At 10:43 21/2/01 +0100, Christoph Wilhelms wrote: >Agreed, but am I right when I say, that, even in future, the target >threads will be STARTED in the order specified in the depend- >parameter.
not necessarily - they start in whatever order seems fun for the implementor - perhaps even random so that bad build files fail fast ;) >This does NOT imply that the first target IS FINISHED >before the second starts. I think it makes sense to be able to >specify this, and I really propose to include this in the AntSpec! This has come up in a number of different guises before. The current solution is to use antcall - and I expect the same solution to be used in the future. Not sure - though we may even add in ordering semantics - not really sure ;) Work on Ant2.0 (where this behavgiour will be determined) has not really been active lately ;) >Tim's workaround will always work, but it's really just a "workaround" >because sometimes I do not want dependency-chains! Look at this one: > >Targets: >cleanclassesdir >compile >buildall > >I do not want <compile> to depend on <cleanclassesdir> but when I >execute <buildall> then first <cleanclassesdir> and then <compile> >shlould be executed! In that case multithreading wouldn't make >sense, too ;-). REMEBER: you can execute single targets, wich aren't >the default target! In GUI-context like Antidote or the >VAJIntegration this is the normal use case! usually I get this by > ant[.bat] buildall cleanall which will execiute target "buildall" and then "cleanall" Cheers, Pete *-----------------------------------------------------* | "Faced with the choice between changing one's mind, | | and proving that there is no need to do so - almost | | everyone gets busy on the proof." | | - John Kenneth Galbraith | *-----------------------------------------------------*
