[molpro-user] compilation troubles on AIX 5.3

[Toon Verstraelen]

Hi,

I didn't use these options either. Despite the title of the bug report 
these options do not have any relation with the problem. They are just 
used for a more verbose output.

Toon

Jyh-Shyong Ho wrote:
> Hi,
> 
> Thanks, I checked the IBM page and found nothing I can do on my part, 
> there is no pointer reference
> in file ftc.c, and the compler parameters -qinfo=cnd or -qinfo=all was 
> not used in compiling ftc.c.
> 
> I'll wait the local IBM engineer to apply all the compiler updates and 
> patches on our IBM P595
> before taking further steps.
> 
> .Jyh-Shyong Ho, Ph.D.
> Research Scientist
> National Center for High Performance Computing
> Hsinchu, Taiwan, ROC
> 
> 
> Toon Verstraelen wrote:
>> Hi,
>>
>> I wrote a mistake in my previous reply. The bug related to ras.h is 
>> the fourth item in the list on the ibm website: 'IY86292: ERROR IN 
>> RAS.H V1.15 -QINFO=CND OR -QINFO=ALL'
>>
>> good luck,
>>
>> Toon
>>
>> Jyh-Shyong Ho wrote:
>>> Thanks for your information. A local IBM engineer is working on setting
>>> up out IBM P595, I'll have him to fix this system bug as soon as 
>>> possible.
>>>
>>> When I finish building Molpro2006.1, I'll let you know if it works.
>>>
>>> Best regards
>>>
>>> Jyh-Shyong Ho, Ph.D.
>>> Research Scientist
>>> National Center for High Performance Computing
>>> Hsinchu, Taiwan, ROC
>>>
>>>
>>>
>>> Toon Verstraelen 提到:
>>>
>>>> Jyh-Shyong Ho wrote:
>>>>
>>>>> Dear Molpro users,
>>>>>
>>>>> We to have some difficulties in building Molpro 2006.1 on our IBM 
>>>>> P595 (AIX 5.3) system,
>>>>> the compilation failed even in building the utilitiy ftc.c:
>>>>>
>>>>> cc -q64 -qmaxmem=-1 -c -DSEEK=.001 -DSPEED=200000 -qarch=auto -DI64 
>>>>> -q64 -qwarn64 -DAIX42 -DAIX5 -DNOALRM -DUSE_LAPI -DHAVE_INTTYPES_H 
>>>>> -DMOLPRO_GETOPTLONG -DMA_ALLOC -DMOLPROC_PAR 
>>>>> -DMOLPROBASE=\"/package/chem/molpro/source/molpro2006.1\" -I. -O2 
>>>>> ftc.c
>>>>> "/usr/include/sys/ras.h", line 357.1: 1506-195 (S) Integral 
>>>>> constant expression with a value greater than zero is required.
>>>>> "ftc.c", line 43.47: 1506-743 (I) 64-bit portability: possible 
>>>>> change of result through conversion of int type into unsigned long 
>>>>> int type.
>>>>> "ftc.c", line 44.38: 1506-743 (I) 64-bit portability: possible 
>>>>> change of result through conversion of int type into unsigned long 
>>>>> int type.
>>>>> "ftc.c", line 55.12: 1506-742 (I) 64-bit portability: possible loss 
>>>>> of digits through conversion of unsigned long int type into int type.
>>>>> "ftc.c", line 61.39: 1506-743 (I) 64-bit portability: possible 
>>>>> change of result through conversion of int type into unsigned long 
>>>>> int type.
>>>>> "ftc.c", line 62.30: 1506-743 (I) 64-bit portability: possible 
>>>>> change of result through conversion of int type into unsigned long 
>>>>> int type.
>>>>>
>>>>
>>>> The problem with ras.h is a bug in the IBM VAC C/C++ compiler, which 
>>>> was fixed in the update of june 2006. Please upgrade your compiler 
>>>> tot the latest patch release. More info on this problem (and the 
>>>> updates) can be found via the ibm website:
>>>>
>>>>
>>>> If the direct link doesn't work, just go to ibm.com, click on 
>>>> 'support and downloads', then in the box with title 'Support by 
>>>> product', click on the link 'software', on the next page click 
>>>> 'search' and then type 'ras.h' in the box under '2. Enter search 
>>>> terms' and press enter. This will produce a list of problems related 
>>>> to ras.h. The first item in this list describes this problem: 
>>>> 'IY90668: SYS/RAS.H DOES NOT #IFDEF OUT "LONG LONG" TYPE FOR C89'. 
>>>> Click on this link and you will find info about compiler updates etc.
>>>>
>>>> Let me know if it works (or not). Can you send a copy of your CONFIG 
>>>> file on the mailing ligts? We would like to compare it with the one 
>>>> we are using. It might be instructive.
>>>>
>>>> best regards,
>>>>
>>>> Toon
>>>
>>