Boost logo

Boost :

Subject: Re: [boost] Permission to proceed?
From: Agustín K-ballo Bergé (kaballo86_at_[hidden])
Date: 2013-10-13 12:39:52


On 13/10/2013 05:20 a.m., Stephen Kelly wrote:
> On 10/13/2013 02:38 AM, Agustín K-ballo Bergé wrote:
>> May I ask you to try and build/test Boost before pushing your changes
>> in the future? That doesn't guarantee it will build everywhere, but at
>> least it will let you catch this kind of simple errors.
>
> I did run most of the tests, but I must have missed function_types
> somehow...
>
> Even with function_types/detail/cv_traits.hpp fixed, I'm still getting
> errors
> stephen_at_hal:~/dev/src/boost-trunk/libs/function_types/test{master}$
> ../../../b2
> ...patience...
> ...found 1531 targets...
> ...updating 8 targets...
> gcc.compile.c++
> ../../../bin.v2/libs/function_types/test/nonmember_ccs.test/gcc-4.8/debug/nonmember_ccs.o
> custom_ccs/nonmember_ccs.cpp:19:5: error: #error "test not supported
> with this compiler/platform"
> # error "test not supported with this compiler/platform"
> ^
>
> I don't see how that test could *not* fail for me, and I don't see
> anything in the Jamfile to restrict that test to certain platforms.

It seems the test checks for compiler specific functionality, and chose
a weird way of acting on compilers that do not have said functionality.
I would have just made the entire test a no-op on those compilers. You
may want to create a ticket for this, so the author checks it.

In any case, running the tests before and after your changes and then
diffing them should be enough to spot those trivial syntax errors. You
certainly don't need to fix test that were already failing.

Regards,

-- 
Agustín K-ballo Bergé.-
http://talesofcpp.fusionfenix.com

Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk