-thread-info in lldbmi does not conform to protocol. Should end with current thread id

[jacdavis]

Summary

-thread-info in lldbmi does not conform to protocol. Should end with current thread id as described here: https://sourceware.org/gdb/onlinedocs/gdb/GDB_002fMI-Thread-Commands.html#GDB_002fMI-Thread-Commands

When printing all threads, the current thread id should be printed afterwards.

Example:
-thread-info

^done,threads=[
{id="2",target-id="Thread 0xb7e14b90 (LWP 21257)",
   frame={level="0",addr="0xffffe410",func="__kernel_vsyscall",
           args=[]},state="running"},
{id="1",target-id="Thread 0xb7e156b0 (LWP 21254)",
   frame={level="0",addr="0x0804891f",func="foo",
           args=[{name="i",value="10"}],
           file="/tmp/a.c",fullname="/tmp/a.c",line="158"},
           state="running"}],
current-thread-id="1"
(gdb)

[zturner]

Hi Jackson,

Two suggestions:

1. Please clang-format the patch. Let me know if you need help getting this working.
2. use C++11 threads instead of pthreads. This makes the tests portable to Windows. Even if the test still fails on Windows for other reasons, at least it's closer.

[jacdavis]

Updated to use C++ 11 threads. Clang-format tried to format the diff in the llvm style. Since this is all lldb code, is that actually what is expected? None of the braces matched etc....

[zturner]

We have an lldb style file in llvm/tools/lldb, it should be picked up
automatically if you run it from the lldb directory. I'm not sure though if
it's based on the folder you're in when you run clang-format or by
searching up the tree from the target file until it finds a rule file

[jacdavis]

Ran formatting tool with lldb format. pthread linker option is still required for C++11 threads so left it in. New test still disabled for windows.

[jacdavis]

last diff missed lldb-mi changes

[zturner]

Looks fine to me. TBH I haven't seen any of the lldb-mi owners around in months, so I'm not sure if they're still active

[ChuckR]

closed with http://reviews.llvm.org/rL265858

[abidh]

In D18880#395530, @zturner wrote:

Looks fine to me. TBH I haven't seen any of the lldb-mi owners around in months, so I'm not sure if they're still active

Sorry for late reply. I have been busy lately but keep an eye on lldb-mi patches.

packages/Python/lldbsuite/test/tools/lldb-mi/threadinfo/Makefile

+LEVEL = ../../../make
+
+LDFLAGS=-pthread
+
+CXX_SOURCES := test_threadinfo.cpp
+
+include $(LEVEL)/Makefile.rules

packages/Python/lldbsuite/test/tools/lldb-mi/threadinfo/TestMiThreadInfo.py

+"""
+Test lldb-mi -thread-info command.
+"""
+
+from __future__ import print_function
+
+import lldbmi_testcase
+from lldbsuite.test.decorators import *
+from lldbsuite.test.lldbtest import *
+from lldbsuite.test import lldbutil
+
+class MiThreadInfoTestCase(lldbmi_testcase.MiTestCaseBase):
+
+    mydir = TestBase.compute_mydir(__file__)
+
+    @skipIfWindows # pthreads not supported on Windows
+    @skipIfFreeBSD # llvm.org/pr22411: Failure presumably due to known thread races
+    def test_lldbmi_thread_info(self):
+        """Test that -thread-info prints thread info and the current-thread-id"""
+
+        self.spawnLldbMi(args = None)
+
+        # Load executable
+        self.runCmd("-file-exec-and-symbols %s" % self.myexe)
+        self.expect("\^done")
+
+        self.runCmd("-break-insert ThreadProc")
+        self.expect("\^done")
+
+        # Run to the breakpoint
+        self.runCmd("-exec-run")
+        self.expect("\^running")
+        self.expect("\*stopped,reason=\"breakpoint-hit\"")
+
+        self.runCmd("-thread-info")
+        self.expect("\^done,threads=\[\{id=\"1\",(.*)\},\{id=\"2\",(.*)\],current-thread-id=\"2\"")
+
+        self.runCmd("-gdb-quit")
+   

packages/Python/lldbsuite/test/tools/lldb-mi/threadinfo/test_threadinfo.cpp

+#include <cstdlib>
+#include <iostream>
+#include <thread>
+
+using namespace std;
+
+void
+ThreadProc()
+{
+    int i = 0;
+    i++;
+}
+
+int
+main()
+{
+    thread t(ThreadProc);
+    t.join();
+
+    return 0;
+}

tools/lldb-mi/MICmdCmdThread.h

     bool m_bThreadInvalid; // True = invalid, false = ok
     VecMIValueTuple_t m_vecMIValueTuple;
     const CMIUtilString m_constStrArgNamedThreadId;
+
+    // mi value of current-thread-id if multiple threads are requested
+    bool m_bHasCurrentThread;
+    CMICmnMIValue m_miValueCurrThreadId;
 };

tools/lldb-mi/MICmdCmdThread.cpp

 // Throws:  None.
 //--
 CMICmdCmdThreadInfo::CMICmdCmdThreadInfo()
-    : m_bSingleThread(false)
-    , m_bThreadInvalid(true)
-    , m_constStrArgNamedThreadId("thread-id")
+    : m_bSingleThread(false),
+      m_bThreadInvalid(true),
+      m_constStrArgNamedThreadId("thread-id"),
+      m_bHasCurrentThread(false)
 {
     // Command factory matches this name with that received from the stdin stream
     m_strMiCmd = "thread-info";

         }
     }
 
+    // -thread-info with multiple threads ends with the current thread id if any
+    if (thread.IsValid())
+    {
+        const CMIUtilString strId(CMIUtilString::Format("%d", thread.GetIndexID()));
+        CMICmnMIValueConst miValueCurrThreadId(strId);
+        m_miValueCurrThreadId = miValueCurrThreadId;
+        m_bHasCurrentThread = true;
+    }
+
     return MIstatus::success;
 }
 

         ++it;
     }
 
-    const CMICmnMIValueResult miValueResult("threads", miValueList);
+    CMICmnMIValueResult miValueResult("threads", miValueList);
+    if (m_bHasCurrentThread)
+    {
+        CMIUtilString strCurrThreadId = "current-thread-id";
+        miValueResult.Add(strCurrThreadId, m_miValueCurrThreadId);
+    }
     const CMICmnMIResultRecord miRecordResult(m_cmdData.strMiCmdToken, CMICmnMIResultRecord::eResultClass_Done, miValueResult);
     m_miResultRecord = miRecordResult;