Re: [AD] X11 unresponsiveness

[Chris <kcat@xxxxxxxxxx>]

Peter Wang wrote:
> For .so's: if you make a symbol global, it doesn't automatically break 
> the ABI.  It only matters if the user makes a binary that references the 
> symbol (directly).

This is good.. this is actually how I thought all DLLs were supposed to 
work, later to find out that it's not the case.

Although, still.. playing devil's advocate.. ABI compatibility is ABI 
compatibility, whether or not the program uses just the official API.

It's a judgement call here, although I personally don't think we should 
continue working on 4.0 unless some major issue shows up. 4.1 is 
different enough already and we should focus on 4.2.

> Regardless of what the docs warn, I refuse to use the polling mechanisms 
> (and if I could, I would deprecate all the relevant parts of the API).  
> Whatever is done to make it work better without polling would be good 
> for people like me.  No, it's not just laziness...

This is pretty much what happens when you use an interrupt-type method 
in a threaded environment. AFAIK, the signal version of the lib should 
only be used when pthreads is not an option for exactly this reason, and 
I think the docs should say that if it doesn't already.

- Kitty Cat