> find_solib: Can't read pathname for load map: Input/output error
EIO? Wow, you rarely, if ever, see *that* on a UNIX-flavored system....
(I think I last saw it ages ago at Sun, and it really *was* an I/O error
from the disk. Of course, those were the days when a disk drive was the
size of a laser printer and had maybe *three hundred megabytes* on it if
it was a really big disk, and the really big disks used an SMD
interface....)
It's probably some weird system call it's making on "/proc" or
something.
Unfortunately, that means it doesn't say what GTK+ routines are on the
call stack; however, this *might* be a GTK+ bug that somebody else also
discovered, and for which they posted a fix to one of the GTK+ mailing
lists.
What version of GTK+ do you have? I forget which version introduced the
bug, but it's in 1.2.6 (I think it goes back to 1.2.3 or so); I've
attached the patch that guy posted.
*** gtktext.c.dist Fri Sep 3 15:20:39 1999
--- gtktext.c Tue Dec 14 01:17:32 1999
***************
*** 903,908 ****
--- 903,909 ----
g_return_if_fail (GTK_IS_TEXT (text));
text->freeze_count++;
+ undraw_cursor (text, FALSE);
}
void
***************
*** 917,922 ****
--- 918,924 ----
recompute_geometry (text);
gtk_widget_queue_draw (GTK_WIDGET (text));
}
+ draw_cursor (text, FALSE);
}
void