afraid that's mostly the case, though maybe I'm forgetting >>> something. Magnus? >>> >>>> Any pointers as to where to look for setup information e.g requirements.txt. >>> >>> Eh, not exactly the same as what you're asking for, but the package we >>> built for it has: >>> >>> Depends: pginfra-wrap-django, pginfra-wrap-eximserver, python-psycopg2, >>> python3-psycopg2, pginfra-wrap-pgbouncer, >>> pginfra-wrap-postgresql-11 | pginfra-wrap-postgresql-12, >>> pginfra-wrap-exim-using-pgsql >> >> pginfra lives where? > > Not sure what you're asking here. > I have yet to find a pginfra project so I am trying to figure out where the packages come from? > That'd certainly be helpful. I would think that such a document would > live in the git repo.. > > I'm also trying to pull in some other folks who I know are playing > around with pglister who might be able to help. > > Thanks, > > Stephen > -- Adrian Klaver adrian.klaver@aklaver.com
Greetings,
* Adrian Klaver (adrian.klaver@aklaver.com) wrote:
> On 8/16/21 7:55 AM, Stephen Frost wrote:
> >* Adrian Klaver (adrian.klaver@aklaver.com) wrote:
> >>On 8/13/21 1:51 PM, Stephen Frost wrote:
> >>>>Am I right that there is no documentation currently?
> >>>
> >>>Yeah.. I'm afraid that's mostly the case, though maybe I'm forgetting
> >>>something. Magnus?
> >>>
> >>>>Any pointers as to where to look for setup information e.g requirements.txt.
> >>>
> >>>Eh, not exactly the same as what you're asking for, but the package we
> >>>built for it has:
> >>>
> >>>Depends: pginfra-wrap-django, pginfra-wrap-eximserver, python-psycopg2,
> >>> python3-psycopg2, pginfra-wrap-pgbouncer,
> >>> pginfra-wrap-postgresql-11 | pginfra-wrap-postgresql-12,
> >>> pginfra-wrap-exim-using-pgsql
> >>
> >>pginfra lives where?
> >
> >Not sure what you're asking here.
>
> I have yet to find a pginfra project so I am trying to figure out where the
> packages come from?
Those are just wrappers that we build that depend on regular packages
but add a bit of configuration and such, but nothing that is likely to
be interesting to this. pgbouncer isn't required, certainly PG is but
it could be 11, 12, or probably others. We use exim for most of our
email and we have a wrapper package to make exim wait for PG to start up
since our exim config queries the PG database for things (the actual
exim config is backed up but isn't part of the package since it's
different on our pglister system from other ones).
As much as I'd like this to all be automated and easy to deploy, it
isn't today. One thing that might help is that the processes which need
to be started are:
/lib/systemd/system/pglister_bounceprocessor.service
/lib/systemd/system/pglister_cleanup.service
/lib/systemd/system/pglister_eximcleaner.service
/lib/systemd/system/pglister_mailprocessor.service
/lib/systemd/system/pglister_mailsender.service
/lib/systemd/system/pglister_moderationprocessor.service
Each of those is basically:
####
[Unit]
Description=pglister: bounce processor
After=postgresql.service
[Service]
ExecStart=/usr/local/pglister/bin/bounceprocessor.py
WorkingDirectory=/usr/local/pglister
Restart=always
RestartSec=30
User=list
Group=list
[Install]
WantedBy=multi-user.target
####
As mentioned, we use exim and so the way mail gets into the system is
through a pipe:
pglister_pipe:
driver = pipe
command = MAILINGLISTPIPESCRIPT
#ignore_status
temp_errors = *
return_output = false
freeze_exec_fail = true
log_defer_output = true
log_output = true
headers_remove = List-Unsubscribe:List-Owner:List-Post:List-Subscribe:List-Help:List-Id:List-Archive
#log_fail_output = true
umask = 022
environment =
user = Debian-exim
message_prefix =
where that command (MAILINGLISTPIPESCRIPT) is:
/usr/local/pglister/bin/inject.py -s $sender_address -d $local_part$local_part_suffix@$domain -m $header_message-id:
There's then some lookups we do from exim to check if we should be
sending things to pglister:
domainlist local_domains = @ : ${lookup pgsql{select domain from domains where domain='${quote_pgsql:$domain}'}
{$value}}
local_parts = mailer-daemon : ${lookup pgsql{select name from mailinglists WHERE name='${quote_pgsql:$local_part}' and
domain='${quote_pgsql:$domain}'}{$value}}
though that's not strictly required, just nicer to bounce things earlier
than having them end up in the pglister inbound queue where they'll sit
until an admin deals with it.
Generally, email comes in via inject.py, which will decide if it's a
bounce or an incoming email and INSERT the email into the appropriate
table and call NOTIFY to wake up either the bouncer processor or the
mail processor. Those will do some work and eventually may create
emails to go out and wake up the mailsender to process the queue of
mails to send.
