gump-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Adam R. B. Jack" <>
Subject Re: Ought we consider a Gump on Brutus running JDK1.5?
Date Fri, 11 Jun 2004 12:08:29 GMT
From: "Brett Porter" <>:

> My 2c: I would have it ready to go when 1.5 goes into first release (so
> it ready now), but not nag anyone until then. While there might be helpful
> stuff, you'd want to avoid nagging people for bugs in the betas too - and
> might be hard to differentiate.

+1. I wasn't thinking of nagging, just testing.

> Niclas Hedhman wrote:
> >Shouldn't each project declare which JDK version to use for the build?
> I think a long-term goal could be for projects to declare minimum and
> maximum java versions and gumps to define in workspace.xml which
> versions of java are available (and where).

I definitely agree w/ the range. And Gump matching itself (or it's
environment, in this case) against the projects.

From: "Leo Simons" <>

> Gump has pretty much always took a "latest of everything" approach.
> Changing that will be difficult. For example, I remember having to
> upgrade from j2sdk_1.4.2_03 to 1.4.2_04 because xerces (was it xerces?)
> wouldn't build otherwise.

Yup, I think it was xerces.

> The problem right now is that lots of stuff won't work against jdk 1.5
> (like Berin writes). I hope that with the final release that will be
> less true. I like Brett's suggestion of waiting for 1.5 final.

It seems to me, the sooner we test that, the sooner we can report bugs to
Sun. :) Seriously though, I am in no great hurry, just didn't want to be
behind if the community were gagging for it.

BTW: It seems to me we ought be able to have N compilers available to Gump,
and we ought categorize output (within a Gump tree) by the compiler (and/or
class format version). Clearly components built on compiler X ought only be
compiled against sub-components of same version. A brute force approach of
one compiler per workspace, and multiple workspaces, is probably the easiest



To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message