Excellent Seb. Will do. Thank you for taking action on this one. Sometimes we (I) cannot see the wood for the trees... I can now :)
On Thu, Jun 19, 2014 at 6:47 PM, Sebastian Nagel <[email protected] > wrote: > Hi Lewis, > > a patch is ready, on my machine all tests pass now. > Currently, I experience problems with Jira: > feel free to open and resolve the issue. > > Cheers, > Sebastian > > > On 06/19/2014 07:58 PM, Lewis John Mcgibbney wrote: > > Hi Seb, > > > > On Thu, Jun 19, 2014 at 1:46 PM, <[email protected] > > <mailto:[email protected]>> wrote: > > > > > > it seems to be related to NUTCH-1714: > > WebPage-owned maps (metadata, headers, etc.) are not > > initialized any more in the constructor. > > > > > > Yeah, I thought we had addressed this in the patch! We should create > empty data structures for > > MapWrapper, ArrayWrapper, etc to work with rather than a null field. > > > > This causes also other tests to fail. > > > > The solution would be to replace > > WebPage page = new WebPage(); > > by > > WebPage page = WebPage.newBuilder().build(); > > in every test where a WebPage object is needed. > > > > Right? > > > > > > Yes, this is the best way to use new Objects. Unlike constructors, > builders will automatically set > > any default values specified in the WebPage/Host (or any others JSON) > schema. Additionally, builders > > validate the data as it set, whereas objects constructed directly will > not cause an error until the > > object is serialized. However, using constructors directly generally > offers better performance, as > > builders create a copy of the datastructure before it is written. > > > > > > > > I'll open a Jira and try to provide a patch. > > > > That would be excellent. I thought that this was related to something > else entirely... e.g. a dodgy > > commit. > > > > Hopefully this is the root of the problem. > > -- *Lewis*
