Julian Sikorski wrote:
> Is that the same as bumping soname whenever new symbols are introduced?
No, it's not. It's better. Symbol versioning is per symbol, not per shared
object, so only the new symbols get new versions and it's all backwards-
compatible. Unfortunately, not all targets support s
Julian Sikorski wrote:
> W dniu 05.01.2013 18:37, Miloslav Trmač pisze:
>> On Sat, Jan 5, 2013 at 11:38 AM, Julian Sikorski
>> wrote:
>>> a bug was recently filed against gnumeric [1], in which the program
>>> refused to run due to user not having updated his whole package set and
>>> a symbol be
W dniu 05.01.2013 18:37, Miloslav Trmač pisze:
> On Sat, Jan 5, 2013 at 11:38 AM, Julian Sikorski wrote:
>> a bug was recently filed against gnumeric [1], in which the program refused
>> to run due to user not having updated his whole package set and a symbol
>> being missing in the older libgsf h
On Sat, Jan 5, 2013 at 11:38 AM, Julian Sikorski wrote:
> a bug was recently filed against gnumeric [1], in which the program refused
> to run due to user not having updated his whole package set and a symbol
> being missing in the older libgsf he had installed.
> How are such issues supposed to b
Hi list,
a bug was recently filed against gnumeric [1], in which the program
refused to run due to user not having updated his whole package set and
a symbol being missing in the older libgsf he had installed.
How are such issues supposed to be handled? Surely, manually introducing
a versioned
