Saw this on the berlin list.
---
'Cloning and the reprogramming of DNA is the first serious step in
becoming one with God.'
- Scientist G. Richard Seed
---------- Forwarded message ----------
Date: 22 Oct 2000 02:24:13 +0300
From: Momchil Velikov <[EMAIL PROTECTED]>
To: [EMAIL PROTECTED]
Subject: [Berlin-design] [RFA] interrupted select() again
Hi,
There are still problems in GGIConsole::next_event(). When
ggiEventSelect() is interrupted, `mask' and `rfdset' variables are
cleared, so the next invocation of ggiEventSelect() deadlocks. This
patch initializes the variables `mask' and `rfdset' inside the
ggiEventSelect() loop.
Regards,
-velco
Index: GGI.cc
===================================================================
RCS file: /cvsroot/berlin/Berlin/src/Berlin/GGI.cc,v
retrieving revision 1.20
diff -u -r1.20 GGI.cc
--- GGI.cc 2000/09/30 00:24:43 1.20
+++ GGI.cc 2000/10/21 22:54:53
@@ -158,16 +158,23 @@
{
Prague::Trace trace("GGI::Console::next_event");
ggi_event event;
- ggi_event_mask mask = ggi_event_mask (emKeyboard | emPtrMove | emPtrButtonPress |
emPtrButtonRelease);
+ ggi_event_mask mask;
ggi_event_mask move_mask = ggi_event_mask (emPtrMove);
int input = fileno(stdin);
Prague::FdSet rfdset;
- rfdset.set(_wakeupPipe[0]);
- if (_autoplay) rfdset.set(input);
+
int nfds = -1;
- do nfds = ggiEventSelect(_visual, &mask, rfdset.max() + 1, rfdset, 0, 0, 0);
+ do
+ {
+ mask = ggi_event_mask (emKeyboard | emPtrMove | emPtrButtonPress |
+emPtrButtonRelease);
+ rfdset.set (_wakeupPipe [0]);
+ if (_autoplay)
+ rfdset.set (input);
+ nfds = ggiEventSelect (_visual, &mask, rfdset.max() + 1, rfdset, 0, 0, 0);
+ }
while (nfds == -1 && errno == EINTR);
+
if (nfds == 0)
{
// no input from the outside world
