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

Posted to dev@shindig.apache.org by Apache Hudson Server <hu...@hudson.apache.org> on 2010/08/28 03:15:53 UTC

Hudson build became unstable: Shindig #1787

See <https://hudson.apache.org/hudson/job/Shindig/1787/changes>

Re: Hudson build became unstable: Shindig #1787

Posted by Paul Lindner <pl...@linkedin.com>.

I take that back, it is a java 1.5 issue -- reproduced it and will look for
a fix today.

On Sun, Aug 29, 2010 at 11:15 AM, Paul Lindner <pl...@linkedin.com>wrote:

> can't reproduce on a local hudson instance with java 1.5 or 1.6... odd.
>
>
> On Sat, Aug 28, 2010 at 9:53 AM, Henry Saputra <he...@gmail.com>wrote:
>
>> Hi Gagan,
>>
>> Hmmm I also has tried clean pull from trunk and didnt see this test fail
>> when running mvn clean install.
>>
>> I am running the build in MBP with snow leopard.
>>
>> - Henry
>>
>> On Sat, Aug 28, 2010 at 9:24 AM, Gagandeep singh <gagan.goku@gmail.com
>> >wrote:
>>
>> > This is strange. In my local client synced to head without any
>> > modifications, shindig seems to build fine without any breakages (im on
>> > Windoze). Why is the CajaCssSanitizerTest failing on hudson machine ?
>> >
>> >
>> https://hudson.apache.org/hudson/job/Shindig/org.apache.shindig$shindig-gadgets/1787/testReport/org.apache.shindig.gadgets.parse.caja/CajaCssSanitizerTest/
>> > <
>> >
>> https://hudson.apache.org/hudson/job/Shindig/org.apache.shindig$shindig-gadgets/1787/testReport/org.apache.shindig.gadgets.parse.caja/CajaCssSanitizerTest/
>> > >
>> >
>> > On Sat, Aug 28, 2010 at 6:45 AM, Apache Hudson Server <
>> > hudson@hudson.apache.org> wrote:
>> >
>> > > See <https://hudson.apache.org/hudson/job/Shindig/1787/changes>
>> > >
>> > >
>> > >
>> >
>>
>
>
>
> --
> Paul Lindner -- plindner@linkedin.com -- linkedin.com/in/plindner
>



-- 
Paul Lindner -- plindner@linkedin.com -- linkedin.com/in/plindner

Re: Hudson build became unstable: Shindig #1787

Posted by Paul Lindner <pl...@linkedin.com>.

can't reproduce on a local hudson instance with java 1.5 or 1.6... odd.

On Sat, Aug 28, 2010 at 9:53 AM, Henry Saputra <he...@gmail.com>wrote:

> Hi Gagan,
>
> Hmmm I also has tried clean pull from trunk and didnt see this test fail
> when running mvn clean install.
>
> I am running the build in MBP with snow leopard.
>
> - Henry
>
> On Sat, Aug 28, 2010 at 9:24 AM, Gagandeep singh <gagan.goku@gmail.com
> >wrote:
>
> > This is strange. In my local client synced to head without any
> > modifications, shindig seems to build fine without any breakages (im on
> > Windoze). Why is the CajaCssSanitizerTest failing on hudson machine ?
> >
> >
> https://hudson.apache.org/hudson/job/Shindig/org.apache.shindig$shindig-gadgets/1787/testReport/org.apache.shindig.gadgets.parse.caja/CajaCssSanitizerTest/
> > <
> >
> https://hudson.apache.org/hudson/job/Shindig/org.apache.shindig$shindig-gadgets/1787/testReport/org.apache.shindig.gadgets.parse.caja/CajaCssSanitizerTest/
> > >
> >
> > On Sat, Aug 28, 2010 at 6:45 AM, Apache Hudson Server <
> > hudson@hudson.apache.org> wrote:
> >
> > > See <https://hudson.apache.org/hudson/job/Shindig/1787/changes>
> > >
> > >
> > >
> >
>



-- 
Paul Lindner -- plindner@linkedin.com -- linkedin.com/in/plindner

Re: Hudson build became unstable: Shindig #1787

Posted by Henry Saputra <he...@gmail.com>.

Hi Gagan,

Hmmm I also has tried clean pull from trunk and didnt see this test fail
when running mvn clean install.

I am running the build in MBP with snow leopard.

