manifoldcf-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Karl Wright <daddy...@gmail.com>
Subject Re: IO exception during indexing: null
Date Mon, 11 Mar 2013 14:05:59 GMT
We already use multipart entity.  But we can't use bytearraybody
because it means the entire document has to be sucked into memory.
That's not acceptable.

I will post again on the HttpClient list, because they did lead me to
believe that using expect-continue would do the right thing.

Karl


On Mon, Mar 11, 2013 at 9:58 AM, Erlend Garåsen <e.f.garasen@usit.uio.no> wrote:
>
> Did an svn up, built and deployed on our prod server, but I'm still getting
> the same error.
>
> I just saw the following comment from you in the HttpClient forum: "Problem
> is that ManifoldCF output connectors get an inputstream handed to them, not
> a file.".
>
> I think I had a similar issue when I upgraded from a previous release of
> HttpClient. By using the HttpPost class in HttpClient, I did something like
> this with a dummy file named used (for multi-form post data):
>
> MultipartEntity multipartEntity = new MultipartEntity();
> multipartEntity.addPart("myfile", new ByteArrayBody(content, contentType,
> "myfile"));
> httpPost.setEntity(multipartEntity);
>
> I'm afraid that we need to use the new HttpEntity implementations.
>
> Erlend
>
>
> On 10.03.13 20.43, Erlend Garåsen wrote:
>>
>> Thanks Karl!
>>
>> I'm sitting on the train this evening, so I will try and report tomorrow.
>>
>> Erlend.
>>
>> Sent from my iPad
>>
>> On 8. mars 2013, at 21:06, Karl Wright <daddywri@gmail.com> wrote:
>>
>>> There's some indication that turning on expect-continue will resolve
>>> this issue.  So that is what I did.  Please synch up trunk and see if
>>> you still get the error.
>>>
>>> Thanks!
>>> Karl
>>>
>>>
>>> On Fri, Mar 8, 2013 at 9:14 AM, Karl Wright <daddywri@gmail.com> wrote:
>>>>
>>>> OK, there IS a problem here, and it is a significant one.  See
>>>> CONNECTORS-661 for details.
>>>>
>>>> I've asked for a workaround proposal on the httpcomponents list; it
>>>> seems to me that pre-emptive auth headers should be possible in
>>>> HttpClient.  In the worst case, we can construct our own header, but
>>>> first let's see what they say.
>>>>
>>>> Karl
>>>>
>>>>
>>>> On Fri, Mar 8, 2013 at 8:20 AM, Erlend Garåsen <e.f.garasen@usit.uio.no>
>>>> wrote:
>>>>>
>>>>> On 08.03.13 13.37, Karl Wright wrote:
>>>>>>
>>>>>>
>>>>>> Actually, let's just cut to the chase, and enable full wire and
>>>>>> context
>>>>>> logging:
>>>>>>
>>>>>> log4j.logger.org.apache.http=DEBUG
>>>>>>
>>>>>> That will let us see what is happening for sure.
>>>>>
>>>>>
>>>>>
>>>>> Working on it.
>>>>>
>>>>> I will send the log directly to you, Karl, since this list may cut the
>>>>> log
>>>>> file.
>>>>>
>>>>>
>>>>> Erlend
>>>>>
>>>>> --
>>>>> Erlend Garåsen
>>>>> Center for Information Technology Services
>>>>> University of Oslo
>>>>> P.O. Box 1086 Blindern, N-0317 OSLO, Norway
>>>>> Ph: (+47) 22840193, Fax: (+47) 22852970, Mobile: (+47) 91380968, VIP:
>>>>> 31050
>
>
>
> --
> Erlend Garåsen
> Center for Information Technology Services
> University of Oslo
> P.O. Box 1086 Blindern, N-0317 OSLO, Norway
> Ph: (+47) 22840193, Fax: (+47) 22852970, Mobile: (+47) 91380968, VIP: 31050

Mime
View raw message