And then there's a django interface to managing the lists, handling
moderation, etc.
Hopefully this helps give a bit more insight into how it works.
Thanks,
Stephen
Attachment
On 8/16/21 9:06 AM, Stephen Frost wrote: > Greetings, > >> I have yet to find a pginfra project so I am trying to figure out where the >> packages come from? > > Those are just wrappers that we build that depend on regular packages > but add a bit of configuration and such, but nothing that is likely to > be interesting to this. pgbouncer isn't required, certainly PG is but > it could be 11, 12, or probably others. We use exim for most of our > email and we have a wrapper package to make exim wait for PG to start up > since our exim config queries the PG database for things (the actual > exim config is backed up but isn't part of the package since it's > different on our pglister system from other ones). > > As much as I'd like this to all be automated and easy to deploy, it > isn't today. One thing that might help is that the processes which need > to be started are: > > /lib/systemd/system/pglister_bounceprocessor.service > /lib/systemd/system/pglister_cleanup.service > /lib/systemd/system/pglister_eximcleaner.service > /lib/systemd/system/pglister_mailprocessor.service > /lib/systemd/system/pglister_mailsender.service > /lib/systemd/system/pglister_moderationprocessor.service > Interesting, but I still don't see where these packages/code/etc can be found? > Hopefully this helps give a bit more insight into how it works. > > Thanks, > > Stephen > -- Adrian Klaver adrian.klaver@aklaver.com
Greetings, * Adrian Klaver (adrian.klaver@aklaver.com) wrote: > On 8/16/21 9:06 AM, Stephen Frost wrote: > >>I have yet to find a pginfra project so I am trying to figure out where the > >>packages come from? > > > >Those are just wrappers that we build that depend on regular packages > >but add a bit of configuration and such, but nothing that is likely to > >be interesting to this. pgbouncer isn't required, certainly PG is but > >it could be 11, 12, or probably others. We use exim for most of our > >email and we have a wrapper package to make exim wait for PG to start up > >since our exim config queries the PG database for things (the actual > >exim config is backed up but isn't part of the package since it's > >different on our pglister system from other ones). > > > >As much as I'd like this to all be automated and easy to deploy, it > >isn't today. One thing that might help is that the processes which need > >to be started are: > > > >/lib/systemd/system/pglister_bounceprocessor.service > >/lib/systemd/system/pglister_cleanup.service > >/lib/systemd/system/pglister_eximcleaner.service > >/lib/systemd/system/pglister_mailprocessor.service > >/lib/systemd/system/pglister_mailsender.service > >/lib/systemd/system/pglister_moderationprocessor.service > > Interesting, but I still don't see where these packages/code/etc can be > found? They're not somewhere public today and I don't know if that's likely to change anytime soon. Thanks, Stephen
Attachment
On 8/16/21 12:45 PM, Stephen Frost wrote: > Greetings, > > * Adrian Klaver (adrian.klaver@aklaver.com) wrote: >> On 8/16/21 9:06 AM, Stephen Frost wrote: >>>> I have yet to find a pginfra project so I am trying to figure out where the >>>> packages come from? >>> >>> /lib/systemd/system/pglister_bounceprocessor.service >>> /lib/systemd/system/pglister_cleanup.service >>> /lib/systemd/system/pglister_eximcleaner.service >>> /lib/systemd/system/pglister_mailprocessor.service >>> /lib/systemd/system/pglister_mailsender.service >>> /lib/systemd/system/pglister_moderationprocessor.service >> >> Interesting, but I still don't see where these packages/code/etc can be >> found? > > They're not somewhere public today and I don't know if that's likely to > change anytime soon. So trying to do what is asked in this message: https://www.postgresql.org/message-id/20210813183537.GZ17906%40tamriel.snowman.net is not possible? > > Thanks, > > Stephen > -- Adrian Klaver adrian.klaver@aklaver.com
Greetings, * Adrian Klaver (adrian.klaver@aklaver.com) wrote: > On 8/16/21 12:45 PM, Stephen Frost wrote: > >* Adrian Klaver (adrian.klaver@aklaver.com) wrote: > >>On 8/16/21 9:06 AM, Stephen Frost wrote: > >>>>I have yet to find a pginfra project so I am trying to figure out where the > >>>>packages come from? > > >>> > >>>/lib/systemd/system/pglister_bounceprocessor.service > >>>/lib/systemd/system/pglister_cleanup.service > >>>/lib/systemd/system/pglister_eximcleaner.service > >>>/lib/systemd/system/pglister_mailprocessor.service > >>>/lib/systemd/system/pglister_mailsender.service > >>>/lib/systemd/system/pglister_moderationprocessor.service > >> > >>Interesting, but I still don't see where these packages/code/etc can be > >>found? > > > >They're not somewhere public today and I don't know if that's likely to > >change anytime soon. > > So trying to do what is asked in this message: > > https://www.postgresql.org/message-id/20210813183537.GZ17906%40tamriel.snowman.net > > is not possible? Erm, no, of course it's possible, just like it's possible to install PG without using apt-get. Further, the packages you're asking about aren't actually packages that include the software, as I said before, they're just wrappers, so they wouldn't actually help you when it comes to installing the actual software anyway. Thanks, Stephen
