Il 26.02.2015 17.22, Brett Ryan ha scritto:
I thought Oleg pointed this out already.
If you build up a byte array for your request that byte array must fully occupy
local memory, streaming this will both send data to the host as its constructed
directly from the marshaller which is not only more efficient but will rule out
any issue resulted by an incorrectly formed in memory representation of your
payload. Who knows, maybe you have a byte array overflow thats not apparent,
which is kinda what I'm suspecting is going on here since your issue only
arrises once the payload reaches a certain size.
I followed Oleg suggestions.
First I logged both wire an header.
Inside the log there's nothing abnormal, the posted xml is not
truncated, tomcat keep reply 400.
Then I changed from StringEntity to ByteArrayEntity, and arranged
content-type with encoding utf-8. This night the customer will change
the jar, tomorrow morning I'll see.
My thought do was some array overflow, but I think that it's not local,
I mean not inside HTTPClient. In other words I don't know whether
there's some buffers in HTTPClient or not. Further more, I don't know
how to adjust ipotetical buffers.
Sent from my iPhone
On 26 Feb 2015, at 23:28, Alessandro Manzoni <manzoni.alessand...@gmail.com>
wrote:
What I would like to understand, is why an InputStreamEntity would be better
than a StringEntity.
---------------------------------------------------------------------
To unsubscribe, e-mail: httpclient-users-unsubscr...@hc.apache.org
For additional commands, e-mail: httpclient-users-h...@hc.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: httpclient-users-unsubscr...@hc.apache.org
For additional commands, e-mail: httpclient-users-h...@hc.apache.org
