You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@commons.apache.org by Gary Gregory <ga...@gmail.com> on 2019/09/25 21:45:13 UTC

[pool] POOL-376 and release?

Hi All,

Is the fix for POOL-376 important enough to warrant an ASAP release?

Gayr

Re: [pool] POOL-376 and release?

Posted by Phil Steitz <ph...@gmail.com>.

On 10/7/19 8:37 AM, Gary Gregory wrote:
> On Sun, Oct 6, 2019 at 6:27 PM Phil Steitz <ph...@gmail.com> wrote:
>
>>
>> On 10/5/19 12:08 PM, Phil Steitz wrote:
>>> On 10/5/19 5:47 AM, Gary Gregory wrote:
>>>> On Sat, Oct 5, 2019 at 8:17 AM sebb <se...@gmail.com> wrote:
>>>>
>>>>> On Sat, 5 Oct 2019 at 02:32, Gary Gregory <ga...@gmail.com>
>>>>> wrote:
>>>>>> Hi Phil and all:
>>>>>>
>>>>>> It looks like you merged from the "old" git repo
>>>>>> https://git-wip-us.apache.org/repos/asf/commons-pool
>>>>>>
>>>>>> I am not sure why we have two named repos but I am basing my work on
>>>>> GitBox
>>>>>> https://gitbox.apache.org/repos/asf/commons-pool.git
>>>>>>
>>>>>> I think these are in fact the same (?) and the confusion (on my part)
>>>>> only
>>>>>> comes in due to seeing the "old" name in the Git commit history.
>>>>>>
>>>>>> I am not sure if this matters aside from the confusion. Can anyone
>>>>>> elucidate?
>>>>> Try browsing to the two URLs.
>>>>>
>>>>> Do you see any difference in the displayed pages (apart from the URL)?
>>>>>
>>>> Not after looking for 30 seconds, and that's exactly my point: it is
>>>> _confusing_. Are they really _exactly_ the same? Which one is the
>>>> canonical one? Are there two repos or one?
>>>> It would help us all if we standardize on GitBox IMO.
>>>
>>> Sorry for the screw-up.  I think when I set up the clone that I used
>>> to make that patch I was following [1].  We should change the URL
>>> there, I guess.
>>>
>>> I will fix the setup on that box and verify that the logs look right.
>> I have checked the logs and everything looks correct.  Once again, sorry
>> for causing confusion by the commit to wip.  I also verified that that
>> pom has been fixed so that when the web site is published next, the
>> right scm URL will be shown.
>>
>> I think the release is good to go now unless anyone sees problems with
>> the fixes master contains.
>>
> Thank you Phil. I will find the time to cut a RC tonight.

I just pushed another commit.  Looking at the fix again, it was possible 
for the race to still happen (startEviction happens between first check 
and removal from idleObjects).

