The problem is that putting it in pluginManagement doesn't help because the builds genuinely do want to use different versions, particularly of internally-developed mojos that change in different (and sometimes incompatible ways).
A typical case is a base library uses a fixed, stable release, but another project has needed further development, so moves to a snapshot. The projects all build in isolation, but in reactor builds (build project and all dependencies) they fail seemingly mysteriously ways, because be base library builds first (and is owned by a separate team who may be unaware) and lock down the plugin. I've also seen it with different teams using different versions of things like the assembly plugin. Anyway... Is 2.1 going to be the next released M2 version, rather than a 2.0.9? I thought 2.1 was still pretty unstable? Would the fix (or something like it) in MNG-3284 work? (if not I'm prepared to do some more work on it, if someone can point me to where I need to look...) TIA, Nigel On Feb 3, 2008 5:56 PM, Brian E. Fox <[EMAIL PROTECTED]> wrote: > This is fixed in 2.1, but if you are constantly having this problem, it > sounds like maybe you need to use pluginManagement in a corp pom to > resolve this issue. > > > -----Original Message----- > From: Nigel Magnay [mailto:[EMAIL PROTECTED] > Sent: Sunday, February 03, 2008 8:44 AM > To: [email protected] > Subject: MNG-3284 > > Hi. > > In our work projects, we're *constantly* being bitten by the 'first > version of a plugin used wins' problem, which seems variously > described in particular MNG JIRA items, mine in particular being > MNG-3284. > > I've had a go at fixing it as it didn't seem that complicated - but - > some of the comments in the code and some of the other JIRA tickets > allude to the problem being possibly bigger than what I've looked at. > Unfortunately I can't tell if these are things that are still issues > as it dates back quite a long way. > > I'd really like it if someone who's more intimate with the workings > could tell me what else would need considering - that way I could > generate some test cases and perhaps have a stab at a wider fix. > > TIA, > Nigel > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
