Thread: 9.2beta web issues

9.2beta web issues

From
Tom Lane
Date:
In bug #6715, Greg Davidson writes
> Please also note these infelicities:
> (1) There's no 9.2beta2 choice in the bug report tool.
> (2) On the beta announcement web page
> http://www.postgresql.org/about/news/1395/
> it says that information on how to test and report bugs with the beta are on
> the web page
> http://www.postgresql.org/developer/beta/
> but no such information is on that page!
        regards, tom lane


Re: 9.2beta web issues

From
Magnus Hagander
Date:
On Thu, Jul 5, 2012 at 11:19 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> In bug #6715, Greg Davidson writes
>> Please also note these infelicities:
>> (1) There's no 9.2beta2 choice in the bug report tool.
>> (2) On the beta announcement web page
>> http://www.postgresql.org/about/news/1395/
>> it says that information on how to test and report bugs with the beta are on
>> the web page
>> http://www.postgresql.org/developer/beta/
>> but no such information is on that page!

IIRC we don't normally add this, because we don't want to "assign
bugs" to things in beta. Also IIRC, didn't Berkus set up a whole
separate list that people are supposed to report beta bugs on? (no, I
don't agree with that, but that was the intention of pgsql-testers -
and I was voted down at the time).

We can certainly look at changing that policy, but that's the *reason*
why it's not on there at all now...


-- Magnus HaganderMe: http://www.hagander.net/Work: http://www.redpill-linpro.com/


Re: 9.2beta web issues

From
Tom Lane
Date:
Magnus Hagander <magnus@hagander.net> writes:
> IIRC we don't normally add this, because we don't want to "assign
> bugs" to things in beta. Also IIRC, didn't Berkus set up a whole
> separate list that people are supposed to report beta bugs on?

Geez, I sure hope not, because if there is one I am not subscribed to it
(AFAIK), and I bet few other hackers are either.  In any case, it's the
height of folly to imagine that people won't try to use the bug report
form for beta bugs.
        regards, tom lane


Re: 9.2beta web issues

From
Magnus Hagander
Date:
On Thu, Jul 5, 2012 at 11:41 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Magnus Hagander <magnus@hagander.net> writes:
>> IIRC we don't normally add this, because we don't want to "assign
>> bugs" to things in beta. Also IIRC, didn't Berkus set up a whole
>> separate list that people are supposed to report beta bugs on?
>
> Geez, I sure hope not, because if there is one I am not subscribed to it
> (AFAIK), and I bet few other hackers are either.  In any case, it's the
> height of folly to imagine that people won't try to use the bug report
> form for beta bugs.

In that case, I definitely suggest we retire it. It's listed on both
http://www.postgresql.org/community/lists/ and in the archives, and is
labeled as the one where to post such reports... I think the idea was
that Josh was going to somehow "filter" the reports before putting
them on -bugs, making sure they were proper, etc. And that the list
would also take *positive* test results, which we definitely don't
want listed as bugs. But I'm definitely +1 for getting rid of the
distinction.


-- Magnus HaganderMe: http://www.hagander.net/Work: http://www.redpill-linpro.com/


Re: 9.2beta web issues

From
Alvaro Herrera
Date:
Excerpts from Magnus Hagander's message of jue jul 05 17:22:03 -0400 2012:
> On Thu, Jul 5, 2012 at 11:19 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:

> >> Please also note these infelicities:
> >> (1) There's no 9.2beta2 choice in the bug report tool.

> IIRC we don't normally add this, because we don't want to "assign
> bugs" to things in beta.

This sounds rather nonsensical to me.

