[Xcb] problem when getting replies

[Jamey Sharp]

On Sat, Nov 11, 2006 at 08:41:03AM -0800, Ian Osgood wrote:
> A difference between Xlib and the raw protocol is that Xlib hides the  
> difference between predefined and custom atoms. The atoms "ATOM",  
> "CARDINAL", "STRING", and "WINDOW" are predefined, so you must not  
> intern them.

That's not true: you can send an InternAtom request for any atom,
including the predefined ones. You may not *want* to -- they're
predefined specifically so you can avoid the round-trip -- but it's
perfectly legal. Xlib doesn't optimize this case, either: it'll deliver
the InternAtom request for you and cache the result like any other.

Vincent, would you provide a stack trace for the segfault, after
building ecore with debugging symbols? That should tell you/us at least
what line the segfault happened on.

The reference implementation of InternAtom returns BadValue if the
only_if_exists parameter is neither 0 nor 1 (you're setting it to 0) or
BadAlloc if it can't allocate memory for a new atom (which seems
unlikely). No other errors appear possible. So I think it's unlikely
that you're getting and ignoring X errors.

Nothing jumps out at me as obviously wrong in the code you quoted. Are
you sure this is even the section of code that's segfaulting?

--Jamey
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://lists.freedesktop.org/archives/xcb/attachments/20061111/841d074d/attachment.pgp