On 1/29/21 2:25 PM, Pavel Hrdina wrote:
On Fri, Jan 29, 2021 at 02:07:31PM +0100, Michal Privoznik wrote:
On 1/29/21 1:45 PM, Pavel Hrdina wrote:
On Fri, Jan 29, 2021 at 09:30:24AM -0300, Daniel Henrique Barboza wrote:
So we agree that timeouts are evil :-) Back in the autotools days we did not
No we don't :) timeouts are useful in test suites to reveal issues that
something takes a long time when it shouldn't.
I view timeout as a prevention mechanism for a test that went into an
endless loop. If we really care about timing then there are way better
techniques than meson's test timeout.
Or put differently, I don't record times for individual test cases and I
would not notice if a test doubles its run time.
Michal