--
Álvaro Herrera <alvherre@commandprompt.com>
The PostgreSQL Company - Command Prompt, Inc.
PostgreSQL Replication, Consulting, Custom Development, 24x7 support


Re: 9.2beta web issues

From
Alvaro Herrera
Date:
Excerpts from Magnus Hagander's message of jue jul 05 17:44:10 -0400 2012:
> On Thu, Jul 5, 2012 at 11:41 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> > Magnus Hagander <magnus@hagander.net> writes:
> >> IIRC we don't normally add this, because we don't want to "assign
> >> bugs" to things in beta. Also IIRC, didn't Berkus set up a whole
> >> separate list that people are supposed to report beta bugs on?
> >
> > Geez, I sure hope not, because if there is one I am not subscribed to it
> > (AFAIK), and I bet few other hackers are either.  In any case, it's the
> > height of folly to imagine that people won't try to use the bug report
> > form for beta bugs.
>
> In that case, I definitely suggest we retire it. It's listed on both
> http://www.postgresql.org/community/lists/ and in the archives, and is
> labeled as the one where to post such reports... I think the idea was
> that Josh was going to somehow "filter" the reports before putting
> them on -bugs, making sure they were proper, etc. And that the list
> would also take *positive* test results, which we definitely don't
> want listed as bugs. But I'm definitely +1 for getting rid of the
> distinction.

That list's traffic seems a bit on the low side -- surely we've had
people testing more recently than September 2011.

--
Álvaro Herrera <alvherre@commandprompt.com>
The PostgreSQL Company - Command Prompt, Inc.
PostgreSQL Replication, Consulting, Custom Development, 24x7 support


Re: 9.2beta web issues

From
Tom Lane
Date:
Alvaro Herrera <alvherre@commandprompt.com> writes:
> That list's traffic seems a bit on the low side -- surely we've had
> people testing more recently than September 2011.

pgsql-testers is what we're talking about?

9/2011 would be when the last beta phase ended, and after that would
have been a period where zero traffic could be expected.  Presumably
people have forgotten about the special list since then, accounting for
the lack of 9.2beta reports in that list.  But even before that, the
traffic looks darn thin.
        regards, tom lane


Re: 9.2beta web issues

From
Magnus Hagander
Date:
On Fri, Jul 6, 2012 at 12:44 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Alvaro Herrera <alvherre@commandprompt.com> writes:
>> That list's traffic seems a bit on the low side -- surely we've had
>> people testing more recently than September 2011.
>
> pgsql-testers is what we're talking about?

Yes.


> 9/2011 would be when the last beta phase ended, and after that would
> have been a period where zero traffic could be expected.  Presumably
> people have forgotten about the special list since then, accounting for
> the lack of 9.2beta reports in that list.  But even before that, the
> traffic looks darn thin.

Yeah, I'm not sure the plan to get people over there ever really
worked. And most likely this was because the hackers weren't really
onboard with the plan...

So let's just get rid of the list (well, move it to inactive lists).

I'm looking at what it takes to add the beta version to the bugs form.
Right now that's driven by the same boolean field that controls which
versions are listed as supported on places like the frontpage of the
site, which is probably not the best idea to do - so I'll go add
another field that controls this, and then get that deployed shortly.

-- Magnus HaganderMe: http://www.hagander.net/Work: http://www.redpill-linpro.com/


Re: 9.2beta web issues

From
Stefan Kaltenbrunner
Date:
On 07/06/2012 12:44 AM, Tom Lane wrote:
> Alvaro Herrera <alvherre@commandprompt.com> writes:
>> That list's traffic seems a bit on the low side -- surely we've had
>> people testing more recently than September 2011.
> 
> pgsql-testers is what we're talking about?
> 
> 9/2011 would be when the last beta phase ended, and after that would
> have been a period where zero traffic could be expected.  Presumably
> people have forgotten about the special list since then, accounting for
> the lack of 9.2beta reports in that list.  But even before that, the
> traffic looks darn thin.

yeah same goes for the (imho completely failed) experiment of using
google docs(...) for collecting test results as recommended on
https://wiki.postgresql.org/wiki/HowToBetaTest (which is linked to from
the main site on http://www.postgresql.org/developer/alpha/).

