[bip] alignment/padding differs

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

[bip] alignment/padding differs

Boost - Users mailing list
Hi,

I'm using interprocess to share mem between different processes.
Unfortunately these processes were built using different compiler
version (gcc3, gcc4, gcc5) and we're seeing different alignment or
padding dependent on which run-time was in charge of creating the mapped
files. Of course we used the same boost version (1.60) while creating
the binaries.

Obviously ip can't handle these different layouts and mostly runs in
indefinite loops while traversing an rb-tree not finding the first shm
object.

Are there any template parameters that enable the code to be accessed
using different run times through forcing the same layout?

Thanks a lot,
Jan.

_______________________________________________
Boost-users mailing list
[hidden email]
http://lists.boost.org/mailman/listinfo.cgi/boost-users
Reply | Threaded
Open this post in threaded view
|

Re: [bip] alignment/padding differs

Boost - Users mailing list
Forgot to mention: all of the involved binaries are 64bit arch.
_______________________________________________
Boost-users mailing list
[hidden email]
http://lists.boost.org/mailman/listinfo.cgi/boost-users
Reply | Threaded
Open this post in threaded view
|

Re: [bip] alignment/padding differs

Boost - Users mailing list
In reply to this post by Boost - Users mailing list
On 22/02/2017 11:44, Jan Boehme via Boost-users wrote:

> Hi,
>
> I'm using interprocess to share mem between different processes.
> Unfortunately these processes were built using different compiler
> version (gcc3, gcc4, gcc5) and we're seeing different alignment or
> padding dependent on which run-time was in charge of creating the mapped
> files. Of course we used the same boost version (1.60) while creating
> the binaries.
>
> Obviously ip can't handle these different layouts and mostly runs in
> indefinite loops while traversing an rb-tree not finding the first shm
> object.
>
> Are there any template parameters that enable the code to be accessed
> using different run times through forcing the same layout?

I don't think so. GCC has a flag -fabi-version, but i don't know if it's
useful in your case.

Ion
_______________________________________________
Boost-users mailing list
[hidden email]
http://lists.boost.org/mailman/listinfo.cgi/boost-users