Hi. This keeps failing for me: More build logs here:
https://people.debian.org/~sanvila/build-logs/gawk/ The failure rate is about 1.5%, which I agree it's very low (but not so low that I can't measure it!). So, if you try to reproduce it, please try a *lot* of times. For example, if you build it 100 times, there is still a probability of 22% that it builds ok those 100 times (that would be (1-0.015)**100). As a summary, I did this: cat * | grep "recipe for target" | sort and the following tests are prone to fail: Makefile:2398: recipe for target 'fts' failed Makefile:3922: recipe for target 'time' failed Makefile:3941: recipe for target 'pass-fail' failed This time I have also used real machines (a bunch of old Pentium 4 at 3.00GHz I have at work), which are known to have a stable clock, so I don't really think that the machine clock is going back and forth. If, as you say, this does not mean that there is a bug in gawk, then we should probably assume instead that those three particular tests are not really meaningful as a way to tell that there is a problem in gawk. So: Could we please disable them? Thanks.
