On Saturday 07 November 2009, Thomas Kindler wrote: > If more people had tried the RCs, the problem might have shown up before > the release.
No; the real issue here was slipping a very bad patch in, late in the RC cycle ... *AND* then not extending the test cycle enough to catch it. Either with another RC, or just by holding off for longer. In retrospect, any patch of the "I don't know *WHY* this reportedly fixes <problem X>" ilk should never be merged, and should be setting off warning sirens and alarms. Rule of thumb: if developers can't explain it, it's likely a bad patch ... fixing symptom by accident (or Heisenberg), not the root cause. As acting Release Manager, I really don't like the idea of needing to prevent everyone else from making commits during the final week or two before release. It's not like I want to increase my workload that much (for a job I didn't even volunteer for) ... but it might come down to such tactics. _______________________________________________ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
