On 2010-11-09, Sander Temme wrote:
> I have updated http://wiki.apache.org/gump/OsxGumpConfig to document
> the installation on Adam. Let me know if I missed something obvious.
Looks good at first glance but I'll read it over once again later.
> - if not self._checkWithDashVersion('NAnt', "no NAnt builds"):
> - if not self._checkWithDashVersion('NAnt.exe', "no NAnt
> + if not self._checkWithDashVersion('NAnt', "no NAnt builds",
> + if not self._checkWithDashVersion('NAnt.exe', "no NAnt
> builds", '-help'):
0.85 is pretty old (0.86 is the latest released version with 0.90 being
> What we're running is the live branch: what is our backport/merge
> policy? Can I commit to this, or should I commit to trunk first and
> then bring it across?
So far I've comitted to trunk, waited for the FreeBSD jail to complete a
run just to be sure I didn't break anything and then merged the change
to live. I don't care to which branch you apply the patch as long as it
becomes part of both live and trunk.
> I also found this necessary to run:
> -from rdflib.URIRef import URIRef
> -from rdflib.Literal import Literal
> -from rdflib.BNode import BNode
> -from rdflib.Namespace import Namespace
> -from rdflib.constants import TYPE
> +from rdflib import URIRef
> +from rdflib import Literal
> +from rdflib import BNode
> +from rdflib import Namespace
> +from rdflib import TYPE
> I don't know enough Python to know (and don't have the cycles to find
> out) what's going on here.
Me neither. It doesn't cause problems on the other systems (but there
is some whining about the RDFDescriber when Gump starts). I'd say
commit it to trunk and we'll see what the jail says.
> Any thoughts on the projects that are currently failing?
Strangely it says NAnt hasn't been found. I'll put it on my TODO list
to investigate. In my environment NUnit doesn't build after the first
updates because Bazaar works not the same way the updater expects it to
- another TODO item that will stay low priority until the main
installation needs Bazaar support.
commons-lang3 has test failures, it is not too uncommon for tests to
only pass on mainstream Java platforms (we have some test projects that
only fail in the FreeBSD jail but pass on the Ubuntu box). I don't know
whether the Mac is a developer platform for commons-lang. I'll add a
report element to the descriptor so we can see the reasons for the
You need to build args4j manually without "--offline" once. Just do a
"maven jar" inside /Users/Shared/gump/workspace/args4j/args4j.
Certain dependencies are installed into Maven 1.x's local repository and
not via the Gump descriptor since otherwise all Maven 1.x projects would
effectively need to depend on Velocity and Jelly (which itself is built
via Maven 1.x).
To unsubscribe, e-mail: general-unsubscr...@gump.apache.org
For additional commands, e-mail: general-h...@gump.apache.org