<rdar://problem/11744001>
Fixed an issue where GDB servers that don't support the thread suffix could get registers states incorrectly due to an incorrect assumption that the current register thread (set using the "Hg%x" packet) will always be cached between runs. Now we clear the cached register thred when the process is resumed.
git-svn-id: https://llvm.org/svn/llvm-project/lldb/trunk@159603 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/source/Plugins/Process/gdb-remote/GDBRemoteCommunicationClient.cpp b/source/Plugins/Process/gdb-remote/GDBRemoteCommunicationClient.cpp
index cf527e2..83d2843 100644
--- a/source/Plugins/Process/gdb-remote/GDBRemoteCommunicationClient.cpp
+++ b/source/Plugins/Process/gdb-remote/GDBRemoteCommunicationClient.cpp
@@ -381,6 +381,7 @@
StringExtractorGDBRemote &response
)
{
+ m_curr_tid = LLDB_INVALID_THREAD_ID;
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet (GDBR_LOG_PROCESS));
if (log)
log->Printf ("GDBRemoteCommunicationClient::%s ()", __FUNCTION__);