I think we should completely drop all references to the google docs
thingy and the -testers list there and redirect people simply to -bugs
and/or hackers.

Comments?



Stefan


Re: 9.2beta web issues

From
Peter Geoghegan
Date:
On 6 July 2012 19:23, Stefan Kaltenbrunner <stefan@kaltenbrunner.cc> wrote:
> On 07/06/2012 12:44 AM, Tom Lane wrote:
> yeah same goes for the (imho completely failed) experiment of using
> google docs(...) for collecting test results as recommended on
> https://wiki.postgresql.org/wiki/HowToBetaTest (which is linked to from
> the main site on http://www.postgresql.org/developer/alpha/).
>
> I think we should completely drop all references to the google docs
> thingy and the -testers list there and redirect people simply to -bugs
> and/or hackers.

+1.

I'm not necessarily opposed to have a web interface, or even a
full-blown bug tracker, but having just read the wiki page about beta
testing, it sort of sounds like this was instituted as an official
policy. That seems like a questionable way to present it to me,
because I wasn't even aware that there was a google form.

-- 
Peter Geoghegan       http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training and Services


Re: 9.2beta web issues

From
Josh Berkus
Date:
>> I think we should completely drop all references to the google docs
>> thingy and the -testers list there and redirect people simply to -bugs
>> and/or hackers.

Oh, dammit.  I thought I took the google thing out when we went beta,
but apparently I didn't save my changes.

People do use -testers.  While -bugs/-hackers is superior for
experienced testers, submitting stuff to those lists can be a bit
intimidating.

-- 
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com




Re: 9.2beta web issues

From
Josh Berkus
Date:
On 7/9/12 4:22 PM, Josh Berkus wrote:
> 
>>> I think we should completely drop all references to the google docs
>>> thingy and the -testers list there and redirect people simply to -bugs
>>> and/or hackers.
> 
> Oh, dammit.  I thought I took the google thing out when we went beta,
> but apparently I didn't save my changes.
> 
> People do use -testers.  While -bugs/-hackers is superior for
> experienced testers, submitting stuff to those lists can be a bit
> intimidating.

See new text on wiki.

-- 
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com




Re: 9.2beta web issues

From
Magnus Hagander
Date:
On Tue, Jul 10, 2012 at 1:34 AM, Josh Berkus <josh@agliodbs.com> wrote:
> On 7/9/12 4:22 PM, Josh Berkus wrote:
>>
>>>> I think we should completely drop all references to the google docs
>>>> thingy and the -testers list there and redirect people simply to -bugs
>>>> and/or hackers.
>>
>> Oh, dammit.  I thought I took the google thing out when we went beta,
>> but apparently I didn't save my changes.
>>
>> People do use -testers.  While -bugs/-hackers is superior for
>> experienced testers, submitting stuff to those lists can be a bit
>> intimidating.

But how does that actually help if nobody who could fix things read it?

> See new text on wiki.

The "what goes in the field" part seems to still talk about a google
doc that now doesn't exist - like "mark multiple options if...".

-- Magnus HaganderMe: http://www.hagander.net/Work: http://www.redpill-linpro.com/


Re: 9.2beta web issues

From
Josh Berkus
Date:
> The "what goes in the field" part seems to still talk about a google
> doc that now doesn't exist - like "mark multiple options if...".

Yeah, needs more revision.  Feel free!