- Henry

On Sat, Aug 28, 2010 at 9:24 AM, Gagandeep singh <ga...@gmail.com>wrote:

> This is strange. In my local client synced to head without any
> modifications, shindig seems to build fine without any breakages (im on
> Windoze). Why is the CajaCssSanitizerTest failing on hudson machine ?
>
> https://hudson.apache.org/hudson/job/Shindig/org.apache.shindig$shindig-gadgets/1787/testReport/org.apache.shindig.gadgets.parse.caja/CajaCssSanitizerTest/
> <
> https://hudson.apache.org/hudson/job/Shindig/org.apache.shindig$shindig-gadgets/1787/testReport/org.apache.shindig.gadgets.parse.caja/CajaCssSanitizerTest/
> >
>
> On Sat, Aug 28, 2010 at 6:45 AM, Apache Hudson Server <
> hudson@hudson.apache.org> wrote:
>
> > See <https://hudson.apache.org/hudson/job/Shindig/1787/changes>
> >
> >
> >
>

Re: Hudson build became unstable: Shindig #1787

Posted by Gagandeep singh <ga...@gmail.com>.

This is strange. In my local client synced to head without any
modifications, shindig seems to build fine without any breakages (im on
Windoze). Why is the CajaCssSanitizerTest failing on hudson machine ?
https://hudson.apache.org/hudson/job/Shindig/org.apache.shindig$shindig-gadgets/1787/testReport/org.apache.shindig.gadgets.parse.caja/CajaCssSanitizerTest/
<https://hudson.apache.org/hudson/job/Shindig/org.apache.shindig$shindig-gadgets/1787/testReport/org.apache.shindig.gadgets.parse.caja/CajaCssSanitizerTest/>

On Sat, Aug 28, 2010 at 6:45 AM, Apache Hudson Server <
hudson@hudson.apache.org> wrote:

> See <https://hudson.apache.org/hudson/job/Shindig/1787/changes>
>
>
>

Re: Hudson build is back to normal : Shindig #1793

Posted by Mat Mannion <M....@warwick.ac.uk>.

A SortedMap approach wouldn't preserve insertion order, but would
instead re-order parameters on the fly with the comparator provided. A
LinkedHashMap (or LinkedMultimap, which seems like a better solution
as you suggest) preserves insertion order, so this would be the
correct way to go if we're looking to preserve the parameter order.

Mat

On 31 August 2010 10:23, Gagandeep singh <ga...@gmail.com> wrote:
> I have a doubt. A fellow colleague tells me that we should retain the order
> in which the keys are inserted into the query params map, because some ppl
> do funny stuff like their own begin and end parameters and mark useful
> parameters in between.
>
> I was wondering if we should go with a SortedMap approach for query params
> instead. I mean do we really need to bother about query param P1 added
> before query param P2 ? Also, tomorrow if some1 refactores some code, the
> tests should not fail because the order of insertion has changed.
>
> As a side note, i think Google uses LinkedListMultimap for query params.
> Isn't that better than LinkedHashMap<String, List<String>> ?
>
> Thoughts ?
>
>
>
> On Mon, Aug 30, 2010 at 1:31 PM, Gagandeep singh <ga...@gmail.com>wrote:
>
>> Nice debugging.
>>
>>
>> On Mon, Aug 30, 2010 at 7:42 AM, Paul Lindner <pl...@linkedin.com>wrote:
>>
>>> The test output showed that the problem was incorrectly ordered query
>>> params.  I validated that the problem occurs only on Java 1.5.  Therefore
>>> it
>>> had to be a HashMap iteration problem somewhere.  So I poked around
>>> UriBuilder and found one remaining use of HashMap and replaced it with
>>> LinkedHashMap.
>>>
>>> On Sun, Aug 29, 2010 at 6:54 PM, Gagandeep singh <gagan.goku@gmail.com
>>> >wrote:
>>>
>>> > Could you also tell us how you found and fixed the problem ?
>>> >
>>> > On Mon, Aug 30, 2010 at 7:24 AM, Gagandeep singh <gagan.goku@gmail.com
>>> > >wrote:
>>> >
>>> > > Thanks for fixing this Paul.
>>> > >
>>> > >
>>> > > On Mon, Aug 30, 2010 at 2:46 AM, Apache Hudson Server <
>>> > > hudson@hudson.apache.org> wrote:
>>> > >
>>> > >> See <https://hudson.apache.org/hudson/job/Shindig/1793/>
>>> > >>
>>> > >>
>>> > >>
>>> > >
>>> >
>>>
>>>
>>>
>>> --
>>> Paul Lindner -- plindner@linkedin.com -- linkedin.com/in/plindner
>>>
>>
>>
>



-- 
Mat Mannion
Web Developer
IT Services
University of Warwick
Coventry
CV4 7AL

Tel: 024 765 74433
Email: M.Mannion@warwick.ac.uk

Re: Hudson build is back to normal : Shindig #1793

Posted by Paul Lindner <pl...@linkedin.com>.

