Hi Daniel, David, Is there any way of disabling some individual tests on specific architectures? Zero cost exception handling on ARM is known to be broken and there isn't anyone I know working on it with high priority, so it's likely to be that way for a long time. For that reason, I'd like to disable them on ARM, so that I can get a green LNT, making it easier to spot regressions. There are other failures, but I'd rather fix them than disable. It's only on this case, because it's a low priority feature AND it's broken, that I'd suggest temporarily disabling the test. Makes sense? Is it possible? cheers, --renato -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20130315/ad970449/attachment.html>
On Mar 15, 2013, at 7:57, Renato Golin <renato.golin at linaro.org> wrote:> Hi Daniel, David, > > Is there any way of disabling some individual tests on specific architectures? > > Zero cost exception handling on ARM is known to be broken and there isn't anyone I know working on it with high priority, so it's likely to be that way for a long time. > > For that reason, I'd like to disable them on ARM, so that I can get a green LNT, making it easier to spot regressions. > > There are other failures, but I'd rather fix them than disable. It's only on this case, because it's a low priority feature AND it's broken, that I'd suggest temporarily disabling the test. > > Makes sense? Is it possible?Hi Renato, Yes, makes sense and is usually possibly. I don't have a reference offhand but the way other tests do this is just filter the test out of the targets based on the target, etc. If you grep for filter-out in the Makefiles there should be a couple of examples. - Daniel> > cheers, > --renato
On 16 March 2013 06:53, Daniel Dunbar <daniel.dunbar at gmail.com> wrote:> I don't have a reference offhand but the way other tests do this is > just filter the test out of the targets based on the target, etc. If > you grep for filter-out in the Makefiles there should be a couple of > examples. >Thanks Daniel, I just added a ifned(ARM) for DIRS=EH on C++ regression Makefile. That's quite simple, worked well and will be easy to revert once EH support is better on ARM. cheers, --renato -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20130318/8a95a0d4/attachment.html>