On Wed, 08 Dec 2004, Stefan Bodewig <[EMAIL PROTECTED]> wrote: > On Mon, 06 Dec 2004, Stefano Mazzocchi <[EMAIL PROTECTED]> wrote: > >> So, here is my first suggestion: split gump in three stages. >> >> 1) metadata aggregation >> 2) build >> 3) build data use > > Sounds good.
One additional thing. I'd love to have part 2 separated into at least three steps that can get invoked indiviually: 2a) SCM update 2b) syncing updated working copy with workspace 2c) building With "traditional Gump" it has been possible to modify classes in the workspace and rebuild using Gump. This has been very useful in resolving Gump problems in the past. Right now I don't see an easy way to do this. For example, I "fixed" the commons-jelly-tags-ant build by patching the jelly-util taglib. I verified it would fix the Gump build by applying my patch locally and only building commons-jelly-tags-util and after that commons-jelly-tags-ant. Using current Gump my local patch would have been blown away by CVS updates or syncs - unless I applied it in what is supposed to be a clean checkout and disconnected from the network. Also, just building commons-jelly-tags-util and commons-jelly-tags-ant without rebuilding Ant and all that seems to be impossible right now (I may be wrong, though). Stefan --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
