This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc 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] |
On 01/30/14 08:55, Andreas Jaeger wrote:
That would, of course, be a disaster. I'm going to trust Carlos & his team that this won't happen. Given glibc's ability to maintain ABI compatibility through the years, I'm not terribly concerned here.The GLIBC 2.19 will be the 2.19 ABI and the GLIBC-2.17 will be the 2.17 ABI. We want to insure that PPC64LE ELF2 binary build on GLIBC-2.17 will run on a 2.18 or 2.19 or later GLIBC. We do not expect a binary build on 2.19 will run on GLIBC 2.18 or 2.17. But this the normal definition.An ABI is more than just the default symbol version. I'm asking for a stability promise of the ABI that will be in glibc. What if the implemented 2.17 glibc used by Carlos is by accident incompatible to the 2.19 ABI? Which ABI will then be broken to get those compatible again.
I actually worry more about how GCC/LD play into the ABI issues -- there's plenty of risk with a newly defined ABI like this that corner cases which affect the ABI definition are still lurking out there. But that's another discussion for another list :(
Jeff
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |