In GDB I can get nice formatting of a POD struct when I use a plain print command:
(gdb) print l[0]
$9 = {rq = 10, eq = 0 }
I would like to use a printf expression in a gdb user-defined function that formats this variable in the same way.
What is the correct syntax to do this? The "natural" way to do this fails
(gdb) printf "%s", l[0]
Value can't be converted to integer.
There have been a few questions where the answer involves putting the output of a GDB command into a convenience variable. Rather than writing Python every time, as in my first answer, here's a convenience function named $execute that does this. You can put it in the file ~/execute.py, and load it into GDB using the source command.
import gdb
class Execute(gdb.Function):
"""Execute the first argument as a command.
Returns a string (C array of char).
Optional second argument chomp=True will remove trailing newlines.
In general, the first argument should be a C string,
and the second argument a C bool."""
def __init__(self):
super(Execute, self).__init__("execute")
def invoke(self, arg, chomp=True):
ret = gdb.execute(arg.string(), to_string=True)
if chomp:
ret = ret.rstrip('\n')
return ret
Execute()
And here's how to use it in your case:
(gdb) source ~/execute.py
(gdb) set $firstelt = $execute("output l[0]")
(gdb) printf "In thread %d, first element is %s.\n", $_thread, $firstelt
In thread 1, first element is {rq = 10, eq = 0}.
The GDB output command does the same thing as print, but it omits the $n = at the beginning.