You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@zookeeper.apache.org by Patrick Hunt <ph...@apache.org> on 2009/02/10 23:59:59 UTC
ACL issue in C client (re 3.1 candidate)
This just came in:
https://issues.apache.org/jira/browse/ZOOKEEPER-309
The patch seems to fix the issue, however it's missing tests -- actually
the cppunit tests aren't testing any acl functionality, this (missing
tests) is a major oversight we need to address asap (post 3.1)!
There is a workaround for this issue - use the async version of the
method. Seems that no one ever used this function before (or never
reported it at least) because it's an obvious problem and will always fail.
I'd like to hold off on fixing this until post 3.1 (3.1.0 or 3.2
whichever comes first) where we can also address the testing issue for
acls. There is a workaround -- Chris you ok with that? What do the rest
of you think (vote).
Patrick
Re: ACL issue in C client (re 3.1 candidate)
Posted by Chris Darroch <ch...@pearsoncmg.com>.
Patrick Hunt wrote:
> There is a workaround for this issue - use the async version of the
> method. Seems that no one ever used this function before (or never
> reported it at least) because it's an obvious problem and will always fail.
>
> I'd like to hold off on fixing this until post 3.1 (3.1.0 or 3.2
> whichever comes first) where we can also address the testing issue for
> acls. There is a workaround -- Chris you ok with that? What do the rest
> of you think (vote).
No question some tests would be good; that said, I'll personally
just continue to use a simple patch because it's a lot easier than
copy-and-pasting the code required to wrap the async version into a
sync one, and my particular current needs require a sync interface.
If there's a reasonable chance of just sliding that one-liner into
the next release, cool, but I can live without and just recommend
other people patch the source (or do the workaround, I suppose).
Chris.
--
GPG Key ID: 366A375B
GPG Key Fingerprint: 485E 5041 17E1 E2BB C263 E4DE C8E3 FA36 366A 375B
Re: ACL issue in C client (re 3.1 candidate)
Posted by Patrick Hunt <ph...@apache.org>.
Sounds like a plan. I've updated the JIRA (also created a 3.1.1 version
in JIRA slated for 4 weeks out -- we can always pull it in if needed)
Patrick
Mahadev Konar wrote:
> +1 to ben's suggestion. We can go ahead with 3.1.0 and make a 3.1.1 (given
> that it wasn't working in the prev releases as well)...
>
> mahadev
>
>
> On 2/10/09 6:31 PM, "Benjamin Reed" <br...@yahoo-inc.com> wrote:
>
>> Chris's patch is very simple and it won't be hard to put some tests in. can we
>> plan a bug release in the next week or two to catch this and any other bugs
>> that might emerge?
>>
>> ben
>> ________________________________________
>> From: Patrick Hunt [phunt@apache.org]
>> Sent: Tuesday, February 10, 2009 2:59 PM
>> To: zookeeper-dev@hadoop.apache.org
>> Cc: Chris Darroch
>> Subject: ACL issue in C client (re 3.1 candidate)
>>
>> This just came in:
>> https://issues.apache.org/jira/browse/ZOOKEEPER-309
>>
>> The patch seems to fix the issue, however it's missing tests -- actually
>> the cppunit tests aren't testing any acl functionality, this (missing
>> tests) is a major oversight we need to address asap (post 3.1)!
>>
>> There is a workaround for this issue - use the async version of the
>> method. Seems that no one ever used this function before (or never
>> reported it at least) because it's an obvious problem and will always fail.
>>
>> I'd like to hold off on fixing this until post 3.1 (3.1.0 or 3.2
>> whichever comes first) where we can also address the testing issue for
>> acls. There is a workaround -- Chris you ok with that? What do the rest
>> of you think (vote).
>>
>> Patrick
>
Re: ACL issue in C client (re 3.1 candidate)
Posted by Mahadev Konar <ma...@yahoo-inc.com>.
+1 to ben's suggestion. We can go ahead with 3.1.0 and make a 3.1.1 (given
that it wasn't working in the prev releases as well)...
mahadev
On 2/10/09 6:31 PM, "Benjamin Reed" <br...@yahoo-inc.com> wrote:
> Chris's patch is very simple and it won't be hard to put some tests in. can we
> plan a bug release in the next week or two to catch this and any other bugs
> that might emerge?
>
> ben
> ________________________________________
> From: Patrick Hunt [phunt@apache.org]
> Sent: Tuesday, February 10, 2009 2:59 PM
> To: zookeeper-dev@hadoop.apache.org
> Cc: Chris Darroch
> Subject: ACL issue in C client (re 3.1 candidate)
>
> This just came in:
> https://issues.apache.org/jira/browse/ZOOKEEPER-309
>
> The patch seems to fix the issue, however it's missing tests -- actually
> the cppunit tests aren't testing any acl functionality, this (missing
> tests) is a major oversight we need to address asap (post 3.1)!
>
> There is a workaround for this issue - use the async version of the
> method. Seems that no one ever used this function before (or never
> reported it at least) because it's an obvious problem and will always fail.
>
> I'd like to hold off on fixing this until post 3.1 (3.1.0 or 3.2
> whichever comes first) where we can also address the testing issue for
> acls. There is a workaround -- Chris you ok with that? What do the rest
> of you think (vote).
>
> Patrick
RE: ACL issue in C client (re 3.1 candidate)
Posted by Benjamin Reed <br...@yahoo-inc.com>.
Chris's patch is very simple and it won't be hard to put some tests in. can we plan a bug release in the next week or two to catch this and any other bugs that might emerge?
ben
________________________________________
From: Patrick Hunt [phunt@apache.org]
Sent: Tuesday, February 10, 2009 2:59 PM
To: zookeeper-dev@hadoop.apache.org
Cc: Chris Darroch
Subject: ACL issue in C client (re 3.1 candidate)
This just came in:
https://issues.apache.org/jira/browse/ZOOKEEPER-309
The patch seems to fix the issue, however it's missing tests -- actually
the cppunit tests aren't testing any acl functionality, this (missing
tests) is a major oversight we need to address asap (post 3.1)!
There is a workaround for this issue - use the async version of the
method. Seems that no one ever used this function before (or never
reported it at least) because it's an obvious problem and will always fail.
I'd like to hold off on fixing this until post 3.1 (3.1.0 or 3.2
whichever comes first) where we can also address the testing issue for
acls. There is a workaround -- Chris you ok with that? What do the rest
of you think (vote).
Patrick