Re: a suggestion about patch "ping"-s

Re: a suggestion about patch "ping"-s

[Doug Evans]

Joel Brobecker writes:
  > Hello,
  >
  > First of all, please continue to ping the GDB maintainers when
  > patches go unreviewed. We try to do our best, but as a volunteer-based
  > project, lots of patches unfortunately wait a bit before review.
  > As a reminder, the generally accepted ping frequency is 2 weeks after
  > first submission, and every week thereafter. Some have even accepted
  > the first ping after a week.
  >
  > The real suggestion I would like to make, when sending a ping, is
  > that people reply to the original email, with the patch being quoted
  > in the reply, rather than sending a completely separate email
  > with URLs to the patches. The latter may sound like a good idea,
  > since it allows to send a nice and compact email. But what it
  > does is also make the ping itself a new email thread, unrelated to
  > the emailing submitting the patch. As a result, there are two email
  > threads where a review can happen.
  >
  > For instance, one maintainer could see the ping first, and therefore
  > send a review by finding and then replying to the original thread.
  > But then, a second reviewer might not notice the review, and therefore
  > review again.
  >
  > Another example happened to me, today. I try to keep gdb-patches
  > emails in a mailbox until I either review them, or see someone else
  > do the review. Then comes a new email, un-connected to the original
  > email, with URLs of 3 patches. I had to do a little bit of research
  > to figure out whether they had already been reviewed or not.
  >
  > IMO, just replying to the original email with "PING" clearly shown
  > in the subject's reply will help us save a bit of time, and we can
  > then in turn invest that time is reviewing more submissions.
  >
  > Thank you!
  > --
  > Joel

+1

Re: a suggestion about patch "ping"-s

[Pedro Alves]

On 01/01/2016 11:24 AM, Joel Brobecker wrote:

> The real suggestion I would like to make, when sending a ping, is
> that people reply to the original email, with the patch being quoted
> in the reply, rather than sending a completely separate email
> with URLs to the patches. The latter may sound like a good idea,
> since it allows to send a nice and compact email. But what it
> does is also make the ping itself a new email thread, unrelated to
> the emailing submitting the patch. As a result, there are two email
> threads where a review can happen.
> 
> For instance, one maintainer could see the ping first, and therefore
> send a review by finding and then replying to the original thread.
> But then, a second reviewer might not notice the review, and therefore
> review again.
> 
> Another example happened to me, today. I try to keep gdb-patches
> emails in a mailbox until I either review them, or see someone else
> do the review. Then comes a new email, un-connected to the original
> email, with URLs of 3 patches. I had to do a little bit of research
> to figure out whether they had already been reviewed or not.

This "separate mail with url" seems to be somewhat standard practice on
the gcc-patches@ list, and it's always puzzled me.  I too find it highly
confusing.

+1 on all you said.

I've added this to:

 https://sourceware.org/gdb/wiki/ContributionChecklist#Ping_and_keep_pinging

Thanks,
Pedro Alves

a suggestion about patch "ping"-s

[Joel Brobecker]

Hello,

First of all, please continue to ping the GDB maintainers when
patches go unreviewed. We try to do our best, but as a volunteer-based
project, lots of patches unfortunately wait a bit before review.
As a reminder, the generally accepted ping frequency is 2 weeks after
first submission, and every week thereafter. Some have even accepted
the first ping after a week.

The real suggestion I would like to make, when sending a ping, is
that people reply to the original email, with the patch being quoted
in the reply, rather than sending a completely separate email
with URLs to the patches. The latter may sound like a good idea,
since it allows to send a nice and compact email. But what it
does is also make the ping itself a new email thread, unrelated to
the emailing submitting the patch. As a result, there are two email
threads where a review can happen.

For instance, one maintainer could see the ping first, and therefore
send a review by finding and then replying to the original thread.
But then, a second reviewer might not notice the review, and therefore
review again.

Another example happened to me, today. I try to keep gdb-patches
emails in a mailbox until I either review them, or see someone else
do the review. Then comes a new email, un-connected to the original
email, with URLs of 3 patches. I had to do a little bit of research
to figure out whether they had already been reviewed or not.

IMO, just replying to the original email with "PING" clearly shown
in the subject's reply will help us save a bit of time, and we can
then in turn invest that time is reviewing more submissions.

Thank you!
-- 
Joel