Tom Lane wrote:
I see that the buildfarm script seems to be running ecpg-check pretty early in the sequence. Considering that the ecpg tests are still far from stable, this seems to be taking away the opportunity to learn as much as we can from a buildfarm run. Could we run the ecpg tests last?
We can, although I have tried to keep the order fairly logical. Are we seeing so many ECPG failures that it's a major blocker? bustard has had one ECPG failure in the last 5 days. We are currently only reporting failures on OpenBSD and the mipsel box.
Anyway, I have made the change in CVS and buildfarm members can upgrade to CVS version 1.69 of run_build.pl. I guess that would mainly matter for the owners of the currently failing 3 boxes. If it proves to be necessary beyond a small time frame I will cut a new release.
An even better idea would be to teach the script about test dependencies so that it could run test steps even when an earlier-but-unrelated test had failed. But I'm sure that's a lot more work.
Yes. It might be possible with the modularisation work I'm contemplating, but that's some way off, and I'm not going to hack it in right now.
cheers andrew ---------------------------(end of broadcast)--------------------------- TIP 9: In versions below 8.0, the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match