I have the problem that was discussed several times here: the application runs when started directly from shell and does not run when I try to start it within the debugger in the same shell. Running inside GDB produces the "cannot open shared object" error.
I've read all the posts and did all the suggestions:
I setup LD_LIBRARY_PATH manually and verified that my application runs and ldd -r passes without errors
I setup solib-search-path and solib-absolute-prefix inside GDB to the same value as LD_LIBRARY_PATH and '/' respectively. All the paths are absolute
I ran GDB with strace to see where GDB looks for the required shared libraries and found that it ignores the list of directories from LD_LIBRARY_PATH / solib-search-path
What can I do?
It's GDB 7.11.1 with RHEL 7
I have the problem
If you do, you've done a very poor job of describing what your problem actually is.
I ran GDB with strace to see where GDB looks for the required shared libraries and found that it ignores the list of directories
GDB doesn't look for any shared libraries until it gets notified by the runtime loader that some shared library has been added to the process.
If you didn't attach to a running inferior, then GDB wouldn't look for any libraries, and you've likely arrived at a completely wrong conclusion.
P.S. GDB doesn't use LD_LIBRARY_PATH at all, and setting either solib-search-path or solib-absolute-prefix is usually not needed either (unless you are doing remote debugging, or analyzing a core that came from a different machine).
Update:
the application runs when started directly from shell and does not run when I try to start it within the debugger
In 99.9% of the cases I've seen, this happens because ~/.bashrc or some such resets LD_LIBRARY_PATH inappropriately. GDB starts a new shell to run the application in, and if that (non-interactive) shell resets the environment, the application may behave differently inside vs. outside of GDB.
The correct solution is to make ~/.bashrc (or whatever shell startup file is appropriate for your shell) to do nothing when running non-interactively.
Related
I'm in the middle of a large debugging project, and every time I start running gdb I have to type b 253.
It would be really nice if I could set my run script so that gdb loads with that breakpoint already set.
To be more explicit: Here are the contents of run.csh:
gdb --args path/to/program arg1 arg2
Can I modify this so that, once I run it, I can just type r and the program breaks on line 253?
Yes. Read documentation of gdb.
You can extend GDB. You can have Canned Sequences of Commands.
You can define or use extensions in Python, in Guile. See also this.
(you might need to recompile GDB itself from source, since sadly not all usual gdb are configured with Guile support)
You can have your .gdbinit file (read about startup files and command files). Btw you might prefer to break in function names, not in line numbers there. Read more about specifying locations.
Actually, many large projects have some .gdbinit (perhaps generated) in their source repository.
Be sure to use a recent version of GDB. The latest one (in March 2018) is GDB 8.1
I'm trying to lunch an application from Eclipse CDT and as I read everywhere, I set LD_LIBRARY_PATH to the directory containing the shared library:
However, when I try to run this configuration:
/home/luca/Dropbox/HKUST/CloudCache/cloudcache/CloudCache/make/CloudCache: error while loading shared libraries: libvl.so: cannot open shared object file: No such file or directory
Just to clarify: I'm 100% sure that libvl.so is in that path, in fact when I try to run the application from command line it works perfectly.
Why this happens?
I wonder if it is being reset somewhere else. You can also set environment variable in the Launch Configuration if you are using that. There are two places in there, one where you have an 'Environment' tab, and another where you choose your 'Build Configuration' which could itself point somewhere else.
I've just been trying to get this to work for the first time in eclipse (Linux 64), and did exactly what you showed and it worked.
A bit late, but still can be useful. You have the option "Append environment to native environment" checked on your screenshot. Very probably you have this variable redefined there. It was my case.
I am trying to debug, with gdb, a plugin to a program. I saw a question earlier that indicated that the directory command in gdb might help. I thought it would help because, when I try to set a breakpoint within the code of the plugin, I get the error that says: "No source file named..." It didn't seem to do anything when I used the directory command with the source path structure. Any ideas? Thanks.
"No source file named..."
GDB will not be able to set a breakpoint until your plugin is actually loaded into the inferior (being debugged) process.
Use (gdb) info shared command to check whether your plugin is already loaded or not.
If it isn't, you can set a "deferred" breakpoint (GDB should be asking you whether you want to set such a breakpoint, assuming you have the default set confirm on setting).
If your plugin is already loaded and visible in info shared output, then you haven't built your plugin with debug info. Rebuild it with -g, and you should be able to set breakpoints in it.
I have a core dump generated by 2 applications -> /usr/bin/python and /usr/bin/app1.
I know the dump can be analyzed by
gdb /path/to/app /path/to/core
but is there a way to include both applications in the arguement?
I did try gdb '/usr/bin/python /usr/bin/app1' core.xxx but that doesnt seem right.
Any suggestions?
I think you cannot achieve what you want with a single invocation of gdb. But you could run gdb twice, in different terminal windows. I did that more than once, and it works quite well (except of course that your own brain could be slightly overloaded).
a gdb process can debug only one single program, with one single debugged process or (for post mortem debug) one single core file.
And a given core file is produced by abnormal termination of one single process (not several), so I don't understand your question.
Apparently, you have a crash in some execution of python probably augmented by your faulty C code. I suggest having a debuggable variant of Python, perhaps by installing the python3-all-dbg package or something similar, then use gdb on it. Of course, compile your C code plugged into Python with debugging enabled. Perhaps you violated some invariant of the Python garbage collector.
I am trying to build a Windows service with MingW. It need thread safe exceptions, so I added the linker flag -mthreads. The application works fine from the command-line, but when I try to start it from services.msc, the 1054 error ("The service did not respond to the start or control request in a timely fashion") is raised. The service starts if I re-build it without the -mthreads flag. How can I get this working with -mthreads?
I suspect -mthreads is bringing in a dependency on a DLL, and that DLL is not on the path when it's running as a service. In my cygwin environment, if I compile a trivial program with "-mno-cygwin -mthreads", I get a dependency on MINGWM10.DLL, which certainly wouldn't be on the path when running as a service. If I try running it with no PATH set, it crashes as it starts to load (and leaves a turd in the Application event log).
I'd be bringing up your exe in Dependency Walker (http://www.dependencywalker.com) to see what you're loading at load-time, and check your Windows Event Log to see if there are any hints there. You're probably going to need to put a copy of the DLLs it needs alongside the executable.
You need mingwm10.dll in the working directory or in [edit: system, not per user] PATH, because C++ programs compiled with -mthread option have that dependency. If you're pretty sure exception will never be thrown by your code nor propogate through your stack, use -fno-exception instead of -mthread to resolve the dependency.
I wonder if you can debug it when it runs as a service. There must be something spooking your program when service host runs it. Perhaps try to attach a debugger to svchost.exe, at least you can see what modules are loaded and maybe which exception causes the crash.
Is your application even starting up at all? Put a call to OutputDebugString (or equivalent) at the start of your main function to see if it even gets that far. (Grab DbgView from SysInternals if you don't have it already.)
If it doesn't get that far, we start checking for the obvious: is it a matter of the application not finding the runtime DLL? It could be that you have the regular runtime in its PATH, but it can't find the MT version. That could explain the behaviour you describe. You may need to copy the MT runtime or update the PATH accordingly.