[PATCH v3.1] New functions pthread_[sg]etattr_default_np for default thread attributes

[Joseph S. Myers]

On Thu, 13 Jun 2013, Siddhesh Poyarekar wrote:

> > I think that's a bad idea.  In practice what goes in the test's .out file 
> > is *more* visible than what goes in the huge pile of make check output.  
> > Arguably stderr as well as stdout should be redirected.
> 
> IIRC the rationale behind that was that such warnings go unnoticed and
> it is assumed that all tests *passed* when in reality some may have
> been skipped.  I think [1] may have been the point where I had noted

The notion that people notice a warning that scrolls off the screen in 
under a second is unrealistic.  If I want details of what happened for a 
particular test, I examine its .out file.  To distinguish passing and 
skipped tests for the overall testsuite status, we need more structured 
information in the .out files about the status of particular test 
assertions, rather than unstructured warnings appearing mixed up with huge 
numbers of long compilation and test-running commands.

-- 
Joseph S. Myers
joseph@codesourcery.com