Re: Epiphany memory corruption problem

[mcatanzaro gnome org]

On Tue, Sep 3, 2019 at 6:45 AM, Bastien Nocera <hadess hadess net> 
wrote:
> Maybe release 3.32.x as a 3.34 instead if that's the problem?

We absolutely could, but have no way to know where the memory 
corruption was actually introduced.

It could be an Epiphany regression (not unlikely). But it could also be 
a WebKit regression (more likely) or even a GTK regression (less 
likely). Re-releasing Epiphany 3.32 won't solve anything unless we're 
sure it's an Epiphany problem, and I think the chances of that are only 
~40% (just guessing).

Since the memory corruption is not reproducible, we can't easily 
determine where the problem is without actually catching it in 
valgrind/asan.

There's actually multiple problems, at least one in the UI process, and 
a different memory corruption problem in the web process (which is 
less-likely to be Epiphany's fault, but still could be).

Really, without being able to catch the problem with valgrind/asan, 
it's too hard to know.

Michael