At 12:42 17/5/01 +0100, Jose Alberto Fernandez wrote: >> From: Diane Holt [mailto:[EMAIL PROTECTED] >> >> >> >> Sure there is. Given 'ant A B', it'll run X,Y,A,Y,X,B; >> running 'ant B A' >> runs Y,X,B,X,Y,A. And if you add a target "Z" that depends on >> both A and >> B, listed as depends="A,B", running 'ant Z' will run >> X,Y,A,B,Z -- listed >> as depends="B,A", it'll run Y,X,B,A,Z. (This is almost >> identical to how >> 'make' would do it, except for the first case, where it would only run >> X,Y,A,B, since X and Y were already run -- 'make b a' runs Y,X,B,A.) >> > >Should ANT behave as make does? Today, the call 'ant A B' is equivalent to: > > ant A > ant B > >two independent execusions, two independent dependency graphs. I have always >thought that it is kind of wrong, that it should behave as make does: > > ant 'A B' > >where <target name="A B" depends="A,B" /> is consider to be a new pseudo >target added to the buildfile. > >By the way, that would also give a hint on the correct meaning of: > > <ant target="A,B" .../> > >which should not be the same as: > > <ant target="A" .... /> > <ant target="B" .... /> > >What do you all think?
indifferent to it - whatever we decide we should clearly doc it though ;) 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 | *-----------------------------------------------------*
