MeshCentral/MeshAgent
1
0
Fork 0
mirror of https://github.com/Ylianst/MeshAgent synced 2026-01-04 17:43:45 +00:00
Code Issues Releases Wiki Activity

Fixed dispatch loop that can happen on process exit

Browse Source
This commit is contained in:
Bryan Roe
2020-04-26 13:03:48 -07:00
parent 19508fa79e
commit 1db2b400d1
1 changed files with 5 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
microscript/ILibDuktape_ChildProcess.c
View File

@@ -34,6 +34,9 @@ typedef struct ILibDuktape_ChildProcess_SubProcess
void *subProcess;
void *chain;
ILibProcessPipe_Process childProcess;
#ifdef WIN32
int dispatchFlags;
#endif
ILibDuktape_readableStream *stdOut;
ILibDuktape_readableStream *stdErr;
@@ -108,7 +111,7 @@ void ILibDuktape_ChildProcess_SubProcess_ExitHandler(ILibProcessPipe_Process sen
if (!ILibMemory_CanaryOK(p)) { return; }
#ifdef WIN32
if (duk_ctx_context_data(p->ctx)->apc_flags == 0)
if (duk_ctx_context_data(p->ctx)->apc_flags == 0 && p->dispatchFlags == 0)
{
// This method was called with an APC, but this thread was running an unknown alertable method when it was interrupted
// So we must unwind the stack, and use a non-apc method to re-dispatch to this thread, becuase we can't risk
@@ -118,6 +121,7 @@ void ILibDuktape_ChildProcess_SubProcess_ExitHandler(ILibProcessPipe_Process sen
// We had to do the APC first, becuase otherwise child_process.waitExit() will not work, becuase that method is blocking
// the event loop thread with an alertable wait object, so APC is the only way to propagate this event
p->exitCode = exitCode;
p->dispatchFlags = 1;
Duktape_RunOnEventLoop(p->chain, duk_ctx_nonce(p->ctx), p->ctx, ILibDuktape_ChildProcess_SubProcess_ExitHandler_sink1, NULL, p);
}
#endif