On 01.03.2004 07:02, Stefano Mazzocchi wrote:
What about just adding an element for licence location to jars.xml? e.g.,
Gump descriptors now contain license information and gump is checking it for you. I think we should use that.
I think gump would be a cool tool if we would use it correctly or if it's more strict - but maybe I only don't see everything necessary. In theory gump descriptor contains all information about projects, sub projects (our blocks), dependencies and so on. But why for example does our gump descriptor not contain information about all JARs like about joost/stx at the end of the file:
<project name="joost"> <url href="http://joost.sourceforge.net"/> <description> Streaming Transformation for XML (STX) library </description> <home nested="src/blocks/stx/lib"/> <jar name="joost-20031219.jar"/> </project>
Is it only a question of maintenance? Or how can the descriptor say that Cocoon core depends on "rhino-cocoondev" but this one is nowhere described? The same is true for "concurrent" and many more. I might be wrong but in theory gump seems to be a perfect replacement for our jars.xml - no need for duplicate effort of maintaining JARs and JAR descriptions. And we don't need to maintain the JAR checker code anymore, just the conversion from gump descriptor to JAR documentation (XDocs).
Is our gump descriptor that bad maintained? Some time ago we also got gump error messages to the list, now we get those only from persons outside of Cocoon. What's the benefit of gump when we don't use it correctly?
Gump has not been able to run cocoon for the last 6 months at least, due to unmet dependencies. As a result, our decriptor is basically dead because we haven't been nagged for that period of time.
The gump community is working hard to make this possible again, but this is very hard work and takes a lot of time and patience.
My proposal (that the board accepted) to move gump to top level status was also to bring more visibility and make projects being more gump-friendly, we'll see how this goes.
I have a RT ready for gump that would allow to solve these issues... but I need to work more on this since it appears to be a pretty tough computational problem of graph analysis.
-- Stefano.
smime.p7s
Description: S/MIME Cryptographic Signature
