GNOME Bugzilla – Bug 60460
tests fail unless adding usleep on IRIX
Last modified: 2008-06-08 16:14:38 UTC
I recently built gnome-vfs/1.0.2 for SGI IRIX 6.5. I found that to make self-tests pass I had to increase the usleep() times in the test code. With the following changes applied "make check" passed: --- ./test/test-async-cancel.c Mon Jan 29 22:17:45 2001 +++ ../gnome-vfs-1.0.2/./test/test-async-cancel.c Wed Sep 12 17:37:26 2001 @@ -107,7 +107,7 @@ } for (i = 0; i < MAX_THREAD_WAIT; i++) { - usleep (1); + usleep (100); gtk_main_iteration_do (FALSE); if (*wait_for_it) { return TRUE; @@ -126,7 +126,7 @@ } for (i = 0; i < MAX_THREAD_WAIT; i++) { - usleep (1); + usleep (10000); gtk_main_iteration_do (FALSE); if (gnome_vfs_backend_get_job_count () == 0) { return TRUE; @@ -145,7 +145,7 @@ } for (i = 0; i < MAX_THREAD_WAIT; i++) { - usleep (1); + usleep (100); if (gnome_vfs_backend_get_job_count () == 0) { return TRUE; } @@ -163,7 +163,7 @@ } for (i = 0; i < MAX_THREAD_WAIT; i++) { - usleep (1); + usleep (100); gtk_main_iteration_do (FALSE); if (get_used_file_descriptor_count () == 0) { return TRUE;
These required time increases seem indicative of a timing bug of some sort in the cancel code. Or maybe the async jobs just take too long. Needs further investigation.
Actually this also happened to me on a linux x86 box
*** Bug 110907 has been marked as a duplicate of this bug. ***
Mass resetting target milestone to clean up gnome-vfs milestones. Please complain if I removed a milestone which should have been kept. Sorry for the spam
Adding the portability keyword.
Do you still see these performance or timing issue?
No feedback to Christian's questions, hence closing as INCOMPLETE. Please reopen if this is still an issue.