-- 
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com




Re: 9.2beta web issues

From
Stefan Kaltenbrunner
Date:
On 07/10/2012 01:22 AM, Josh Berkus wrote:
> 
>>> I think we should completely drop all references to the google docs
>>> thingy and the -testers list there and redirect people simply to -bugs
>>> and/or hackers.
> 
> Oh, dammit.  I thought I took the google thing out when we went beta,
> but apparently I didn't save my changes.
> 
> People do use -testers.  While -bugs/-hackers is superior for
> experienced testers, submitting stuff to those lists can be a bit
> intimidating.

hmm I'm still wondering if it is a good thing, the description of the
list on the archives is "For active testers of PostgreSQL Alphas and
Betas to report test results." or if you ask mj2:

"The pgsql-testers mailing list is for users committed to helping with
Alpha and Beta testing of new versions of PostgreSQL. It is also use for
automated processing of testing results. "

apart from the typo in the second one(what automated processing btw - is
that google docs?) I basically parse both version as "this is the
official way to report bugs in beta/alpha", I'm not sure this is
actually understood by every community member and a sensible split of
traffic between -bugs and hackers...


Stefan


Re: 9.2beta web issues

From
Josh Berkus
Date:
> apart from the typo in the second one(what automated processing btw - is
> that google docs?) I basically parse both version as "this is the
> official way to report bugs in beta/alpha", I'm not sure this is
> actually understood by every community member and a sensible split of
> traffic between -bugs and hackers...

Yeah, we should update that description.

-- 
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com




Re: 9.2beta web issues

From
Alvaro Herrera
Date:
Excerpts from Josh Berkus's message of mar jul 10 13:55:26 -0400 2012:
>
> > apart from the typo in the second one(what automated processing btw - is
> > that google docs?) I basically parse both version as "this is the
> > official way to report bugs in beta/alpha", I'm not sure this is
> > actually understood by every community member and a sensible split of
> > traffic between -bugs and hackers...
>
> Yeah, we should update that description.

Just let me know a suitable wording and I'll put it up in both places.

--
Álvaro Herrera <alvherre@commandprompt.com>
The PostgreSQL Company - Command Prompt, Inc.
PostgreSQL Replication, Consulting, Custom Development, 24x7 support


Re: 9.2beta web issues

From
Josh Berkus
Date:
On 7/10/12 11:06 AM, Alvaro Herrera wrote:
> 
> Excerpts from Josh Berkus's message of mar jul 10 13:55:26 -0400 2012:
>>
>>> apart from the typo in the second one(what automated processing btw - is
>>> that google docs?) I basically parse both version as "this is the
>>> official way to report bugs in beta/alpha", I'm not sure this is
>>> actually understood by every community member and a sensible split of
>>> traffic between -bugs and hackers...
>>
>> Yeah, we should update that description.
> 
> Just let me know a suitable wording and I'll put it up in both places.
> 

Archives:

"For active testers of PostgreSQL Alphas and Betas who want to discuss
testing."

MJ:

"The pgsql-testers mailing list is for users committed to helping with
Alpha and Beta testing of new versions of PostgreSQL to discuss testing."

BTW, I'm mainly interested in preserving the pgsql-testers mailing list
for future use; I'm still hoping to find time to get a serious volunteer
QA effort together, in which case it will need a mailing list.  So we
could kill off pgsql-testers now, but I'll probably just be asking to
revive it next year if you do.

-- 
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com




Re: 9.2beta web issues

From
Bruce Momjian
Date:
On Tue, Jul 10, 2012 at 11:15:00AM -0700, Josh Berkus wrote:
> BTW, I'm mainly interested in preserving the pgsql-testers mailing list
> for future use; I'm still hoping to find time to get a serious volunteer
> QA effort together, in which case it will need a mailing list.  So we
> could kill off pgsql-testers now, but I'll probably just be asking to
> revive it next year if you do.

I have never been a fan of our cycle of "let's try something --- oh,
that didn't work --- let's not tell anyone --- (years later) why is this
here".  I vote we remove it, and in the future, if we decide not to use
something, let's not assume some amazing idea is going to fix it ---
let's just remove it proactively.

--  Bruce Momjian  <bruce@momjian.us>        http://momjian.us EnterpriseDB
http://enterprisedb.com
 + It's impossible for everything to be true. +


Re: 9.2beta web issues

From
Stefan Kaltenbrunner
Date:
On 07/10/2012 08:29 PM, Bruce Momjian wrote:
> On Tue, Jul 10, 2012 at 11:15:00AM -0700, Josh Berkus wrote:
>> BTW, I'm mainly interested in preserving the pgsql-testers mailing list
>> for future use; I'm still hoping to find time to get a serious volunteer
>> QA effort together, in which case it will need a mailing list.  So we
>> could kill off pgsql-testers now, but I'll probably just be asking to
>> revive it next year if you do.
> 
> I have never been a fan of our cycle of "let's try something --- oh,
> that didn't work --- let's not tell anyone --- (years later) why is this
> here".  I vote we remove it, and in the future, if we decide not to use
> something, let's not assume some amazing idea is going to fix it ---
> let's just remove it proactively.

I fully agree - if we ever figure out how that "serious QA" thingy is
going to work we can easily add a dedicated and maybe even more
appropriately named list. refacturing a list that we have people
subscribed to under a specific purpose into something that might be
something entirely different "next year" seems completely wrong anyway...


Stefan


Re: 9.2beta web issues

From
Tom Lane
Date:
Stefan Kaltenbrunner <stefan@kaltenbrunner.cc> writes:
> I fully agree - if we ever figure out how that "serious QA" thingy is
> going to work we can easily add a dedicated and maybe even more
> appropriately named list. refacturing a list that we have people
> subscribed to under a specific purpose into something that might be
> something entirely different "next year" seems completely wrong anyway...

+1.  Adding a mailing list is not exactly the hardest part of getting
a formal QA effort going.  Moreover, if the previous rationale for having
pgsql-testing was that it was a place where newbies could feel sheltered
while posting bug reports, how are the existing subscribers going to
relate to what Josh is now suggesting?  Repurposing this list for that
usage seems entirely inappropriate.
        regards, tom lane


Re: 9.2beta web issues

From
Josh Berkus
Date:
> +1.  Adding a mailing list is not exactly the hardest part of getting
> a formal QA effort going.  Moreover, if the previous rationale for having
> pgsql-testing was that it was a place where newbies could feel sheltered
> while posting bug reports, how are the existing subscribers going to
> relate to what Josh is now suggesting?  Repurposing this list for that
> usage seems entirely inappropriate.

Well, the ones who've spoken up would be fine with it.  But I'm not sure
how many people are on now.

If we kill the list, though, I'll be directing people to submit
*successful* test reports (i.e. "this worked") to pgsql-bugs.  Please
make sure that's what you want.

-- 
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com




Re: 9.2beta web issues

From
Stefan Kaltenbrunner
Date:
On 07/11/2012 10:46 PM, Josh Berkus wrote:
> 
>> +1.  Adding a mailing list is not exactly the hardest part of getting
>> a formal QA effort going.  Moreover, if the previous rationale for having
>> pgsql-testing was that it was a place where newbies could feel sheltered
>> while posting bug reports, how are the existing subscribers going to
>> relate to what Josh is now suggesting?  Repurposing this list for that
>> usage seems entirely inappropriate.
> 
> Well, the ones who've spoken up would be fine with it.  But I'm not sure
> how many people are on now.
> 
> If we kill the list, though, I'll be directing people to submit
> *successful* test reports (i.e. "this worked") to pgsql-bugs.  Please
> make sure that's what you want.

well checking the archives on pgsql-testers (on a quick look) only
turned up a SINGLE "successful" test report in the last two years among
a hundred or so of "this is a bug or looks broken" ones which are just
fine for -bugs, so I dont see how that will be a problem...


Stefan


Re: 9.2beta web issues

From
Tom Lane
Date:
Josh Berkus <josh@agliodbs.com> writes:
> If we kill the list, though, I'll be directing people to submit
> *successful* test reports (i.e. "this worked") to pgsql-bugs.  Please
> make sure that's what you want.