Attachment
On 8/16/21 12:50 PM, Stephen Frost wrote: > Greetings, > > * Adrian Klaver (adrian.klaver@aklaver.com) wrote: >> On 8/16/21 12:45 PM, Stephen Frost wrote: >>> * Adrian Klaver (adrian.klaver@aklaver.com) wrote: >>>> On 8/16/21 9:06 AM, Stephen Frost wrote: >>>>>> I have yet to find a pginfra project so I am trying to figure out where the >>>>>> packages come from? >> >>>>> >>>>> /lib/systemd/system/pglister_bounceprocessor.service >>>>> /lib/systemd/system/pglister_cleanup.service >>>>> /lib/systemd/system/pglister_eximcleaner.service >>>>> /lib/systemd/system/pglister_mailprocessor.service >>>>> /lib/systemd/system/pglister_mailsender.service >>>>> /lib/systemd/system/pglister_moderationprocessor.service >>>> >>>> Interesting, but I still don't see where these packages/code/etc can be >>>> found? >>> >>> They're not somewhere public today and I don't know if that's likely to >>> change anytime soon. >> >> So trying to do what is asked in this message: >> >> https://www.postgresql.org/message-id/20210813183537.GZ17906%40tamriel.snowman.net >> >> is not possible? > > Erm, no, of course it's possible, just like it's possible to install PG > without using apt-get. Further, the packages you're asking about aren't > actually packages that include the software, as I said before, they're > just wrappers, so they wouldn't actually help you when it comes to > installing the actual software anyway. So pginfra is not needed for this, it is just the setup code to build the infrastructure that runs on the live server. Basically you need the programs named in the pginfra-* packages(pgbouncer optional), posted earlier, installed on your machine and then the pglister code dropped into say a virtenv. > > Thanks, > > Stephen > -- Adrian Klaver adrian.klaver@aklaver.com
Greetings, * Adrian Klaver (adrian.klaver@aklaver.com) wrote: > On 8/16/21 12:50 PM, Stephen Frost wrote: > >* Adrian Klaver (adrian.klaver@aklaver.com) wrote: > >>On 8/16/21 12:45 PM, Stephen Frost wrote: > >>>* Adrian Klaver (adrian.klaver@aklaver.com) wrote: > >>>>On 8/16/21 9:06 AM, Stephen Frost wrote: > >>>>>>I have yet to find a pginfra project so I am trying to figure out where the > >>>>>>packages come from? > >> > >>>>> > >>>>>/lib/systemd/system/pglister_bounceprocessor.service > >>>>>/lib/systemd/system/pglister_cleanup.service > >>>>>/lib/systemd/system/pglister_eximcleaner.service > >>>>>/lib/systemd/system/pglister_mailprocessor.service > >>>>>/lib/systemd/system/pglister_mailsender.service > >>>>>/lib/systemd/system/pglister_moderationprocessor.service > >>>> > >>>>Interesting, but I still don't see where these packages/code/etc can be > >>>>found? > >>> > >>>They're not somewhere public today and I don't know if that's likely to > >>>change anytime soon. > >> > >>So trying to do what is asked in this message: > >> > >>https://www.postgresql.org/message-id/20210813183537.GZ17906%40tamriel.snowman.net > >> > >>is not possible? > > > >Erm, no, of course it's possible, just like it's possible to install PG > >without using apt-get. Further, the packages you're asking about aren't > >actually packages that include the software, as I said before, they're > >just wrappers, so they wouldn't actually help you when it comes to > >installing the actual software anyway. > > So pginfra is not needed for this, it is just the setup code to build the > infrastructure that runs on the live server. > > Basically you need the programs named in the pginfra-* packages(pgbouncer > optional), posted earlier, installed on your machine and then the pglister > code dropped into say a virtenv. Yeah, and then you need to configure your MTA to run the inject.py as I mentioned in the prior email (or just run it by hand for testing) after you're got the django environment set up and such. Then just have to start up the various daemons that are mentioned above, either through systemd or just by hand, to see them running. Yes, there's probably more than just these items, but I don't know offhand what else is needed. What would be great is for someone to take these steps and then see what else needs to be done and to document all of it. Thanks, Stephen
Attachment
