Thread: Re: [patch]socket_timeout in interfaces/libpq

Re: [patch]socket_timeout in interfaces/libpq

From
Michael Paquier
Date:
On Wed, Jun 26, 2019 at 04:13:36AM +0000, nagaura.ryohei@fujitsu.com wrote:
> I don't think that the rest one of my proposals has been rejected
> completely, so I want to restart discussion.

I recall on the matter that there was consensus that nobody really
liked this option because it enforced a cancellation on the
connection.
--
Michael

Attachment

RE: [patch]socket_timeout in interfaces/libpq

From
"nagaura.ryohei@fujitsu.com"
Date:
Hi, Michael-san.

> From: Michael Paquier <michael@paquier.xyz>
> On Wed, Jun 26, 2019 at 04:13:36AM +0000, nagaura.ryohei@fujitsu.com wrote:
> > I don't think that the rest one of my proposals has been rejected
> > completely, so I want to restart discussion.
> I recall on the matter that there was consensus that nobody really liked this option
> because it enforced a cancellation on the connection.
It seems that you did not think so at that time.
# Please refer to [1]

I don't think all the reviewers are completely negative.
I think some couldn't judge because  lack of what kind of problem I was going to solve and the way to solve it, so I
restartedto describe them in this time. 

[1] https://www.postgresql.org/message-id/20190406065428.GA2145%40paquier.xyz
Best regards,
---------------------
Ryohei Nagaura





Re: [patch]socket_timeout in interfaces/libpq

From
Michael Paquier
Date:
On Wed, Jun 26, 2019 at 11:56:28AM +0000, nagaura.ryohei@fujitsu.com wrote:
> It seems that you did not think so at that time.
> # Please refer to [1]
>
> I don't think all the reviewers are completely negative.

I recall having a negative impression on the patch when first looking
at it, and still have the same impression when looking at the last
version.  Just with a quick look, assuming that you can bypass all
cleanup operations normally taken by pqDropConnection() through a
hijacking of pqWait() is not fine as this routine explicitely assumes
to *never* have a timeout for its wait.
--
Michael

Attachment

Re: [patch]socket_timeout in interfaces/libpq

From
Michael Paquier
Date:
On Tue, Sep 10, 2019 at 03:38:21PM +0900, Michael Paquier wrote:
> I recall having a negative impression on the patch when first looking
> at it, and still have the same impression when looking at the last
> version.  Just with a quick look, assuming that you can bypass all
> cleanup operations normally taken by pqDropConnection() through a
> hijacking of pqWait() is not fine as this routine explicitely assumes
> to *never* have a timeout for its wait.

By the way, Fabien, you are marked as a reviewer of this patch since
the end of June.  Are you planning to review it?
--
Michael

Attachment

Re: [patch]socket_timeout in interfaces/libpq

From
Fabien COELHO
Date:
> By the way, Fabien, you are marked as a reviewer of this patch since the 
> end of June.  Are you planning to review it?

Not this round.

-- 
Fabien.



Re: [patch]socket_timeout in interfaces/libpq

From
Michael Paquier
Date:
On Wed, Sep 11, 2019 at 04:25:07PM +0200, Fabien COELHO wrote:
> Not this round.

You have registered yourself as a reviewer of this patch since the end
of June.  Could you please avoid that?  Sometimes people skips patches
when they see someone already registered to review it.

The patch applies cleanly so I am movingit to next CF.

(FWIW, I still have the same impression as upthread, looking again at
the patch, but let's see if there are other opinions.)
--
Michael

Attachment

Re: [patch]socket_timeout in interfaces/libpq

From
Fabien COELHO
Date:
Michaël,

>> Not this round.
>
> You have registered yourself as a reviewer of this patch since the end
> of June.  Could you please avoid that?  Sometimes people skips patches
> when they see someone already registered to review it.

Yep. ISTM that I did a few reviews on early versions of the patch, which 
was really a set of 3 patches.

