[Bug gdb/31669] wrong "info symbol" output, symbol shadowing, __dso_handle

["dominik.b.czarnota+bugzilla at gmail dot com" <sourceware-bugzilla@sourceware.org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=31669

--- Comment #5 from Disconnect3d <dominik.b.czarnota+bugzilla at gmail dot com> ---
(In reply to Andreas Schwab from comment #4)
> Since __dso_handle does not have a size and no other hint is available, gdb
> can only assume that it extends until the next symbol.

How does this logic really works? Is this that some kind of map entry is
created for each memory address in such a case?

Is there any way to disable this behavior or tell GDB to assume a size for
unknown sizes to be sizeof(pointer) or something like this?

Eventually, I think we want to have Python API for the two use cases I noted
earlier, so that we can "get symbol for given address" and "find symbols around
addresses".

For now I guess we can workaround this issue in Pwndbg by ignoring returned
symbols with "+" but this may be buggy...

I bet that if there is a symbol with known size, it can also appear/resolve in
GDB with symbol+offset as it is done here, heh.

-- 
You are receiving this mail because:
You are on the CC list for the bug.