[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Question about a potential false positive error



On 02/10/22 13:18, Felix Lechner wrote:

Hello Felix, 

Thanks for the quick reply and assistance!

> > libz is dynamically linked as a shared library:
> 
> The check looks for strings [2] with this pattern. [3]
> 
> Your build system could have linked the executable to libz in addition
> to embedding the library. Is that what happened?

Ahh, so I spent some time looking at strings instead of symbol tables
and I think I figured out where  it is coming from. Crash is a kernel
debugger which requires a lot of functionality from gdb. In fact it is
so tightly aligned with a specific version of gdb that crash compiles
its own gdb as part of the build. Crash has been tied to gdb-7.6 for
many years. However, the new upstream version of crash that I'm
packaging is a major version change form 7.x to 8.x and has been
re-synced with gdb-10.2. It looks that version of gdb does include
zlib directly where gdb-7.6 did not. So crash is embedding gdb and gdb
has statically linked zlib.

> Could Lintian be correct?

OK, so yes it looks like lintian is indeed correct. Because crash is
such a unique package with the way it uses an embedded gdb, is not a
daemon or a library and used as standalone binary for kenrel
debugging, I think it should qualify for a exemption to the policy. I
believe that is handled with a lintian override. What is the proper
procedure to request and get crash approved for an override? 

Troy

Attachment: signature.asc
Description: PGP signature


Reply to: