Re: changing Ivy retrieve behaviour

2009-03-16 Thread Maarten Coene
Subject: Re: changing Ivy retrieve behaviour Or maybe we could adapt the current conflict strategy to only to select the latest version of the same artifact (same name, type, extension, extra attributes) and throw an error in any other situation. That sounds good indeed, and I agree in this case

Re: changing Ivy retrieve behaviour

2009-03-12 Thread Xavier Hanin
Subject: Re: changing Ivy retrieve behaviour On Mon, Mar 9, 2009 at 6:53 PM, Gilles Scokart gscok...@gmail.com wrote: I think failing the build is indeed the best aproach. But I fear some build that might be broken. Isn't it? I think some builds would fail, so I'd classify this as non

Re: changing Ivy retrieve behaviour

2009-03-10 Thread Xavier Hanin
On Mon, Mar 9, 2009 at 6:53 PM, Gilles Scokart gscok...@gmail.com wrote: I think failing the build is indeed the best aproach. But I fear some build that might be broken. Isn't it? I think some builds would fail, so I'd classify this as non backward compatible change. So I think the best

Re: changing Ivy retrieve behaviour

2009-03-10 Thread Maarten Coene
or less backwards compatible, and we'll solve a very common problem Maarten - Original Message From: Xavier Hanin xavier.ha...@gmail.com To: Ant Developers List dev@ant.apache.org Sent: Tuesday, March 10, 2009 5:06:43 PM Subject: Re: changing Ivy retrieve behaviour On Mon, Mar 9, 2009

Re: changing Ivy retrieve behaviour

2009-03-09 Thread Gilles Scokart
I think failing the build is indeed the best aproach. But I fear some build that might be broken. Isn't it? I'm also wondering if it is not possible to have warning in the retrieve task when some kind of patterns are used. Gilles Scokart 2009/3/9 Maarten Coene maarten_co...@yahoo.com Hi,

Re: changing Ivy retrieve behaviour

2009-03-09 Thread Stefan Bodewig
On 2009-03-09, Gilles Scokart gscok...@gmail.com wrote: 2009/3/9 Maarten Coene maarten_co...@yahoo.com I would like to change this so that Ivy throws an error if multiple different artifacts map to the same file. Is this a condition a user can fix? Or is the reason bad upstream metadata?