Band-aid fix for test hangs:

base/process_util_posix.cc

Index: base/process_util_posix.cc
diff --git a/base/process_util_posix.cc b/base/process_util_posix.cc
index 6d41fe6cd66c034a26ce80848af7bdc76051eabd..a40d812607e6ea0eedf2d12addb2c01a6ebb5c41 100644
--- a/base/process_util_posix.cc
+++ b/base/process_util_posix.cc
@@ -141,16 +141,23 @@ void StackDumpSignalHandler(int signal, siginfo_t* info, ucontext_t* context) {
   if (debug::BeingDebugged())
     debug::BreakDebugger();
 
-  DLOG(ERROR) << "Received signal " << signal;
-  debug::StackTrace().PrintBacktrace();
+  char buf[1024];
+  size_t len;
+
+  // NOTE: Even |snprintf()| is not on the approved list for signal
+  // handlers, but buffered I/O is definitely not on the list due to
+  // potential for |malloc()|.
+  len = static_cast<size_t>(
+      snprintf(buf, sizeof(buf), "Received signal %d\n", signal));
+  ignore_result(write(STDERR_FILENO, buf, std::min(len, sizeof(buf) - 1)));
+
+  // TODO(phajdan.jr): Re-enable backtrace printing after fixing the hang.
+  // debug::StackTrace().PrintBacktrace();
 
   // TODO(shess): Port to Linux.
 #if defined(OS_MACOSX)

[Scott Hess - ex-Googler, 2012/10/30 23:47:59]

I would prefer to have the PrintBacktrace() enabled on OSX, since it doesn't use
tcmalloc anyhow.

I agree that we should be using backtrace_symbols_fd() or whatever it is.  But
that will lose out on demangling on Linux, I think.

   // TODO(shess): Port to 64-bit.
 #if ARCH_CPU_32_BITS
-  char buf[1024];
-  size_t len;
-
   // NOTE: Even |snprintf()| is not on the approved list for signal
   // handlers, but buffered I/O is definitely not on the list due to
   // potential for |malloc()|.