[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [ecf-dev] filetransfer tests
- From: Scott Lewis <slewis@xxxxxxxxxxxxx>
- Date: Sat, 12 May 2012 09:16:28 -0700
- Delivered-to: firstname.lastname@example.org
- User-agent: Mozilla/5.0 (Windows NT 5.1; rv:12.0) Gecko/20120428 Thunderbird/12.0.1
I realized that the plugin Markus listed was from Orbit (and not one of
Yes...the target platform would have to contain the httpcomponents (4.0)
if my hypothesis was true.
So...we still have a bit of a mystery (as to why the tests are not
running the same locally...for me...as they are on in the build).
BTW...by not running the same, I mean:
1) I see the httpclient tracing...since the system parameters for
httpclient tracing are on...and I don't see them in the build output
2) The 7 test failures in the auto build...all having to do with
cancel...do not happen for me.
On 5/12/2012 9:08 AM, Scott Lewis wrote:
On 5/12/2012 1:25 AM, Markus Alexander Kuppe wrote:
On 05/11/2012 07:57 PM, Scott Lewis wrote:
I suspect that somehow the new httpclient4 provider was being used
rather than the httpclient 3.1 provider (I initially committed the
httpclient4 code with higher priority than the httpclient 3.1
provider). That is, it seems as if the httpclient4 provider is being
used...when it should be using httpclient 3.1 (and it certainly does
the httpclient 3.1 provider in my local running of the launch config).
The target platform only contains
org.apache.commons.httpclient_3.1.0.v201012070820.jar. Wouldn't it
contain 4.x if your assumption is true?
I suppose...if the tests use the target platform as the source for
plugins (instead of target platform + workspace).
But isn't that (v201012...) an old version? There has been a
significant bug fix since then...perhaps that's why some of the cancel
tests are failing (I fixed a cancel bug with the httpclient provider
during the winter).
ecf-dev mailing list