[PATCH 2/2] Do not call release_pack_memory in malloc wrappers when GIT_TRACE is used

[Johannes Sixt <j6t@kdbg.org>]

This avoids a potential race condition when async procedures are
implemented as threads where release_pack_memory() can be called from
different threads without locking under memory pressure.

Signed-off-by: Johannes Sixt <j6t@kdbg.org>
---
 It is worthwhile to have this in np/malloc-threading (and in maint or
 master) even if js/async-thread is not in master, yet, because on
 Windows we already have async procedures implemented in threads.

 trace.c |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)

diff --git a/trace.c b/trace.c
index 4229ae1..1e560cb 100644
--- a/trace.c
+++ b/trace.c
@@ -25,6 +25,10 @@
 #include "cache.h"
 #include "quote.h"
 
+void do_nothing(size_t unused)
+{
+}
+
 /* Get a trace file descriptor from GIT_TRACE env variable. */
 static int get_trace_fd(int *need_close)
 {
@@ -72,6 +76,7 @@ void trace_printf(const char *fmt, ...)
 	if (!fd)
 		return;
 
+	set_try_to_free_routine(do_nothing);	/* is never reset */
 	strbuf_init(&buf, 64);
 	va_start(ap, fmt);
 	len = vsnprintf(buf.buf, strbuf_avail(&buf), fmt, ap);
@@ -103,6 +108,7 @@ void trace_argv_printf(const char **argv, const char *fmt, ...)
 	if (!fd)
 		return;
 
+	set_try_to_free_routine(do_nothing);	/* is never reset */
 	strbuf_init(&buf, 64);
 	va_start(ap, fmt);
 	len = vsnprintf(buf.buf, strbuf_avail(&buf), fmt, ap);
-- 
1.7.1.64.ga1799.dirty