> -----Original Message----- > From: Tim Williams [mailto:william...@gmail.com] > Sent: Friday, 2 April 2010 11:13 PM > To: dev@forrest.apache.org > Subject: Re: forrest-sample-2 FAILED and build test failed: Could not > resolve locationmap location > > On Fri, Apr 2, 2010 at 8:22 AM, Tim Williams <william...@gmail.com> > wrote: > > On Thu, Apr 1, 2010 at 3:27 AM, David Crossley <cross...@apache.org> > wrote: > >>> Automated build for forrest-sample-2 FAILED > >>> [java] > >>> [java] > >>> [java] X [0] linkmap.html > BROKEN: Could not resolve locationmap location. > >> > >> > >> I get that same error doing 'build.sh test' locally. > >> > >> The build of this Dispatcher sample were okay on the > >> zone server before today. > > > > I'm getting the same thing. It looks like a mounted locationmap > can't > > be resolved but it appears to be dispatcher-related code. I don't > > know that code at all but a quick look seems like the resolver field > > in the child class (RecursiveDirectoryTraversalAction) is hiding the > > intended resolver in the parent class (AbstractTraversal). Can you > > comment out the child class' resolver field and see if that helps? > It > > seems to get me past that one but then introduces tons of other > > "Invalid byte 1 of 1-byte UTF-8 sequence" errors. The timing seems > > right too, it would have been introduced with r929463. > > I went ahead and applied that because I feel sure that's the right > thing to do. I'm now getting errors like: > > [Fatal Error] :6:25: Invalid byte 1 of 1-byte UTF-8 sequence. > X [0] linkmap.pdf BROKEN: Invalid > byte 1 of 1-byte UTF-8 sequence. > > ... for every pdf file. I don't have time to look into that one right > now though, can someone see if they get the same?
Yes, I am getting it too, but here we are again, a Windows only bug it seems. I dont get this on any linux box - which is why our zone doesnt complain. Only thing I could come up with via Go-ogle was this thread featuring some known names: http://www.mail-archive.com/j-...@xerces.apache.org/msg00030.html Consensus there seems to be that Windows default to latin rather than utf8 and that OutputStreamWriter should be used in place of FileWriter. Why this is a recent thing I dont know, havent had time to look much further, but I do know I didnt get these errors when recently testing the Java 1.5/1.6 scenarios before the vote. I also went back before Thorstens recent commits and the errors remain - not that I thought for one minute they were related, but tested it anyway. Gav... > Thanks, > --tim
