Pádraig Brady <[email protected]> writes: > That shouldn't cause any issues elsewhere so it's good to push, > but we might look at making that test more robust after release.
I agree. Sorry, I probably should have mentioned that in my email. On GNU/Hurd, it basically ran until your disk was full, which we should try obviously try to avoid in all circumstances. It is a bit tricky here, though. Since we want to save the process ID for 'dd' to signal later, we can't use timeout. Not easily, at least. Since the test hasn't really been changed in a long time, and the issue only occurs on GNU/Hurd, and occurs regularly, I preferred just making that tiny change. I didn't want to risk introducing a bug with added complexity just before the release. Collin
