[aspectc-user] Errors in the compilation process of Puma on Tru64 Unix
Olaf.Spinczyk at informatik.uni-erlangen.de
Mon Mar 29 17:45:26 CEST 2004
Gaetano.Vivo at libero.it wrote:
> Hello, I have tried the compilation of Puma on my system (Tru64 UNIX 5.1A on alpha) in the hope of being able to then compile ac++ on the same machine.
> The compilation of puma has happened second the following steps:
> 1) execution of ' make weave' on linux-machine.
> 2) copy of sources generates on the machine target (Tru64 UNIX)
> 3) execution of ' make compile' on the target-machine.
> in the course of the last step the following errors have been taken place:
> 1) The sources generated during the phase of weave have produced one variable of type ' long long' that create ambiguity in the choice of the operator ' < < ' during the operations of I/O. ( I have resolved this problem modifying by hand the code generated from the phase of weave transforming all the variable ones of type long long in the type long. )
Well, as g++ normally supports 'long long' this seems to be a problem
with the <iostream> header file you are using. If you can't fix the
problem on this side you can edit src/basics/Limits.h. That's the point
where we configure whether 'long long' is used.
> 2) The sources generated during the phase of weave have produced template functions that generate functions similar to following:
> void& foo();
> or that they use variable of type void * like pointers to objects as:
> void * p;
This is a known problem. Please try it again with ac++-0-8pre2. We will
publish a bug fix release 0.8.1 as soon as possible.
> I do not know as to resolve this problem for which they are completely firm.
> I hope you can help me. Thanks in advance and excused my English.
> PS. The compiler i have used is GCC 3.3.2
I hope that helps. Please stay in touch.
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 3142 bytes
Desc: S/MIME Cryptographic Signature
More information about the aspectc-user