[patch, testsuite] test newlib.string/strcmp-1.c takes too long
Sun May 6 17:46:00 GMT 2012
On 02/05/12 14:36, Corinna Vinschen wrote:
> strcmp is one of the first candidates for a target-specific rewrite in
> assembler. So there's a good chance that you see a rewrite as soon as
> a new target gets supported. Shouldn't we rather opt on the safe side
> by default?
This is a regression testsuite, it's run regularly by large numbers of
people (we run it every night for multiple permutations). We don't
modify strcmp every day though.
When you're actively modifying strcmp, running comprehensive tests is a
good idea, but I don't see the need to retry every combination every
day. I think we should try the short version. If folks modifying
strcmp find that there are errors detected by the long version that
aren't caught be the short one, then we should re-visit...
I'll also note that the test is quite new; so we're still better off
than we were before...
More information about the Newlib