M3 regression: reactor calculates wrong build order

[Jörg Schaible <jo...@gmx.de>]

Hi,

I'd like to call your attention again to a M3 regression that prevents us 
from switching to M3 in general. In MNG-5207 I've reported before that the 
reactor might calculate the wrong build order. If your local repo does not 
already contain any of the SNAPSHOTs that are build, the reactor simply 
fails doe to a missing dependency. However, if the SNAPSHOT is already 
available it will silently pack stale SNAPSHOTs (into WARs, EARs or ZIPs).

Since I had today some spare time in the office (I'll leave for holidays, so 
I won't be able to response for some time), therefore I've added now an IT 
for Maven's core-it-suite that demonstrates the problem. I hope this will 
help to identify where the build order calculation fails in M3. M221 does it 
right.

Regards,
Jörg


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org