Tim Düsterhus

> In your gdb with the core ? I'm pretty sure that ipc_fd[1] is higher than maxsock which might be around 200 or so. You're wrong: ``` (gdb) p proc_self->ipc_fd[1]...

When downgrading back to 2.4 I'm seeing: ``` Jul 14 09:25:46 *snip* haproxy[19710]: [NOTICE] (19710) : Reloading HAProxy Jul 14 09:25:46 *snip* haproxy[19710]: [WARNING] (19710) : Failed to connect to...

As production is running fine and the error only appeared ONCE during the upgrade, I'll abstain from applying any patches, I wouldn't be able to test them anyway.

> `efl_type` is first masked with `0xf0` so the lower bits are always 0. I'm not sure if I'm clear here but the important things is that there is no...

> `-` Can you please fill in the `haproxy -vv` field properly?

> I'm tempted to revert the patch. That's fine. My code reports can generally be considered low priority, no need to break something that already works fine and just could...

> This is somewhat niche and I understand if it's not something that would be fixed within this package This is a bug in the server / ASP.NET. The body...

> Might it be worth to build some kind of optional filter for the post method to prevent this kind of problem ? I believe this is best solved by...

Possibly related: #2933

> Given this work, wouldn't it make sense to _not_ merge this PR? I believe there is _some value_ in having the interfaces and exceptions available, e.g. to create a...