cgdb terminal handling problems

Arthur Norman
Fri Jan 17 16:40:00 GMT 2020

I have a trivial C++ program, which was tested using
    g++ -g hello.cpp -o hello
    cgdb ./hello
    break main

What gets displayed at that stage looks like the following where some of 
the long strings of escaped have been wrapped. There are funny characters
marking "Line 5" which is where the breakpoint was.

1: #include <iostream>
3: int main(int argc, char *argv[])
4: {
5├>    std::cout << "Hello" << std::endl;
6:     return 0;
7: }
Breakpoint 1,[67X[67C[?25h[?25l[80X[80C[?25h[?25lmain[76X[76C[?25h[?25l ([78X[78
[80X[80C[?25h[?25larg-value *[?25h[?25l
0xffffcc40[70X[70C[?25h[?25l)[79X[79C[?25h[?25l at[77X[77C[?25h[?25lhello.cpp[71

and you can see the mess of terminal-control escapes that unhelpfully appear. 
My terminal there is then not as responsive as I would like so progressing is 
hard. I can go ^Z to exit and then use kill to terminate the cgdb session.
The above trasncript was collected by using COPY&PASTE from the terminal 
screen. But all those "[80C" etc etc appear as characters on my terminal 
just as such. When I go "echo $TERM" I see the default "xterm" and SHELL 
is /bin/bash - all that is just as per the default when cygwin was 

To ensure that this was not a matter of other interfering applications I 
installed a totally fresh Windows 10 and then (perhaps apart from virtualbox 
tools) only cygwin. For Cygwin I only install cgdb, gdb and g++, using the 
latest non-testing version of each.

But having done that I tried other versions og gdb & cgdb without joy.

This sort of behaviour in cgdb seems to have arisen some time in the last few 
weeks. If I back off to cygwin 3.1.1-1 I still see trouble but with 
3.0.7-1 things appear to behave better for this. However with a larger 
real program I just got
(gdb) step
__wrap__Znam (sz=80000000) at 
(gdb) s
A problem internal to GDB has been detected,
further debugging may prove unreliable.

Since that may be a metter of using an old cygwin I will wait before I try 
to produce a version of that worth reporting!

But this looks like an underlying cygwin change upsetting gdb or cgdb???

[In some code of my own I was bitten because something has fairly recently 
changes and the Windows call GetConsoleScreenBufferInfo now succeeds in a 
cygwin program running under mintty while it used not to - I had used that test 
to judge whether I could or should use Windows-mode console handling on 
Unix-style... I sort of wonder if this is a variant on my problem there?]

-------------- next part --------------

Problem reports:
Unsubscribe info:

More information about the Cygwin mailing list