On Tue, Apr 26, 2011 at 10:38 PM, Markus Jelsma <[email protected]> wrote: > Hi, > > Of course, but since the agent.* params are mandatory (fetcher will abort when > not specified) we could then add to the error message that these params (like > all) must be set in nutch-site. New users would then keep using nutch-site, at > least that's the idea ;) > > I think that if nutch-default is set to read-only, users that try to modify > will indeed immediately change to write permission and continue to use the > wrong config. > > Cheers, > >> Hi Markus >> >> Any param overridden by the users should be in nutch-site.xml, not just >> http.agent, so why make an exception for it? Moreover that will not >> necessarily prevent people from using nutch-default.xml >> >> Maybe we could set nutch-default to readonly? Could be changed by the user >> but this might nudge them in the right direction >> >> Julien >> >> On 26 April 2011 16:55, Markus Jelsma <[email protected]> wrote: >> > Bump. Thoughts? >> > >> > On Thursday 14 April 2011 12:49:54 Markus Jelsma wrote: >> > > Hi guys, >> > > >> > > Maybe a last convenience would be to precopy the mandatory http.agent >> > > properties to nutch-site. This would, in my opinion, encourage users >> > > not >> > >> > to >> > >> > > set the properties in nutch-default but where it should, in nutch-site. >> > > Thoughts? >> > > >> > > Cheers, >> > >> > -- >> > Markus Jelsma - CTO - Openindex >> > http://www.linkedin.com/in/markus17 >> > 050-8536620 / 06-50258350 >
I would suggest that these properties are set to sensible values in 'conf/nutch-default.xml' itself. I have found it inconvenient to override these properties every time I have installed Nutch. IMHO it would be good to have a working configuration available with the source code and distribution. Regards, Susam Pal
