[Coin-discuss] BCP with Xpress under cygwin

[Andreas Waechter]

Hi Laci,

True.  I can set up tests in CoinUtils that try to find out the 64bit 
integer type (signed/unsigned?), and introduce a CoinInt64 type in 
CoinTypes.

Where would I change CoinUtils?  In devel, or in stable...?

Thanks

Andreas

On Wed, 11 Oct 2006, Laszlo Ladanyi wrote:

>
>
> On Wed, 11 Oct 2006, Andreas Waechter wrote:
>
>> Hi Laci,
>>
>>> As temporary fix in your own copy just go into CoinUtils/src and modify
>>> CoinTypes.hpp to define a 64 bit integer to something on the non-MSC_VER part
>>> of the test. My guess is that
>>>    typedef long long int64_t;
>>> will work.
>>>
>>> Andreas, can you add a macro that figures out what type provides 64-bit
>>> integers (and what needs to be included to get it) and in CoinTypes.hpp we
>>> could use the result of that?
>>
>> Well, I don't think all systems necessarily have a 64bit integer.
>> Usually, the header stdint.h is supposed to declare integer types, but
>> even the IEEE standard doesn't necessarily know about 64bit:
>>
>> http://www.opengroup.org/onlinepubs/000095399/basedefs/stdint.h.html
>>
>> I suggest instead to correct the CoinIndexedVector.cpp file so that it
>> does not use platform dependent types.  This seems to be the only place
>> where int64_t appears.
>>
>> Cheers
>>
>> Andreas
>
> OK, so what to do when ther *is* a 64-bit integer? Pretty much every platform
> provides it in some form, even if it is not part of the standard. Granted, in
> our code we should use CoinInt64 or something similar, but if there is a
> native 64-bit int on the platform then it should just be a simple typedef.
> And for the odd platform that does not provide such type we can code it up in
> any ugly inefficient way.
>
> What do you think? (Besides that I'm so good at giving work to others :-).
>
> --Laci
>
> PS: If you'd like I can try to cook up a test...
>
>>
>>>
>>> Thanks,
>>> --Laci
>>>
>
> _______________________________________________
> Coin-discuss mailing list
> Coin-discuss at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/coin-discuss
>