Missing -lpython2.7 when linking ceph-dencoder

Missing -lpython2.7 when linking ceph-dencoder

[Willem Jan Withagen]

Hi Kefu,

This change has happend relatively early in my vacation because all the
build that are left sort of hang on this. ceph-dencoder just misses out
on a large number of Py object that it can not link. And that again
comes from linking boost_python.

Now I remember we had this problem when we reworked the
find_package(python) stuff.
Any suggestions other than starting to bisect this?

--WjW

Re: Missing -lpython2.7 when linking ceph-dencoder

[Willem Jan Withagen]

On 26-5-2017 16:12, Willem Jan Withagen wrote:
> Hi Kefu,
> 
> This change has happend relatively early in my vacation because all the
> build that are left sort of hang on this. ceph-dencoder just misses out
> on a large number of Py object that it can not link. And that again
> comes from linking boost_python.
> 
> Now I remember we had this problem when we reworked the
> find_package(python) stuff.
> Any suggestions other than starting to bisect this?

Right,

It is the other way around....
Boost linking just always includes boost_python which in turn misses
-lpython to be successful.
So now the question is, why am I getting boost_python there?

--WjW

Re: Missing -lpython2.7 when linking ceph-dencoder

[Willem Jan Withagen]

On 26-5-2017 16:37, Willem Jan Withagen wrote:
> On 26-5-2017 16:12, Willem Jan Withagen wrote:
>> Hi Kefu,
>>
>> This change has happend relatively early in my vacation because all the
>> build that are left sort of hang on this. ceph-dencoder just misses out
>> on a large number of Py object that it can not link. And that again
>> comes from linking boost_python.
>>
>> Now I remember we had this problem when we reworked the
>> find_package(python) stuff.
>> Any suggestions other than starting to bisect this?
> 
> Right,
> 
> It is the other way around....
> Boost linking just always includes boost_python which in turn misses
> -lpython to be successful.
> So now the question is, why am I getting boost_python there?

The standard include Boost code in Cmake returns all available libraries.
So for now I'm manually removing
	/usr/local/lib/libboost_python.so
from the Boost_LIBRARIES.

Now REMOVE_ITEM does not use regual expressions, so it is hardcoded.
Which looks sort of clumsy.

I'll submit a PR for this.

--WjW

Re: Missing -lpython2.7 when linking ceph-dencoder

[kefu chai]

On Fri, May 26, 2017 at 11:32 PM, Willem Jan Withagen <wjw@digiware.nl> wrote:
> On 26-5-2017 16:37, Willem Jan Withagen wrote:
>> On 26-5-2017 16:12, Willem Jan Withagen wrote:
>>> Hi Kefu,
>>>
>>> This change has happend relatively early in my vacation because all the
>>> build that are left sort of hang on this. ceph-dencoder just misses out
>>> on a large number of Py object that it can not link. And that again
>>> comes from linking boost_python.

Willem, could you pastebin the error message somewhere, so i can take
a look at it?

>>>
>>> Now I remember we had this problem when we reworked the
>>> find_package(python) stuff.
>>> Any suggestions other than starting to bisect this?
>>
>> Right,
>>
>> It is the other way around....
>> Boost linking just always includes boost_python which in turn misses
>> -lpython to be successful.
>> So now the question is, why am I getting boost_python there?
>
> The standard include Boost code in Cmake returns all available libraries.
> So for now I'm manually removing
>         /usr/local/lib/libboost_python.so
> from the Boost_LIBRARIES.
>
> Now REMOVE_ITEM does not use regual expressions, so it is hardcoded.
> Which looks sort of clumsy.
>
> I'll submit a PR for this.
>
> --WjW
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html



-- 
Regards
Kefu Chai