It looks like your way indeed solves the problem as I am no longer able to observe the issue I saw earlier and I simply have not fully understood the way you rewrote the code. Thanks for cleaning it up then. Michał.
On 4 July 2018 at 09:41, Alexandre Julliard julliard@winehq.org wrote:
Michał Janiszewski janisozaur@gmail.com writes:
Hi, I would like to complain the commit mentioned by the bot is not the
commit I sent out.
My patch guarded from access to unallocated memory, while the commit
mentioned and attributed to me tests if there are any bits to set/clear and proceeds to use the memory it does not own anyway.
Please revert a1ed500836806ef46fc3eafa7ce5f8621938cbc1, I shall come up
with a test for the scenario I'm talking about.
I believe it does the same thing in a simpler way, but if you have a test demonstrating that it's still broken, please submit it and I'll fix the code.
-- Alexandre Julliard julliard@winehq.org