On Sat, Apr 18, 2015 at 11:43:53AM +0200, Landry Breuil wrote: > On Sat, Apr 18, 2015 at 07:15:08AM +0200, Sébastien Marie wrote: > > On Sat, Apr 18, 2015 at 06:17:18AM +0200, Sébastien Marie wrote: > > > Hi, > > > > > > This patch include: > > > - the previous patch to disambiguate linkage (not yet commited) > > > - update to rust-1.0.0beta1 > > > > > > > It should be better with patch included. > > > > And the update isn't "beta1" but "beta2"... > > One test failed, dunno if its expected:
not really "expected" but stack-overflow detection is a moving target... I don't think it is a regression. The problem per se isn't too grave: the test check if stack overflow is correctly detected as "stack overflow" by rust. THe normal behaviour is when a SIGSEGV is catch, the program check if it is due to stack-overflow. If yes, then an error-message is issued before killing the program. Else the program is just killed. The stack overflow detection in rust is implemented using pthread_stackseg_np(3). But it only "provides a rough estimation of stack bounds". So the detection could fail. The program is killed without a nice message saying: thread '<main>' has overflowed its stack > failures: > [run-pass] run-pass/out-of-stack.rs > > test result: FAILED. 2021 passed; 1 failed; 25 ignored; 0 measured > > Other than that looks good to me. > So it isn't a blocking problem for me. Thanks for your feedback. -- Sébastien Marie
