[Fusion] Lots of test failures on develop
I'm getting a lot of test failures on the develop branch of Fusion on OS X. There are too many to paste the output here. Note that master is just fine. Are the unit tests being run automatically each time a pull request is merged? Louis
On 11/12/14, 12:04 AM, Louis Dionne wrote:
I'm getting a lot of test failures on the develop branch of Fusion on OS X.
There are too many to paste the output here. Note that master is just fine. Are the unit tests being run automatically each time a pull request is merged?
Which compiler? All tests pass at least on Clang here. I just did a sync and full test. I'll look into this, Louis. There's been a lot of clean up and bug fixing code changes lately. Thanks for reporting, Louis. Regards, -- Joel de Guzman http://www.ciere.com http://boost-spirit.com http://www.cycfi.com/
Joel de Guzman
On 11/12/14, 12:04 AM, Louis Dionne wrote:
I'm getting a lot of test failures on the develop branch of Fusion on OS X.
There are too many to paste the output here. Note that master is just fine. Are the unit tests being run automatically each time a pull request is merged?
Which compiler? All tests pass at least on Clang here. I just did a sync and full test. I'll look into this, Louis. There's been a lot of clean up and bug fixing code changes lately. Thanks for reporting, Louis.
That was on Apple's Clang. It turned out that the problem was the BOOST_CXX14_CONSTEXPR macro being used without ever being defined (I grepped all of Boost for it and couldn't find a definition). I replaced it by BOOST_CONSTEXPR, which is used in a lot of other places and is most likely what the author of the faulty commit intended. I'll be creating a pull request in the next minutes. Louis
On 11/12/14, 7:01 AM, Louis Dionne wrote:
Joel de Guzman
writes: On 11/12/14, 12:04 AM, Louis Dionne wrote:
I'm getting a lot of test failures on the develop branch of Fusion on OS X.
There are too many to paste the output here. Note that master is just fine. Are the unit tests being run automatically each time a pull request is merged?
Which compiler? All tests pass at least on Clang here. I just did a sync and full test. I'll look into this, Louis. There's been a lot of clean up and bug fixing code changes lately. Thanks for reporting, Louis.
That was on Apple's Clang. It turned out that the problem was the BOOST_CXX14_CONSTEXPR macro being used without ever being defined (I grepped all of Boost for it and couldn't find a definition). I replaced it by BOOST_CONSTEXPR, which is used in a lot of other places and is most likely what the author of the faulty commit intended.
I'll be creating a pull request in the next minutes.
Thanks, Louis! Merged. Regards, -- Joel de Guzman http://www.ciere.com http://boost-spirit.com http://www.cycfi.com/
On 11/11/2014 5:01 PM, Louis Dionne wrote:
That was on Apple's Clang. It turned out that the problem was the BOOST_CXX14_CONSTEXPR macro being used without ever being defined (I grepped all of Boost for it and couldn't find a definition).
Here: https://github.com/boostorg/config/blob/master/include/boost/config/suffix.h... Regards, -- Agustín K-ballo Bergé.- http://talesofcpp.fusionfenix.com
On 11/12/14, 7:38 AM, Agustín K-ballo Bergé wrote:
On 11/11/2014 5:01 PM, Louis Dionne wrote:
That was on Apple's Clang. It turned out that the problem was the BOOST_CXX14_CONSTEXPR macro being used without ever being defined (I grepped all of Boost for it and couldn't find a definition).
Here: https://github.com/boostorg/config/blob/master/include/boost/config/suffix.h...
Reverted. Thanks for spotting this, Agustín. Makes sense. I wondered why I am not getting the same error as Louis. Same compiler, same platform. Regards, -- Joel de Guzman http://www.ciere.com http://boost-spirit.com http://www.cycfi.com/
Louis Dionne wrote:
It turned out that the problem was the BOOST_CXX14_CONSTEXPR macro being used without ever being defined (I grepped all of Boost for it and couldn't find a definition).
It's possible that this macro is on Config's develop branch and hasn't yet found its way into master. If so, its use may well be correct.
Peter Dimov
Louis Dionne wrote:
It turned out that the problem was the BOOST_CXX14_CONSTEXPR macro being used without ever being defined (I grepped all of Boost for it and couldn't find a definition).
It's possible that this macro is on Config's develop branch and hasn't yet found its way into master. If so, its use may well be correct.
Yes, that was the problem. Sorry for the fuss. How do you guys keep your submodules in sync in the modularized Boost? Do you use something like git submodule foreach git checkout develop && git pull from time to time, or is there a better way? Louis
On 11/12/2014 9:53 AM, Louis Dionne wrote:
Peter Dimov
writes: Louis Dionne wrote:
It turned out that the problem was the BOOST_CXX14_CONSTEXPR macro being used without ever being defined (I grepped all of Boost for it and couldn't find a definition).
It's possible that this macro is on Config's develop branch and hasn't yet found its way into master. If so, its use may well be correct.
Yes, that was the problem. Sorry for the fuss. How do you guys keep your submodules in sync in the modularized Boost? Do you use something like
git submodule foreach git checkout develop && git pull
from time to time, or is there a better way?
I suggest you checkout each submodule as 'master' and then test your own work as 'develop'. If you have dependencies on 'develop' you can easily check them out as such. In general trying to test your own work against everything else as 'develop' is just too much trouble IMO.
participants (5)
-
Agustín K-ballo Bergé -
Edward Diener -
Joel de Guzman -
Louis Dionne -
Peter Dimov