OK - I think we can do without the definefunctions attribute though.
I'll just make the implicit script class a FunctionSet behind the scenes
and remove the constraint that you have to have a ScriptMain.
What about the namespace?
Jarek
I think add the namespace attribute but make it non-required. Just the
global namespace if its present. Since none of the built-in functions
reside in the global namespace that shouldn't be a problem.
Ian
OK - I think we can do without the definefunctions attribute though.
I'll just make the
I think add the namespace attribute but make it non-required. Just the
global namespace if its present. Since none of the built-in functions
reside in the global namespace that shouldn't be a problem.
I'm afraid there's no global namespace ;-( And parser does rely on the fact
that every
Thanks James - but it looks like we're going to see about creating a
general task outputs framework using in-memory xml representations. I'd
like to see where that discussion goes before committing this patch.
Ian
James C. Papp wrote:
Ian, here is a new version that adds attribute
Or, it could default to something like globals (globals::my-function()).
Anyway this expression thing is very good stuff... Thanks.
Rodrigo
- Original Message -
From: Ian MacLean [EMAIL PROTECTED]
To: Jaroslaw Kowalski [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
Sent: Thursday, January 08,
Hi.
I was in the process of doing some testing for Clayton on the cvs
tasks. I donwloaded the latest nightly build (2004-01-04) and the
latest nantcontrib (dated Dec 26).
When I compile nant using the binaries that it comes with (my old
version of nant did not work since it lacked EE), I get
Hi all,
I successfully include XPathObjectNavigator from Bamboo.Prevalence.
Script like this:
exec program=foo.exe commandline=1 2 3 failonerror=false
result taskoutput=/ExecTask/ResultCode property=resultcode/
/exec
echo message=${resultcode}/
now correctly writes:
[exec]
Hmm, that's no good. In my world it's better with no version at all then
wrong version.
It's giving me a lot of headache as I can't know, or prove what version I'm
using. Nor can
I distribute this to my team as I can not later on know weather or not they
have upgraded
or not.
It ought to also give
The only problem with an implied namespace like that is making it clear
to the user that thats the namespace he/she will need to use. Maybe
clear documentation will be enough.
Ian
Jaroslaw Kowalski wrote:
How about local ? (local to the build file).
Jarek
- Original Message -
From:
I hope so Nick.
personal-aside
This is still on the list, but as we are a collection of individual
developers, each with our own interests, it sometimes turns out that one of
those individuals can re-focus the direction of the group. I think all it
would take to get filterchains going is for
10 matches
Mail list logo