Chris Frey wrote:
> it breaks compatibility with compiled
> programs (they must be relinked at the minimum).
You have to anyway.
Anytime you touch something in the header files of a C++ template
library, you need to rebuild. The only time you can get away with
upgrading just the .so file with a C++ template library is if only the
bodies of functions in .cc files changed.
Looking back on the versions I've released, every single one has changed
header files.
> in the description that caught my imagination: It stated that mysql++
> may in the future support other databases as well, besides mysql.
I observe that that dream has existed since Kevin Atkinson was the
library's maintainer. Six years ago, it was transferred to MySQL AB's
maintainership, which effectively killed that dream.
The library name isn't the only thing that has changed recently. You'll
notice that sqlplus.hh got removed, replaced by mysql++.hh. I also plan
to rename the sqlplusint directory in the tarball to src, to match GNU
naming conventions.
> I don't see any reason why that future support couldn't happen.
I'm not going to rename it on a "maybe". That just leads people to
believe, like you, that things are different than they really are.
I'm not against the idea the idea of reviving the dream, but until I see
some concrete plans and real patches, the library should be named for
what it is, not what it might be someday if someone gets ambitious.