No, it is not, and please don't threaten us with abusing other lists
because we are questioning whether this one is still worth anything
or not.
        regards, tom lane


Re: 9.2beta web issues

From
Alvaro Herrera
Date:
Excerpts from Stefan Kaltenbrunner's message of mié jul 11 16:52:02 -0400 2012:
>
> On 07/11/2012 10:46 PM, Josh Berkus wrote:
> >
> >> +1.  Adding a mailing list is not exactly the hardest part of getting
> >> a formal QA effort going.  Moreover, if the previous rationale for having
> >> pgsql-testing was that it was a place where newbies could feel sheltered
> >> while posting bug reports, how are the existing subscribers going to
> >> relate to what Josh is now suggesting?  Repurposing this list for that
> >> usage seems entirely inappropriate.
> >
> > Well, the ones who've spoken up would be fine with it.  But I'm not sure
> > how many people are on now.
> >
> > If we kill the list, though, I'll be directing people to submit
> > *successful* test reports (i.e. "this worked") to pgsql-bugs.  Please
> > make sure that's what you want.
>
> well checking the archives on pgsql-testers (on a quick look) only
> turned up a SINGLE "successful" test report in the last two years among
> a hundred or so of "this is a bug or looks broken" ones which are just
> fine for -bugs, so I dont see how that will be a problem...

... so maybe there is no point in asking people to submit successful
test reports.  After all, what value is there in them?

If we want to have more successful test reports, ask people to add
machines to the buildfarm instead, and/or enable the 9.2 branch.  Just
for fun, here's the list of animals that are running 9.1 but not 9.2:

anchovy anole baiji bushpig castoroides cormorant dugong fennec grebe
hamerkop hawker jaguarundi lyrebird mastodon mongoose mule mussel
narwhal okapi polecat protosciurus raven reindeer smew smilodon

--
Álvaro Herrera <alvherre@commandprompt.com>
The PostgreSQL Company - Command Prompt, Inc.
PostgreSQL Replication, Consulting, Custom Development, 24x7 support


Re: 9.2beta web issues

From
Josh Berkus
Date:
> well checking the archives on pgsql-testers (on a quick look) only
> turned up a SINGLE "successful" test report in the last two years among
> a hundred or so of "this is a bug or looks broken" ones which are just
> fine for -bugs, so I dont see how that will be a problem...

Oh?  I remembered more that that.  A non-issue then.

-- 
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com




Re: 9.2beta web issues

From
Josh Berkus
Date:
On 7/11/12 2:34 PM, Josh Berkus wrote:
> 
>> well checking the archives on pgsql-testers (on a quick look) only
>> turned up a SINGLE "successful" test report in the last two years among
>> a hundred or so of "this is a bug or looks broken" ones which are just
>> fine for -bugs, so I dont see how that will be a problem...
> 
> Oh?  I remembered more that that.  A non-issue then.

... so let me post to the list before you kill it.


-- 
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com




Re: 9.2beta web issues

From
Josh Berkus
Date:
>> Oh?  I remembered more that that.  A non-issue then.
> 
> ... so let me post to the list before you kill it.

OK, list shutdown notice posted.  If there's some way you can keep the
archives but shut down the list that would be nice.

-- 
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com




Re: 9.2beta web issues

From
Alvaro Herrera
Date:
Excerpts from Josh Berkus's message of mié jul 11 17:39:03 -0400 2012:
>
> >> Oh?  I remembered more that that.  A non-issue then.
> >
> > ... so let me post to the list before you kill it.
>
> OK, list shutdown notice posted.  If there's some way you can keep the
> archives but shut down the list that would be nice.

I forgot to follow up here, but the list has been shut down (i.e. if you
email it, you'll get an error message back) and moved to the inactive
lists section.  Archives have been kept online as requested.

--
Álvaro Herrera <alvherre@commandprompt.com>
The PostgreSQL Company - Command Prompt, Inc.
PostgreSQL Replication, Consulting, Custom Development, 24x7 support