[Patch, libgfortran] Obvious ABI bugfix committed

[Janne Blomqvist]

FX Coudert wrote:
> When I saw the title of that mail, before reading it, I thought "ouch, I 
> forgot to ask Thomas to add his new pack/unpack symbols to gfortran.map, 
> and Janne is fixing it". But I now see that I have committed yet another 
> mistake :(
> Thomas, could you add your symbols to the gfortran.map file as well? (in 
> GFORTRAN_1.1) 

IIRC Thomas patch didn't require any new symbols to be exported. There 
is the existing exported symbols internal_pack/unpack or something like 
that, which in turn contains a switch statement that selects the correct 
routine.

> Janne, is there a way we could test this so that it shows 
> up during regtesting, and we avoid forgetting it again and again?

Well, in this case I just diffed the trunk and 4.3 gfortran.map files.

I wrote the one below in another mail about the F2003 I/O stuff:

The C++ library has some ABI tests in 
libstdc++-v3/testsuite/libstdc++-abi , perhaps libgfortran could do 
something similar? Another option could be to run the 4.3 testsuite 
using the libgfortran from 4.4 (maybe as easy as copying the 4.4 
libgfortran into the directory where 4.3 is installed?).

-- 
Janne Blomqvist