AddressSanitizer: SEGV on unknown address (3.0.0-beta0 8a225b6be0, civ2civ3, qt)
The problem goes away when I re-compile without ASAN or UBSAN. Since I always compile with *SAN and usually do a 'git pull; make' about once a day, I suspect the problem was introduced in one of the most recent commits.
Can you reproduce this consistently from current S3_0, then? Can you test when (which commit) it started?
Reply To cazfi
Can you reproduce this consistently from current S3_0, then?
Yup.
Reply To cazfi
Can you test when (which commit) it started?
Already in the process of doing that. Well, I'm busy re-reading my git cheat-sheets and notes to work out how to rewind my checked-out copy to a particular commit.
Strangely, it wasn't that menu.cpp commit. Since it takes me more than an hour to recompile, it is going to take me a while to find it. But should I bother? The problem has disappeared in 4aba0703a1. The only version (that I've found) that displays the problem is 8a225b6be0.
Reply To chippo
Since it takes me more than an hour to recompile, it is going to take me a while to find it. But should I bother? The problem has disappeared in 4aba0703a1. The only version (that I've found) that displays the problem is 8a225b6be0.
If it comes and goes like that, there's no point in wasting time in testing with which commits it happens. That won't show the root cause that is there even in the commits that seemingly work.
Reply To cazfi
If it comes and goes like that, there's no point in wasting time in testing with which commits it happens. That won't show the root cause that is there even in the commits that seemingly work.
Agreed. If it rears its ugly head again, I can put some more effort into it.
I got this SEGV:
And there was a core, so:
Is it possible to attach files to a ticket while you are creating it? If so, I am not working out how. :-(