[rkward/frameworks] rkward/rbackend/rkwarddevice: Comment

[Thomas Friedrichsmeier]

Git commit f8515deb1ddd7394f82a57048f279a10d45a5d66 by Thomas Friedrichsmeier.
Committed on 19/05/2017 at 09:02.
Pushed by tfry into branch 'frameworks'.

Comment

M  +3    -2    rkward/rbackend/rkwarddevice/rkgraphicsdevice_stubs.cpp

https://commits.kde.org/rkward/f8515deb1ddd7394f82a57048f279a10d45a5d66

diff --git a/rkward/rbackend/rkwarddevice/rkgraphicsdevice_stubs.cpp b/rkward/rbackend/rkwarddevice/rkgraphicsdevice_stubs.cpp
index 98b8d22d..03639821 100644
--- a/rkward/rbackend/rkwarddevice/rkgraphicsdevice_stubs.cpp
+++ b/rkward/rbackend/rkwarddevice/rkgraphicsdevice_stubs.cpp
@@ -364,8 +364,9 @@ static void RKD_Close (pDevDesc dev) {
 	// it (also, as usual; in its MinGW-compiled code), we get a crash. Whatever the _exact_ reason. So what we do here is
 	// reset R's pointer, and do the free, ourselves. Fortunately, we can do this, and fortunately, the pointer is no longer
 	// needed at this point. At least not in R 3.2.3...
-        // If (or when) this breaks, we could try to just call some other device's init-routine, then hijack that device. Or out-source
-        // the RKGraphicsDevice backend init code into an R package...
+	// If (or when) this breaks, we could try to just call some other device's init-routine, then hijack that device. Or out-source
+	// the RKGraphicsDevice backend init code into an R package...
+	// Or can we use R's Calloc/Malloc, instead? -> Manual caution not to use free() (only Free()), on that, though.
 	static_cast<RKGraphicsDeviceDesc*> (dev->deviceSpecific)->rgdevdesc->dev = NULL;
 	free (dev);
 #endif