This is the mail archive of the gcc-help@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: -frepo in 4.4 snapshot fails


Edek PieÅkowski wrote:
Ian Lance Taylor wrote:

Basically you have to provide enough information for somebody with no
access to your system to reproduce the problem.

I think I would need some advice.


I tried *.ii files approach, but during recompiling it wants to recompile
original .cpp files. Is there a way to tell g++ to recompile .ii files during
-frepo relinking/recompiling? That would be a simple way of getting it done.


What I did so far:
I kind of reproduced the error on another system, x86, not x86_64 this time, with latest 4.4 snapshot. I say "kind of", because the type of error is the same, and relates to the same file, ("was assigned to /home/e1/tmp/frepo/trunk/build/obj/util/impl/xml/XmlImplementationFactory.rpo', but was not defined during recompilation, or vice versa"), but the actual symbol is different. It is also a different symbol on x86_64 when I compile with -save-temps than when I compile without -save-temps.


Then lots of undefined references follow.

On x86 I'm trying to remove everything unneeded. I am working on removing references to log4cxx, but I think I cannot really remove boost (smart_ptr). Would it be ok to post a bug report and say "you need boost to reproduce (and log4cxx). Just the includes, not the libraries..."? Or should I work on extracting just the needed headers from boost and bundle together what is left?


And is it all worth it, maybe in a snapshot such things are usual?


Thanks,
Edek


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]