Thread: Problem with moderation of messages with patched attached.

Hi Dave,

> Message to list pgsql-hackers held for moderation due to 'Size 1MB (1061796 bytes) is larger than threshold 1000KB
(1024000bytes)', notice queued for 2 moderators
 

Thanks! Does anyone know if cfbot understands .patch.gz and/or .tgz ?

-- 
Best regards,
Aleksander Alekseev

On Thu, Mar 03, 2022 at 04:24:03PM +0300, Aleksander Alekseev wrote:
> 
> Thanks! Does anyone know if cfbot understands .patch.gz and/or .tgz ?

There's a FAQ link on the cfbot main page that answers this kind of questions.

Hi Dave,

> Then someone will complain if their patch is 2.1MB! How often are messages legitimately over 1MB anyway, even with a
patch?I don't usually moderate -hackers, so I don't know if this is a common thing or not.
 
>
> I'll ping a message across to the sysadmin team anyway; I can't just change that setting without buy-in from the rest
ofthe team.
 

IMO, current limits are OK. The actual problem is that when the
message gets into moderation, the notice to the author doesn't contain
the reason:

> Your message to pgsql-hackers with subject
> "Re: Add 64-bit XIDs into PostgreSQL 15"
> has been held for moderation.
>
> It will be delivered to the list recipients as soon as it has been
> approved by a moderator.
>
> If you wish to cancel the message without delivery, please click
> this link: ....

Any chance we could include the reason in the message? I foresee that
otherwise such kinds of questions will be asked over and over again.

--
Best regards,
Aleksander Alekseev

On Thu, Mar 03, 2022 at 01:37:35PM +0000, Dave Page wrote:
>
> Then someone will complain if their patch is 2.1MB! How often are messages
> legitimately over 1MB anyway, even with a patch? I don't usually moderate
> -hackers, so I don't know if this is a common thing or not.

It's not common, most people send compressed versions.

Also, gigantic patchsets tend to be hard to maintain and rot pretty fast, so
authors also sometimes maintain a branch on some external repository and just
send newer versions on the ML infrequently.

Hi again,

> Any chance we could include the reason in the message? I foresee that
> otherwise such kinds of questions will be asked over and over again.

A link to the list of common reasons should work too.

-- 
Best regards,
Aleksander Alekseev

Pavel Borisov <pashkin.elfe@gmail.com> writes:
>> The mail system doesn't have the capability to apply different moderation
>> rules for people in that way I'm afraid.

> Maybe then 2MB for everyone?

Maybe your patch needs to be split up?  You're going to have a hard time
finding people who want to review or commit such large chunks.

            regards, tom lane

On Thu, Mar 03, 2022 at 10:17:06AM -0500, Tom Lane wrote:
> Pavel Borisov <pashkin.elfe@gmail.com> writes:
> >> The mail system doesn't have the capability to apply different moderation
> >> rules for people in that way I'm afraid.
> 
> > Maybe then 2MB for everyone?
> 
> Maybe your patch needs to be split up?  You're going to have a hard time
> finding people who want to review or commit such large chunks.

I think it's the total attachment size, not a single file.  So while splitting
up the patchset even more would still be a good idea, compressing the files
before sending them to hundred of people would be an even better one.

Greetings,

* Aleksander Alekseev (aleksander@timescale.com) wrote:
> My last email to pgsql-jobs@ was moderated in a similar fashion. To my
> knowledge that mailing list is not pre-moderated. So it may have the same
> problem, and not only with patches. (We use regular Google Workspace.)

-jobs is moderated.

Thanks,

Stephen

Greetings,

* Pavel Borisov (pashkin.elfe@gmail.com) wrote:
> > On Thu, 3 Mar 2022 at 13:28, Pavel Borisov <pashkin.elfe@gmail.com> wrote:
> >> The mail system doesn't have the capability to apply different moderation
> >>> rules for people in that way I'm afraid.
> >>>
> >> Maybe then 2MB for everyone? Otherwise it's not so convenient. Lead to
> >> answers before the questions in the thread [1], seems weird.
> >
> > Then someone will complain if their patch is 2.1MB! How often are messages
> > legitimately over 1MB anyway, even with a patch? I don't usually moderate
> > -hackers, so I don't know if this is a common thing or not.

I do pay attention to -hackers and no, it doesn't come up very often.

> Authors in the mentioned thread [1] bump into this issue while posting all
> 11 versions of a patchset. It is little bit more than 1MB. We can try to
> use .gz and if this doesn't work we report it again.

This patch set really shoudl be broken down into smaller independent
pieces that attack different parts and not be all one big series of
patches.

Thanks,

Stephen

Hi,

On 2022-03-03 13:37:35 +0000, Dave Page wrote:
> On Thu, 3 Mar 2022 at 13:28, Pavel Borisov <pashkin.elfe@gmail.com> wrote:
> 
> > The mail system doesn't have the capability to apply different moderation
> >> rules for people in that way I'm afraid.
> >>
> > Maybe then 2MB for everyone? Otherwise it's not so convenient. Lead to
> > answers before the questions in the thread [1], seems weird.
> >
> 
> Then someone will complain if their patch is 2.1MB! How often are messages
> legitimately over 1MB anyway, even with a patch? I don't usually moderate
> -hackers, so I don't know if this is a common thing or not.

I don't think it's actually that rare. But most contributors writing that
large patchsets know about the limit and work around it - I gzip patches when
I see the email getting too large. But it's more annoying to work with for
reviewers.

It's somewhat annoying. If you e.g. append a few graphs of performance changes
and a patch it's pretty easy to get into the range where compressing won't
help anymore.

And sure, any limit may be hit by somebody. But 1MB across the whole email
seems pretty low these days.

Greetings,

Andres Freund

On 3/19/22 14:48, Andres Freund wrote:
> Hi,
>
> On 2022-03-03 13:37:35 +0000, Dave Page wrote:
>> On Thu, 3 Mar 2022 at 13:28, Pavel Borisov <pashkin.elfe@gmail.com> wrote:
>>
>>> The mail system doesn't have the capability to apply different moderation
>>>> rules for people in that way I'm afraid.
>>>>
>>> Maybe then 2MB for everyone? Otherwise it's not so convenient. Lead to
>>> answers before the questions in the thread [1], seems weird.
>>>
>> Then someone will complain if their patch is 2.1MB! How often are messages
>> legitimately over 1MB anyway, even with a patch? I don't usually moderate
>> -hackers, so I don't know if this is a common thing or not.
> I don't think it's actually that rare. But most contributors writing that
> large patchsets know about the limit and work around it - I gzip patches when
> I see the email getting too large. But it's more annoying to work with for
> reviewers.
>
> It's somewhat annoying. If you e.g. append a few graphs of performance changes
> and a patch it's pretty easy to get into the range where compressing won't
> help anymore.
>
> And sure, any limit may be hit by somebody. But 1MB across the whole email
> seems pretty low these days.
>

Of course we could get complaints no matter what level we set the limit
at. I think raising it to 2Mb would be a reasonable experiment. If no
observable evil ensues then leave it that way. If it does then roll it
back. I agree that plain uncompressed patches are easier to deal with in
general.


cheers


andrew

--
Andrew Dunstan
EDB: https://www.enterprisedb.com