* publish a performance time before starting, so that others wait for us with valid values: this enables running with larger audio buffers
* don't publish a performance time before having a non null old realtime (at least two buffer exchanges)
* don't dereference config pointer before checking it is not null
* if we don't find an input for a buffer, we also recycle the buffer
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27578 a95241bf-73f2-0310-859d-f6bbb57e9c96
even in not unlikely situations.
* GrepWindow::_SelectFilesInTracker() was still leaking entry_refs in the
success code path.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27577 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Removed unused fClean member
* Got rid of fSkipSetFlags member by simply calling the BView::SetFlags()
directly where fSkipSetFlags was supposed to prevent the custom
implementation.
* Added some debugging facilities.
* Used the layout friendly constructors of BControl where appropriate.
* Used B_FOLLOW_ALL for the child text input, it should be more correct.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27576 a95241bf-73f2-0310-859d-f6bbb57e9c96
it had been removed from the list. Now, I removed the flag again (but kept
the flags field for now), and solved in a way that doesn't have this problem.
* This might have a positive effect on #2682.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27574 a95241bf-73f2-0310-859d-f6bbb57e9c96
VMware this doesn't produce very good results though. In my tests I only
got twice as many hits as with 1 ms, although it should be roughly ten
times as many. Might work better on real hardware.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27572 a95241bf-73f2-0310-859d-f6bbb57e9c96
yield any hits anyway. The size 0 functions are hand-coded assembly
functions (syscalls etc.).
* Increased the requested sampling interval to 1 ms. Should be
specifiable per command line.
* Improved the output when done profiling a thread. The hit functions
are sorted and printed in a table.
* Some cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27571 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Moved the rock ridge attribute parsing code to a separate function.
* Removed superfluous malloc() vs. realloc() code.
* Checked allocations in the rock ridge parsing code.
* Cleanup, renamed some variables, structures, functions, etc.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27569 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Factored a few functions out of strace that can be reused.
* Added the beginnings of a "profile" command. It is very much work in
progress, though it is already able to profile the main thread of a
program.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27533 a95241bf-73f2-0310-859d-f6bbb57e9c96
still pretty much work in progress.
* Introduced init_thread_debug_info() which is used instead of
clear_thread_debug_info() when the thread is created. The latter
requires former initialization.
* user_debug_thread_deleted() is now already invoked in thread_exit(),
not in the undertaker.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27531 a95241bf-73f2-0310-859d-f6bbb57e9c96
disabled, as long as a fault handler is installed for the thread. It
allows user_memcpy() to be invoked with interrupts disabled -- in this
case it will simply fail, when the address is valid, but not mapped yet.
This is a more desirable behavior for debug facilities, though in all
other cases it is likely a bug. We should probably introduce a thread
flag to discriminate these situations.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27530 a95241bf-73f2-0310-859d-f6bbb57e9c96
