* Stepan Kasal wrote on Thu, May 12, 2005 at 02:57:46PM CEST: > On Thu, May 12, 2005 at 02:02:54PM +0200, Ralf Wildenhues wrote: > > * Harald Dunkel wrote on Thu, May 12, 2005 at 01:43:51PM CEST: > > > Anyway, if I introduce a line > > > > > > myexe: some_internal_tool_used_at_build_time > > > > > > then Automake shows the same misbehaviour: It assumes that > > > I would like to redefine the build rule for myexe. > > well, this behaviour is documented in the Automake manual, > in node ``Extending'': > > : Note that Automake does not make any difference between rules with > : commands and rules that only specify dependencies. So it is not > : possible to append new dependencies to an `automake'-defined target > : without redefining the entire rule. ACK. > But Ralf said: > > > Well, that's just how `make' syntax works (portable make != GNU make). > > You mean that some make implementations don't allow you to specify extra > dependencies in a rule with no commands (so called "separated > dependencies")? Nono, I did not mean that (but now that you point it out I see that I missed the fact that his line quoted above looks differently from before). Harald had, in one of his previous posts, a line like this: target: CC=gcc3 and obviously, `CC=gcc3' was not intended to be a dependency of `target', but some kind of syntax to change a variable esp. for the rules of updating `target'. *This* intended behavior is a broken assumption. > I'm not convinced about this. Node "(autoconf.info)Limitations of Make" > speaks about problems when you combine separated dependencies with so > called "single suffix rules". But generally, separated dependencies > should work. Sure. Thanks for clarifying this point. Regards, Ralf _______________________________________________ Autoconf mailing list Autoconf@xxxxxxx http://lists.gnu.org/mailman/listinfo/autoconf
