Fix various cases where re-init could get things blocked. Remove the
duplicate storage of the active paths; now we always check the primary
data, and hence it can’t be out of sync.
Also remove the obsolete persistent cache code.
Fixes some of the issues discussed in:
https://sourceforge.net/p/flightgear/codetickets/2308/
Further improvements still to come, especially to retry on a better
schedule for intermittent connections.
The deadlock is possibly caused by the thread not being joined, or because of an implementation difference with phtreads.
So this fix adds a call to release the background thread when terminating it and then also joins the thread to await for the completion of the background thread.
As before this works fine under Win32 (x64)
An exclusive thread is one that is suited to being used where the thread needs to be activated at a certain point, and also that the code activating the thread may also need to wait for thread completion.
Example of this is the new background garbage collection for Nasal. The thread will be activated at the end of the frame processing and at the start of the next frame the thread can be awaited - thus allowing the thread to work in parallel with the rendering.
Since there's no external lock, it was impossible to check the queue was non-empty before a pop, without a race condition. To avoid this, tolerate an empty queue and return a default constructed value.
For all source directories outside of simgear/scene,
which really depends very hard on osg, avoid using osg
classes. This should reenable the use of some basic and
scenegraph independent parts of simgear without the
the need for osg.
Reimplement the threading stuff using either pthreads
or win32 threads. These simple classes should help us
to stay osg independent for simgears core classes.
- turn four #include paths from the "foo" form to <foo>
The quotes form is normally only used for headers with path relative
to the including file's path, though the standard doesn't strictly
mandate this. This is consistent with the rest of sg, it makes the
code's intent clearer and helps to find headers. (And it's a few
milliseconds faster, too.)
the MSVC and MipsPro warning stuff).
As a result of this patch, simgear/sg_traits.h can be deleted. So can SGCMath.h,
but I'll do that separately.
There is one more 'mechanical' change to come - getting rid of SG_USING_STD(X),
but I want to keep that separate from everything else. (There's another mechnica
l change, replacing <math.h> with <cmath> and so on *everywhere*, but one step a
t a time)
- don't unlock an already unlocked mutex (Someone wanted to be on
the safe side with this, but the result is undefined and makes
pthread_mutex_destroy fail. Reference: manpage for
pthread_mutexattr_gettype/The Open Group[1]: "Attempting to
unlock a mutex of this type which is not locked results in
undefined behaviour.")
- re-enabled all subsystem destructors again (this has been disabled
because fgfs hung on exit, due to the mutex destroy failure from
above.)
Reference:
http://www.opengroup.org/onlinepubs/007908799/xsh/pthread_mutexattr_gettype.html
- new FSF address
- removed a few hundred trailing spaces
- fixed a few $Id$ lines
- copied two license headers from *.hxx files to their respective
*.cxx counterparts
- added two test aps to .cvsignore
- don't unlock an already unlocked mutex (Someone wanted to be on
the safe side with this, but the result is undefined and makes
pthread_mutex_destroy fail. Reference: manpage for
pthread_mutexattr_gettype/The Open Group[1]: "Attempting to
unlock a mutex of this type which is not locked results in
undefined behaviour.")
- re-enabled all subsystem destructors again (this has been disabled
because fgfs hung on exit, due to the mutex destroy failure)
* Use "const string&" rather than "string" in function calls when appropriate.
* Use "const Point3D&" instead of "Pint3D" in function calls when appropriate.
* Improved course calculation in calc_gc_course_dist()
* Safer thread handling code.
Vassilii Khachaturov:
Dont use "const Point3D&" for return types unless you're absolutely sure.
Erik Hofman:
* Use SGD_(2)PI(_[24]) as defined in simgear/constants.h rather than
calculating it by hand every time.
gcc 3.4 has changed the rules for unqualified template name lookup. This
affects SGQueue.hxx. The changes I've made are backwards compatible with
earlier gcc versions. Everything else compiles pretty much okay except for a
few warnings. The resultant executable seems a bit faster too.
- Removed some old cruft.
- Removed some support for older versions of automake which technically was
correct, but caused the newer automakes to squawk warnings during an
initial sanity check (which isn't done very intelligently.)
NOTE: this fix is technically not correct for older version of automake.
These older version use the variable "INCLUDES" internally and could have
them already set to an important value. That is why we were appending
our values to them. However, newer versions of automake don't set this
value themselves so it is an error to append to a non-existant variable.
We seem to "get away" with overwriting the value on older versions of
automake, but if you have problems, consider upgrading to at least
automake-1.5.
- automake-1.4 sets default values for INCLUDES which we can't
overwrite.
- automake-1.5 renames this to DEFAULT_INCLUDES and leaves INCLUDES
open for the developer to use.
Thus for automake-1.4 we are forced to 'append' to INCLUDES and in
automake-1.5 we can just set the value to whatever we like.
Unfortunately, the behaviors of the two versions are mutually
incompatible.
The solution I am committing now works for both versions but
automake-1.5 generates a lot of spurious warning messages that are
annoying, but not fatal.
