Issue
What were you trying to do that didn't work?
I was trying to debug a C++ library with the system tools (not a toolset).
GDB only showed me the raw data for an STL container instead of the pretty-printed ones.
A quick test with a 5 line program shows that it worked fine in that setting.
My actual setting is a swig-wrapped c++ library. I'm importing the library into python and stopping in the C++ code. The pretty-printers were missing there.
One more observation: if I start gdb on a simple test program and enter `info pretty-printers` before running the program, the pretty-printers aren't loaded. They do appear to be loaded after the program starts running.
Is it possible that the hook to load the pretty-printers isn't triggered when the main debugged program is python instead of c++?
Please provide the package NVR for which bug is seen:
(watnlp) [jlquinn@bertdev01 engines]$ rpm -q gdb
gdb-8.2-20.el8.x86_64
How reproducible:
It showed up for me debugging a python package swig-wrapped c++ library. I had to manually import the pretty printers in .gdbinit.
But it works fine with a standard c++ program on the command line.
If I add the following to .gdbinit, then I get the pretty printers within the swig c++ library too:
python
import sys
sys.path.insert(0, '/usr/share/gcc-8/python')
from libstdcxx.v6.printers import register_libstdcxx_printers
register_libstdcxx_printers (None)
end
Steps to reproduce
Create a swig-wrapped c++ library
Import into a conda python (I used 3.10.13)
Run the python script in gdb
break in the c++ library
in gdb, print a vector<float>
Expected results
(gdb) p x
$1 = std::vector of length 5, capacity 5 = {1, 2, 3, 4, 5}
Actual results
(gdb) p/r x
$2 = {<std::Vector_base<float, std::allocator<float> >> = {_M_impl = {<std::allocator<float>> = {<_gnu_cxx::new_allocator<float>> =
{<No data fields>}
, <No data fields>},
_M_start = 0x613eb0, _M_finish = 0x613ec4, _M_end_of_storage = 0x613ec4}}, <No data fields>}
I used p/r here to duplicate it in a quick 5 line program.
