Ofc you cannot debug your own code with strace. You need gdb for that. But from what I
understand, you are causing the external libraries to crash. These calls go through the middle man…the system!
Absolutely not. I’m calling them directly. Why would I need to do a system call to call a function in the library directly linked to my executable?.. For example, when you call printf(), you call exactly printf(), not syscall(something), which then takes you to printf(). Of course, later printf will call syscall() or write(), which will call syscall(), but your program calls exatctly printf(), not syscall()…
Have you tried valgrind? It is special for memory leaks/corruptions.
I doesn’t use sources, so it can be used with 3rd party executables or
libraries. It is the main reason to develop in Linux instead of Windows:)
HTH
Nikos
:)) Such talks always sound a bit funny to me. Let me explain. You don’t have to convince me to use Linux for development instead of Windows. I’m actually used to both, for many years. Moreover, by now I’d say I don’t have a preference and just follow my customer requirements. Moreover, I’ve programmed a lot of cross-platform stuff, which had to work on both.
As of development tools, you just have to admit it, it is hard to beat Microsoft. I have heard RedHat puts lot of efforts into gcc, but somehow they didn’t manage to get optimized code (including register variables) matched well in the debugger…
valgrind… Well, of course I’m aware of valgrind and I’ve used it before. But it is too slow. When I once needed to track memory in the application which allocates/frees memory very often and totals allocated memory size to gigabytes, valgrind is not the choice, because it is too slow and you’ll end up in DAYS of running under valgrind. I remember I found some open-source memory allocation tracing library which can be injected using LD_PRELOAD and then written custom scripts to detect leaks from its output and show the leak location.
But after this lyrics I have to get back to the main point. The purpose of this discussion topic is to get information, where did Canonical put those debug symbols? But now I’m slowly getting one more question - were they ever published? I’m getting more and more suspected, that they’ve just missed it.
1 Like
This topic was automatically closed after 30 days. New replies are no longer allowed.