Our current UriBuilder has nice reflective properties like this:

String foo = "http://foo.com/a?z=1&a=2";

UriBuilder.parse(foo).toString().equals(foo);

I think this behavior is worth preserving

I would like to see a Multimap implementation of query params.  Note that
ImmutableListMultimaps also retain their insertion order.


On Tue, Aug 31, 2010 at 2:23 AM, Gagandeep singh <ga...@gmail.com>wrote:

> I have a doubt. A fellow colleague tells me that we should retain the order
> in which the keys are inserted into the query params map, because some ppl
> do funny stuff like their own begin and end parameters and mark useful
> parameters in between.
>
> I was wondering if we should go with a SortedMap approach for query params
> instead. I mean do we really need to bother about query param P1 added
> before query param P2 ? Also, tomorrow if some1 refactores some code, the
> tests should not fail because the order of insertion has changed.
>
> As a side note, i think Google uses LinkedListMultimap for query params.
> Isn't that better than LinkedHashMap<String, List<String>> ?
>
> Thoughts ?
>
>
>
> On Mon, Aug 30, 2010 at 1:31 PM, Gagandeep singh <gagan.goku@gmail.com
> >wrote:
>
> > Nice debugging.
> >
> >
> > On Mon, Aug 30, 2010 at 7:42 AM, Paul Lindner <plindner@linkedin.com
> >wrote:
> >
> >> The test output showed that the problem was incorrectly ordered query
> >> params.  I validated that the problem occurs only on Java 1.5.
>  Therefore
> >> it
> >> had to be a HashMap iteration problem somewhere.  So I poked around
> >> UriBuilder and found one remaining use of HashMap and replaced it with
> >> LinkedHashMap.
> >>
> >> On Sun, Aug 29, 2010 at 6:54 PM, Gagandeep singh <gagan.goku@gmail.com
> >> >wrote:
> >>
> >> > Could you also tell us how you found and fixed the problem ?
> >> >
> >> > On Mon, Aug 30, 2010 at 7:24 AM, Gagandeep singh <
> gagan.goku@gmail.com
> >> > >wrote:
> >> >
> >> > > Thanks for fixing this Paul.
> >> > >
> >> > >
> >> > > On Mon, Aug 30, 2010 at 2:46 AM, Apache Hudson Server <
> >> > > hudson@hudson.apache.org> wrote:
> >> > >
> >> > >> See <https://hudson.apache.org/hudson/job/Shindig/1793/>
> >> > >>
> >> > >>
> >> > >>
> >> > >
> >> >
> >>
> >>
> >>
> >> --
> >> Paul Lindner -- plindner@linkedin.com -- linkedin.com/in/plindner
> >>
> >
> >
>



-- 
Paul Lindner -- plindner@linkedin.com -- linkedin.com/in/plindner

Re: Hudson build is back to normal : Shindig #1793

Posted by Gagandeep singh <ga...@gmail.com>.

I have a doubt. A fellow colleague tells me that we should retain the order
in which the keys are inserted into the query params map, because some ppl
do funny stuff like their own begin and end parameters and mark useful
parameters in between.

I was wondering if we should go with a SortedMap approach for query params
instead. I mean do we really need to bother about query param P1 added
before query param P2 ? Also, tomorrow if some1 refactores some code, the
tests should not fail because the order of insertion has changed.

As a side note, i think Google uses LinkedListMultimap for query params.
Isn't that better than LinkedHashMap<String, List<String>> ?

Thoughts ?

