Hi Rama,
It's very weird :(
can you pls. give us the testcase to reproduce this?
Thanks.
WBR,
Alexey.
On 11/22/2011 03:05 PM, Rama wrote:
> Hi Alexey,
>
> i run into a big ussue with 1.9.41, making me impossible to test it :(
>
> Seems that there was some change on how httpReq.getReader(); is
> handled (or on how headers is handled)
>
> i got an exception "invalid character encoding 'a' " while doing
> getReader() if the client is flash player.
>
> with 1.9.36 there isn't such problem.
>
> Also, on java.net the pages of gws are outdated (javadoc+source code +
> repos with jars) and i wasn't able to crawl on
> source code to get track a bit better the issue.
>
> imho, there is some default character encoding handling (maybe when
> missing on request) and this default is 'a' that sound quite strange
> to me.
>
> Ciao,
> R.
>
> 2011/11/21 Oleksiy Stashok<oleksiy.stashok_at_oracle.com>:
>> Hi Rama,
>>
>> 1.9.41 has been released, artifacts could be found on the maven central
>> http://repo1.maven.org/
>>
>> WBR,
>> Alexey.
>>
>> On 11/18/2011 02:10 PM, Rama wrote:
>>> Sure, there is a .jar for that?
>>>
>>> warm regards
>>>
>>>
>>> 2011/11/14 Oleksiy Stashok<oleksiy.stashok_at_oracle.com>:
>>>> Hi Rama,
>>>>
>>>> can you pls. check the latest 1.9.41-SNAPSHOT?
>>>>
>>>> Thanks.
>>>>
>>>> WBR,
>>>> Alexey.
>>>>
>>>> On 11/14/2011 02:05 PM, Oleksiy Stashok wrote:
>>>>> Hi Rama,
>>>>>
>>>>> it's a bug, I mean (1), the (2), IMO, comes from (1).
>>>>> Can I ask you to file an issue?
>>>>>
>>>>> I will fix it asap.
>>>>>
>>>>> Thank you.
>>>>>
>>>>> WBR,
>>>>> Alexey.
>>>>>
>>>>> On 11/14/2011 10:41 AM, Rama wrote:
>>>>>> Hi there, i have some question regarding async adapter in gws 1.9.X
>>>>>> series.
>>>>>>
>>>>>>
>>>>>> let's suppose that i'll chain a AsyncFilter to my webserver and i
>>>>>> suspend the request as follow (pseudocode)
>>>>>>
>>>>>>
>>>>>> public Result doFilter(AsyncExecutor asyncExecutor) {
>>>>>> queue.add(asyncExecutor,System.currenttimemillis());
>>>>>> return Result.INTERRUPT;
>>>>>> }
>>>>>>
>>>>>>
>>>>>> later i'll iterate on the queue and i'll check if the timeout is>
>>>>>> than 30.000 and if yes, i'll execute the request
>>>>>>
>>>>>> if (systemcurrent> eta+30000) {
>>>>>> a.execute();
>>>>>> a.getAsyncTask().doTask();
>>>>>> }
>>>>>>
>>>>>>
>>>>>> basically i expect that all the requests get delayed by 30sec, but
>>>>>> this doesn't happen :(
>>>>>> on safari for istance, request get delayed by 20sec (no idea why)
>>>>>>
>>>>>>
>>>>>> Also, if i set
>>>>>> "this.getSelectorThread().setKeepAliveTimeoutInSeconds(1);"
>>>>>> i got connection closed
>>>>>> on client side.
>>>>>>
>>>>>> Why keepalive is related to the async stuff?
>>>>>> If yes, what's happen when i reach the max_keepalive_num_request with
>>>>>> async?!?
>>>>>>
>>>>>> honestly i don't get why keepalivetimeout influence async stuff.
>>>>>>
>>>>>>
>>>>>> So basically, summarizing the questions
>>>>>>
>>>>>> 1) why keepalive_timeout influence async request?
>>>>>> 2) why async request finish BEFORE what they should? there is any
>>>>>> parsing related to browser header that i don't get?
>>>>>>
>>>>>>
>>>>>> tnx!
>>