On Mon, Apr 15, 2013 at 06:45:04PM -0700, Jonathan Nieder wrote:

Jeff King wrote:

quoted

--- a/run-command.c
+++ b/run-command.c

@@ -603,7 +603,7 @@ static NORETURN void die_async(const char *err, va_list params)
 {
 	vreportf("fatal: ", err, params);
 
-	if (!pthread_equal(main_thread, pthread_self())) {
+	if (!running_main_thread()) {
 		struct async *async = pthread_getspecific(async_key);
 		if (async->proc_in >= 0)
 			close(async->proc_in);

@@ -614,6 +614,19 @@ static NORETURN void die_async(const char *err, va_list params)
 
 	exit(128);
 }
+
+int running_main_thread(void)
+{
+	return pthread_equal(main_thread, pthread_self());
+}

Would it make sense to do something like

	return !main_thread_set || pthread_equal(...);

in case someone tries to call this before the first start_async call?

I considered that, but assumed that pthread_equal would always return
false against the zeroed out static main_thread. Mostly because we
appeared not to be bothering with such a check already.  But actually,
in the existing code, it is only checked from die_async, which is only
put into place when we _do_ start a thread, so we would never hit this
code path without main_thread being set.

So yes, a check probably would make sense. However, I think I'm
convinced that we should drop this in favor of using thread-local
storage for the "dying" counter. So this patch can just go away.

Thanks for reviewing this version, though.

-Peff