On Mon, Aug 30, 2010 at 1:31 PM, Gagandeep singh <ga...@gmail.com>wrote:

> Nice debugging.
>
>
> On Mon, Aug 30, 2010 at 7:42 AM, Paul Lindner <pl...@linkedin.com>wrote:
>
>> The test output showed that the problem was incorrectly ordered query
>> params.  I validated that the problem occurs only on Java 1.5.  Therefore
>> it
>> had to be a HashMap iteration problem somewhere.  So I poked around
>> UriBuilder and found one remaining use of HashMap and replaced it with
>> LinkedHashMap.
>>
>> On Sun, Aug 29, 2010 at 6:54 PM, Gagandeep singh <gagan.goku@gmail.com
>> >wrote:
>>
>> > Could you also tell us how you found and fixed the problem ?
>> >
>> > On Mon, Aug 30, 2010 at 7:24 AM, Gagandeep singh <gagan.goku@gmail.com
>> > >wrote:
>> >
>> > > Thanks for fixing this Paul.
>> > >
>> > >
>> > > On Mon, Aug 30, 2010 at 2:46 AM, Apache Hudson Server <
>> > > hudson@hudson.apache.org> wrote:
>> > >
>> > >> See <https://hudson.apache.org/hudson/job/Shindig/1793/>
>> > >>
>> > >>
>> > >>
>> > >
>> >
>>
>>
>>
>> --
>> Paul Lindner -- plindner@linkedin.com -- linkedin.com/in/plindner
>>
>
>

Re: Hudson build is back to normal : Shindig #1793

Posted by Gagandeep singh <ga...@gmail.com>.

Nice debugging.

On Mon, Aug 30, 2010 at 7:42 AM, Paul Lindner <pl...@linkedin.com> wrote:

> The test output showed that the problem was incorrectly ordered query
> params.  I validated that the problem occurs only on Java 1.5.  Therefore
> it
> had to be a HashMap iteration problem somewhere.  So I poked around
> UriBuilder and found one remaining use of HashMap and replaced it with
> LinkedHashMap.
>
> On Sun, Aug 29, 2010 at 6:54 PM, Gagandeep singh <gagan.goku@gmail.com
> >wrote:
>
> > Could you also tell us how you found and fixed the problem ?
> >
> > On Mon, Aug 30, 2010 at 7:24 AM, Gagandeep singh <gagan.goku@gmail.com
> > >wrote:
> >
> > > Thanks for fixing this Paul.
> > >
> > >
> > > On Mon, Aug 30, 2010 at 2:46 AM, Apache Hudson Server <
> > > hudson@hudson.apache.org> wrote:
> > >
> > >> See <https://hudson.apache.org/hudson/job/Shindig/1793/>
> > >>
> > >>
> > >>
> > >
> >
>
>
>
> --
> Paul Lindner -- plindner@linkedin.com -- linkedin.com/in/plindner
>

Re: Hudson build is back to normal : Shindig #1793

Posted by Paul Lindner <pl...@linkedin.com>.

The test output showed that the problem was incorrectly ordered query
params.  I validated that the problem occurs only on Java 1.5.  Therefore it
had to be a HashMap iteration problem somewhere.  So I poked around
UriBuilder and found one remaining use of HashMap and replaced it with
LinkedHashMap.

On Sun, Aug 29, 2010 at 6:54 PM, Gagandeep singh <ga...@gmail.com>wrote:

> Could you also tell us how you found and fixed the problem ?
>
> On Mon, Aug 30, 2010 at 7:24 AM, Gagandeep singh <gagan.goku@gmail.com
> >wrote:
>
> > Thanks for fixing this Paul.
> >
> >
> > On Mon, Aug 30, 2010 at 2:46 AM, Apache Hudson Server <
> > hudson@hudson.apache.org> wrote:
> >
> >> See <https://hudson.apache.org/hudson/job/Shindig/1793/>
> >>
> >>
> >>
> >
>

-- 
Paul Lindner -- plindner@linkedin.com -- linkedin.com/in/plindner

Re: Hudson build is back to normal : Shindig #1793

Posted by Gagandeep singh <ga...@gmail.com>.

Could you also tell us how you found and fixed the problem ?

On Mon, Aug 30, 2010 at 7:24 AM, Gagandeep singh <ga...@gmail.com>wrote:

> Thanks for fixing this Paul.
>
>
> On Mon, Aug 30, 2010 at 2:46 AM, Apache Hudson Server <
> hudson@hudson.apache.org> wrote:
>
>> See <https://hudson.apache.org/hudson/job/Shindig/1793/>
>>
>>
>>
>