On Fri, Jul 11, 2025 at 09:53:10AM +0100, Lorenzo Stoakes wrote:
On Fri, Jul 11, 2025 at 09:11:47AM +0100, Mark Brown wrote:
One thing to watch out for with peering into the private header files of other selftests is that it's a routine source of build and sometimes runtime failures, people have a tendency to update one selftest without thinking that other selftests might be peering at their code. The cross tree aspect can make it painful to deal with the resulting issues.
I take it from the lack of reported issues this hasn't happened in reality.
That's a general comment about this pattern over the selftests as a whole rather than this specific header - if this one has been working well then great (I certainly didn't run into it myself). In general I'd say this pattern is up there as one of the most common individual sources of build breaks in the selftests, and often has a relatively high level of friction getting things fixed compared to the others.