[CLN-list] CLN 1.2.1 -- silent ABI change?

[Richard B. Kreckel]

Dear Alexei,

Alexei Sheplyakov wrote:
> I've tried to install (and use) CLN 1.2.1. My programs linked with CLN
> 1.2.0 stopped working due to undefined symbols:

Oops.

Indeed: <http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=473494>.

> $ ginsh
> ginsh: symbol lookup error: /usr/local/lib/libginac-1.3.so.2: undefined symbol: _ZN3cln23cl_double_to_DF_pointerERKNS_11dfloatjanusE
> 
> I think the SONAME should be changed to indicate incompatible changes.
> 
> diff --git a/configure.ac b/configure.ac
> index 051223e..9188ad9 100644
> --- a/configure.ac
> +++ b/configure.ac
> @@ -91,8 +91,8 @@ dnl * if any functions/classes have been added, increment CL_AGE,
>  dnl * if backwards compatibility has been broken, set CL_AGE to 0.
>  dnl $(CL_CURRENT):$(CL_REVISION):$(CL_AGE) results in
>  dnl libcln.so.$(CL_CURRENT)-$(CL_AGE)
> -CL_CURRENT=5
> -CL_REVISION=1
> +CL_CURRENT=6
> +CL_REVISION=0
>  CL_AGE=0
>  dnl make substitutions
>  AC_SUBST(CL_CURRENT)

<slab_my_forehead>
Oh, no! Oh, no, no nooooo!!! This entire library versioning scheme sucks 
so big time. I've read all the docs and it didn't help because I missed 
that change. It is theoretical mumbo-jumbo unless one has an actual ABI 
testsuite and uses it on a regular basis. This is especially true for 
C++ libraries.
</slab_my_forehead>

Anyway, I have two options now:

a) Release CLN-1.2.2 with changed soname, although the ABI did not 
change wrt. CLN-1.2.1.

b) Just ignore it and release CLN-1.2.2 as libcln.so.5. Seriously. I am 
under the impression that the amount of trouble incurred by just 
re-building GiNaC for all distros is dwarfed by the trouble induced by a 
soname bump. In Debian, it always takes many weeks till we've hinted 
depending packages into the testing distribution after a soname bump.

I'll make up my. Until then, feel free to lobby me.

Best wishes
    -richy.
-- 
Richard B. Kreckel
<http://www.ginac.de/~kreckel/>