Bug#1125722: dpkg-gensymbols: error if symbols file is too lax

Otto Kekäläinen <[email protected]> Sun, 18 Jan 2026 17:57:03 -0800
Newsgroups gmane.linux.debian.devel.dpkg.bugs
Message-ID <CAOU6tAApLY0QpOZDAsHCVcc8Vc00j0X4b5Vm98bG9a4tdSTq5Q__26167.0913243265$1768787963$gmane$org@mail.gmail.com>
> > > The main purpose of the symbols file is to assist in accurate dependency
> > > generation. That it can also be used to check whether upstreams (that
> > > might not have solid processes to handle library changes), have not
> > > messed up the their symbol modifications, it's a plus.
> >
> > I didn't quite understand what is the benefit with a symbols file that
> > is all wildcards. The package might as well not have any symbols file
> > as far as I see if it is totally void of actual symbol names.
>
> As I mentioned above, the main purpose is to assist with accurate
> dependency generation, at the symbol level to be able to get the lowest
> possible versioned dependencies instead of having to use the version
> for the last ABI change in the library (with shlibs), but please see
> dpkg-shlibdeps(1), where the role of the symbols file (and the simpler
> shlibs file) is explained in more detail.

Both the dpkg-gensymbols and dpkg-shlibdeps man pages explain that the
purpose of the .symbols file is to track new - and in particular
removed - symbols.

I didn't get any sensible explanation when and how a .symbols file
would detect any ABI changes if the contents in its entirety are just
wildcards that can match any symbol.

I do however not expect an answer from you, as you already replied 3
messages without actually answering, so I will not just leave the
topic as is.