Phil

>
> Gary
>
>
>> Phil
>>> Phil
>>>
>>> [1] http://commons.apache.org/proper/commons-pool/scm.html
>>>
>>>
>>>> Gary
>>>>
>>>>
>>>>>> Thank you,
>>>>>> Gary
>>>>>>
>>>>>> On Fri, Oct 4, 2019 at 9:06 PM Phil Steitz <ph...@gmail.com>
>>>>> wrote:
>>>>>>> On 10/1/19 4:27 PM, Gary Gregory wrote:
>>>>>>>> On Tue, Oct 1, 2019 at 5:03 PM Phil Steitz <ph...@gmail.com>
>>>>>>> wrote:
>>>>>>>>> Good news.  I think I now understand the actual root cause for
>>>>>>>>> POOL-376.  Bad news is the fix that I committed masks but does not
>>>>>>>>> really fix the problem.  I will update the ticket and commit a full
>>>>> fix
>>>>>>>>> this evening.  I will try to get a test case but that is going
>>>>>>>>> to be
>>>>>>>>> tricky because it requires a race between the evictor and a
>>>>>>>>> borrower
>>>>>>>>> under the right conditions.
>>>>>>>>>
>>>>>>>> Great news! Thank you Phil.
>>>>>>> Sorry, it was POOL-326 that I was still missing a test for.  I just
>>>>>>> added that and a real fix for the issue. See comments on the
>>>>>>> ticket for
>>>>>>> what was going on there and the unit test I added to the GKOP tests.
>>>>>>> Review of the fix would be good before rolling the release.
>>>>>>>
>>>>>>> Phil
>>>>>>>> Gary
>>>>>>>>
>>>>>>>>
>>>>>>>>> Phil
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On 9/28/19 3:56 PM, Gary Gregory wrote:
>>>>>>>>>> On Sat, Sep 28, 2019, 16:43 Phil Steitz <ph...@gmail.com>
>>>>> wrote:
>>>>>>>>>>> Well,  I don’t have one as I don’t have a test case in hand that
>>>>>>> creates
>>>>>>>>>>> the condition other than my hacked version of [performance] that
>>>>>>>>> reliably
>>>>>>>>>>> reproduces it before my last commit (and doesn’t after it).  I
>>>>> have a
>>>>>>>>> plane
>>>>>>>>>>> ride tomorrow when I can make another go at it.  So let’s say
>>>>> give me
>>>>>>> 48
>>>>>>>>>>> hours and if I still have no test case, I would say cut the
>>>>> release
>>>>>>>>> without
>>>>>>>>>>> it.
>>>>>>>>>>>
>>>>>>>>>> Sound good.
>>>>>>>>>>
>>>>>>>>>> Gary
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>> Phil
>>>>>>>>>>>
>>>>>>>>>>>> On Sep 28, 2019, at 2:10 PM, Gary Gregory <
>>>>> garydgregory@gmail.com>
>>>>>>>>>>> wrote:
>>>>>>>>>>>> Phil (sorry for too post; phone),
>>>>>>>>>>>>
>>>>>>>>>>>> May you give me an ETA so I can plan my time accordingly?
>>>>>>>>>>>>
>>>>>>>>>>>> Thank you,
>>>>>>>>>>>> Gary
>>>>>>>>>>>>
>>>>>>>>>>>>> On Thu, Sep 26, 2019, 20:22 Gary Gregory <
>>>>> garydgregory@gmail.com>
>>>>>>>>>>> wrote:
>>>>>>>>>>>>>> On Thu, Sep 26, 2019 at 5:57 PM Phil Steitz <
>>>>> phil.steitz@gmail.com
>>>>>>>>>>> wrote:
>>>>>>>>>>>>>>> On 9/25/19 6:10 PM, Gary Gregory wrote:
>>>>>>>>>>>>>>> On Wed, Sep 25, 2019 at 9:05 PM Phil Steitz <
>>>>>>> phil.steitz@gmail.com>
>>>>>>>>>>>>>> wrote:
>>>>>>>>>>>>>>>>> On 9/25/19 5:47 PM, Gary Gregory wrote:
>>>>>>>>>>>>>>>>> On Wed, Sep 25, 2019 at 8:32 PM Phil Steitz <
>>>>>>>>> phil.steitz@gmail.com>
>>>>>>>>>>>>>>>> wrote:
>>>>>>>>>>>>>>>>>> I would say yes, but I would also like to add a fix for
>>>>>>>>>>>>>>>>>> the
>>>>>>>>>>> similarly
>>>>>>>>>>>>>>>>>> nasty POOL-326.  I can do that in the next 24 hours. While
>>>>> I
>>>>>>>>> still
>>>>>>>>>>>>>> don't
>>>>>>>>>>>>>>>>>> have a test case hitting it and I am not satisfied with my
>>>>>>>>>>>>>> understanding
>>>>>>>>>>>>>>>>>> of why the createCount counter gets messed up, the fix in
>>>>> my
>>>>>>> last
>>>>>>>>>>>>>>>>>> comment on that ticket (check the size of idleObjects
>>>>> instead
>>>>>>> of
>>>>>>>>>>>>>> relying
>>>>>>>>>>>>>>>>>> on createCount) will eliminate the NPE.  I think we should
>>>>> make
>>>>>>>>>>> that
>>>>>>>>>>>>>>>>>> change and push a release with that fix bundled too.
>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> OK, sounds good. I'll wait for your go signal.
>>>>>>>>>>>>>>>> I just pushed the fix for POOL-326.
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> OK, I should be able to get to an RC tomorrow. Hopefully
>>>>> someone
>>>>>>>>> else
>>>>>>>>>>>>>> can
>>>>>>>>>>>>>>> validate the fix on their set up...
>>>>>>>>>>>>>> I am also still working on a test case.
>>>>>>>>>>>>>>
>>>>>>>>>>>>> I will hold off...
>>>>>>>>>>>>>
>>>>>>>>>>>>> Gary
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>> Phil
>>>>>>>>>>>>>>> Gary
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> Phil
>>>>>>>>>>>>>>>>> Garye
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>> Phil
>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>> Hi All,
>>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>> Is the fix for POOL-376 important enough to warrant an
>>>>> ASAP
>>>>>>>>>>> release?
>>>>>>>>>>>>>>>>>>> Gayr
>>>>>>>>>>>>>>>>>>>
>>>>> ---------------------------------------------------------------------
>>>>>>>>>>>>>>>>>> To unsubscribe, e-mail:
>> dev-unsubscribe@commons.apache.org
>>>>>>>>>>>>>>>>>> For additional commands, e-mail:
>>>>> dev-help@commons.apache.org
>>>>> ---------------------------------------------------------------------
>>>>>>>>>>>>>>>> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
>>>>>>>>>>>>>>>> For additional commands, e-mail:
>> dev-help@commons.apache.org
>>>>>>>>>>>>>>>>
>>>>>>> ---------------------------------------------------------------------
>>>>>>>>>>>>>> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
>>>>>>>>>>>>>> For additional commands, e-mail: dev-help@commons.apache.org
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>> ---------------------------------------------------------------------
>>>>>>>>>>> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
>>>>>>>>>>> For additional commands, e-mail: dev-help@commons.apache.org
>>>>>>>>>>>
>>>>>>>>>>>
>>>>> ---------------------------------------------------------------------
>>>>>>>>> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
>>>>>>>>> For additional commands, e-mail: dev-help@commons.apache.org
>>>>>>>>>
>>>>>>>>>
>>>>>>> ---------------------------------------------------------------------
>>>>>>> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
>>>>>>> For additional commands, e-mail: dev-help@commons.apache.org
>>>>>>>
>>>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
>>>>> For additional commands, e-mail: dev-help@commons.apache.org
>>>>>
>>>>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
>> For additional commands, e-mail: dev-help@commons.apache.org
>>
>>