> The patch applies cleanly so I am movingit to next CF.
>
> (FWIW, I still have the same impression as upthread, looking again at
> the patch, but let's see if there are other opinions.)

AFAICR, I was partly dissuated to pursue reviews by your comment that 
somehow the feature had no clear consensus, so I thought that the patch 
was implicitely rejected.

Although I work for free, I try to avoid working for nothing:-)

It is still unclear from your above comment whether the patch would ever 
get committed, so this does not motivate spending time on it.

-- 
Fabien.

RE: [patch]socket_timeout in interfaces/libpq

From
"nagaura.ryohei@fujitsu.com"
Date:
Hi, Michael-san.

Sorry, I have missed your e-mail...

> From: Michael Paquier <michael@paquier.xyz>
> On Wed, Jun 26, 2019 at 11:56:28AM +0000, nagaura.ryohei@fujitsu.com wrote:
> > It seems that you did not think so at that time.
> > # Please refer to [1]
> >
> > I don't think all the reviewers are completely negative.
>
> I recall having a negative impression on the patch when first looking at it, and still
> have the same impression when looking at the last version.  Just with a quick
> look, assuming that you can bypass all cleanup operations normally taken by
> pqDropConnection() through a hijacking of pqWait() is not fine as this routine
> explicitely assumes to *never* have a timeout for its wait.
I couldn't understand what you meant.
Do you say that we shouldn't change pqWait() behavior?
Or should I modify my patch to use pqDropConnection()?

Best regards,
---------------------
Ryohei Nagaura





Re: [patch]socket_timeout in interfaces/libpq

From
David Steele
Date:
On 11/29/19 12:22 AM, nagaura.ryohei@fujitsu.com wrote:
> 
>> From: Michael Paquier <michael@paquier.xyz>
>> On Wed, Jun 26, 2019 at 11:56:28AM +0000, nagaura.ryohei@fujitsu.com wrote:
>>> It seems that you did not think so at that time.
>>> # Please refer to [1]
>>>
>>> I don't think all the reviewers are completely negative.
>>
>> I recall having a negative impression on the patch when first looking at it, and still
>> have the same impression when looking at the last version.  Just with a quick
>> look, assuming that you can bypass all cleanup operations normally taken by
>> pqDropConnection() through a hijacking of pqWait() is not fine as this routine
>> explicitely assumes to *never* have a timeout for its wait.
 >
> I couldn't understand what you meant.
> Do you say that we shouldn't change pqWait() behavior?
> Or should I modify my patch to use pqDropConnection()?

This patch no longer applies: http://cfbot.cputube.org/patch_27_2175.log

CF entry has been updated to Waiting on Author.

More importantly it looks like there is still no consensus on this  
patch, which is an uncommitted part of a previous patch [1].

Unless somebody chimes in I'll mark this Returned with Feedback at the  
end of the CF.

Regards,
-- 
-David
david@pgmasters.net

[1]  
https://www.postgresql.org/message-id/raw/20190406065428.GA2145%40paquier.xyz



Re: [patch]socket_timeout in interfaces/libpq

From
David Steele
Date:
On 3/24/20 10:58 AM, David Steele wrote:
> On 11/29/19 12:22 AM, nagaura.ryohei@fujitsu.com wrote:
 >
>> I couldn't understand what you meant.
>> Do you say that we shouldn't change pqWait() behavior?
>> Or should I modify my patch to use pqDropConnection()?
> 
> This patch no longer applies: http://cfbot.cputube.org/patch_27_2175.log
> 
> CF entry has been updated to Waiting on Author.
> 
> More importantly it looks like there is still no consensus on this 
> patch, which is an uncommitted part of a previous patch [1].
> 
> Unless somebody chimes in I'll mark this Returned with Feedback at the 
> end of the CF.

Marked Returned with Feedback.

Regards,
-- 
-David
david@pgmasters.net