Re: [AD] X Multimonitor support |
[ Thread Index |
Date Index
| More lists.liballeg.org/allegro-developers Archives
]
On 2008-12-14, Thomas Fjellstrom <tfjellstrom@xxxxxxxxxx> wrote:
> On December 14, 2008, Thomas Fjellstrom wrote:
> > On December 14, 2008, Peter Wang wrote:
> > > On 2008-12-14, Evert Glebbeek <eglebbk@xxxxxxxxxx> wrote:
> > > > On 14-Dec-08, at 8:13 PM, Peter Wang wrote:
> > > > > ex_monitorinfo hangs for me. Any ideas?
> > > >
> > > > Do you have xinerama enabled in xorg.conf?
> > > > I had the sane problem (in a VM, admittedly), which was fixed when I
> > > > enabled xinerama (I didn't realise it was disabled at first).
> > > > Annoyingly enough though, the program did run properly in gdb... not
> > > > sure what went on there.
> > >
> > > I solved (I think) the problem by adding locks around the Xinerama
> > > calls, so there was probably some race condition.
> >
> > How? Its not like the function that creates the connection and then queries
> > the extensions is running in multiple threads, it makes no sense to me. To
> > me, that sounds like broken X. A single thread, making regular calls into X
> > should not lock up. Its the code that sets up the X stuff for peats sake,
> > nothing has a pointer to any of it yet.
>
> Ok, so its not locking up in the initial setup.. I'm tired.. Anyhow, I still
> don't see why its locking up when querying xinerama info later on, what could
> possibly cause it to lock up in that example?
Our X background thread would already be running. Maybe it is calling
XNextEvent or something concurrently. I don't have time to investigate now.
Stop assuming it's a broken installation. I've been running xinerama on
this system for about a year without trouble. It's not like it's the
first time we've had problems with threads on X which go away when we
put in locks (as we should).
Peter