On Mon, Aug 16, 2021 at 9:45 PM Stephen Frost <sfrost@snowman.net> wrote: > > Greetings, > > * Adrian Klaver (adrian.klaver@aklaver.com) wrote: > > On 8/16/21 9:06 AM, Stephen Frost wrote: > > >>I have yet to find a pginfra project so I am trying to figure out where the > > >>packages come from? > > > > > >Those are just wrappers that we build that depend on regular packages > > >but add a bit of configuration and such, but nothing that is likely to > > >be interesting to this. pgbouncer isn't required, certainly PG is but > > >it could be 11, 12, or probably others. We use exim for most of our > > >email and we have a wrapper package to make exim wait for PG to start up > > >since our exim config queries the PG database for things (the actual > > >exim config is backed up but isn't part of the package since it's > > >different on our pglister system from other ones). > > > > > >As much as I'd like this to all be automated and easy to deploy, it > > >isn't today. One thing that might help is that the processes which need > > >to be started are: > > > > > >/lib/systemd/system/pglister_bounceprocessor.service > > >/lib/systemd/system/pglister_cleanup.service > > >/lib/systemd/system/pglister_eximcleaner.service > > >/lib/systemd/system/pglister_mailprocessor.service > > >/lib/systemd/system/pglister_mailsender.service > > >/lib/systemd/system/pglister_moderationprocessor.service > > > > Interesting, but I still don't see where these packages/code/etc can be > > found? > > They're not somewhere public today and I don't know if that's likely to > change anytime soon. The systemd service files are. They're in the suspiciously named "systemd" directory in the repo... But yes, a general list of requirements is definitely missing, along with proper detailed instructions. I definitely recall a mythical "someone" offering to do a test install and actually write that all up, btu I can't remember who it was and clearly that project was never finished :/ -- Magnus Hagander Me: https://www.hagander.net/ Work: https://www.redpill-linpro.com/
On Fri, Aug 13, 2021 at 7:18 PM Stephen Frost <sfrost@snowman.net> wrote: > > Greetings, > > * Adrian Klaver (adrian.klaver@aklaver.com) wrote: > > On 8/13/21 10:03 AM, Stephen Frost wrote: > > >* Adrian Klaver (adrian.klaver@aklaver.com) wrote: > > >>On 8/13/21 9:53 AM, Magnus Hagander wrote: > > >> > > >>>Nope, I think the common factor is they're cced to multiple lists. We > > >>>discourage that pretty strongly and for that reason all such emails are > > >>>moderated. > > >> > > >>I don't suppose it is possible to include that information in the moderation > > >>message? If nothing else as a hint. Muscle memory says hit Reply All. If I > > >>got a reminder of the policy I could cancel the message and resend to a > > >>single list. > > > > > >Isn't implemented today ... but +1 on figuring out a way to do that. > > > > Could it not be added as: > > > > Hint: You may be getting this message if your post was cross posted. Either > > wait for moderator approval or cancel the post and resend to a single list. > > > > to here?: > > > > https://gitlab.com/pglister/pglister/-/blob/master/mailtemplates/moderation_response.txt > > No, I don't think just adding it to every moderation notice that is sent > out would be very good. If we're going to do this then we should make > it so that the notice actually says why it was moderated, if the issue > was cross-posting and then have a custom notice for that. I think that would be a really bad idea to add to *every* notice. Doing it explicitly for moderated-because-of-multi-list-cc I can definitely get on board with. But I think it would be counterproductive to include the details of why something is moderated when it is for example moderated based on a regexp we've (permanently or temporarily) added to deal with somebody abusing the lists. That would give them a very direct blueprint for how to try to get around said block. Likewise if we did it for "everything except regexp matches", because then that would be giving away the same information based on lack of information. Another option might be to include it in the *approval* message. Yes, the information would be more delayed, but at least it would be delivered. And in that case those that were moderated and later discarded or rejected would then just not get the information. -- Magnus Hagander Me: https://www.hagander.net/ Work: https://www.redpill-linpro.com/
Hello! On 16/08/2021 16:55, Stephen Frost wrote: >> Yeah, it seems before changes to code can be made there needs to be some >> sort of README/INSTALL that outlines the process for getting a working >> install to develop against. Is there a place this lives or could live? > > That'd certainly be helpful. I would think that such a document would > live in the git repo.. > > I'm also trying to pull in some other folks who I know are playing > around with pglister who might be able to help. I've been working on an ansible script to install pglister: https://gitlab.com/cmatte/ansible-pglister It's still a work-in-progress. For the moment, it installs a web server (with certificates), postgresql, exim and (partially) jango. I still have to move to a generic authentication system, finish jango install and test exim integration. If anyone wants to participate in this effort, feel free to join and send PRs. Hopefully I can also write a readme when I have a better understanding of the whole process. Cheers, -- Célestin Matte