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org

Re: [pool] POOL-376 and release?

Posted by sebb <se...@gmail.com>.

On Mon, 7 Oct 2019 at 08:40, Mark Thomas <ma...@apache.org> wrote:
>
> On 05/10/2019 18:59, Gary Gregory wrote:
> > On Sat, Oct 5, 2019 at 1:19 PM Phil Steitz <ph...@gmail.com> wrote:
> >
> >>
> >> On 10/5/19 5:47 AM, Gary Gregory wrote:
> >>> On Sat, Oct 5, 2019 at 8:17 AM sebb <se...@gmail.com> wrote:
> >>>
> >>>> On Sat, 5 Oct 2019 at 02:32, Gary Gregory <ga...@gmail.com>
> >> wrote:
> >>>>> Hi Phil and all:
> >>>>>
> >>>>> It looks like you merged from the "old" git repo
> >>>>> https://git-wip-us.apache.org/repos/asf/commons-pool
> >>>>>
> >>>>> I am not sure why we have two named repos but I am basing my work on
> >>>> GitBox
> >>>>> https://gitbox.apache.org/repos/asf/commons-pool.git
> >>>>>
> >>>>> I think these are in fact the same (?)
>
> They are now (after a server failure a little while ago). They point to
> the same box. gitbox.a.o is the new name, git-wip-us.a.o is the old name.

FTR: I thought git-wip-us was redirected to gitbox once all the
git-wip repos had been migrated?
i.e. before the recent git.a.o failure (I thought that was only used
for GitHub RO mirrors?)

> Even if they weren't the same box, if the hash of the tag matches then
> that should provide sufficient reassurance that the tags are identical.

If not, then there are bigger problems...

> Mark
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org