https://bugs.kde.org/show_bug.cgi?id=492663
--- Comment #1 from Paul Floyd <pjfl...@wanadoo.fr> --- (In reply to mh from comment #0) > but the whole ordeal is incredibly fragile and making assumptions that I'm > sure future linkers and/or dynamic loaders will keep breaking. I think that the code is overly cautious. There are two possible errors. One would be false positive reading debuginfo for a binary that is getting loaded other than as part of execution, link loading or dlopen. I don't imagine user code mmapping binaries is at all common, so a false positive isn't a disaster. Then there are all the false negatives where a binary really does get loaded but we fail to read the debuginfo. That's much more common. For the loading of the tool and the guest exe I think we should always trigger the read of debuginfo. The tool can't have failed to load and if the guest exe fails to load it will terminate very quickly. Loading guest shared libraries is trickier. The problem is that we are counting mmaps to see when the library has been fully loaded. There are a lot of things that can go wrong there. -- You are receiving this mail because: You are watching all bug changes.
