Fix 32-bit OS X buildbot.

I guess it's not at risk of being clobbered by a longjmp() in a way that
affects the behavior of the code, but the older GCC on that buildbot
doesn't do the dataflow analysis to figure it out.

Change-Id: I770380e2a22d00aeccf5937203bc70968712d37f
Reviewed-on: https://code.wireshark.org/review/11611
Reviewed-by: Guy Harris <guy@alum.mit.edu>
This commit is contained in:
Guy Harris 2015-11-06 11:34:11 -08:00
parent bebc416db9
commit fdeab5a22f
1 changed files with 2 additions and 1 deletions

View File

@ -1395,7 +1395,8 @@ dissect_dcerpc_guid(tvbuff_t *tvb, packet_info *pinfo, proto_tree *tree, void *d
int (*volatile sub_dissect)(tvbuff_t *tvb, int offset, packet_info *pinfo, proto_tree *tree, dcerpc_info *di, guint8 *drep) = NULL;
proto_item *pi, *sub_item;
proto_tree *sub_tree;
guint length, reported_length;
volatile guint length;
guint reported_length;
volatile gint offset = 0;
tvbuff_t *volatile stub_tvb;
tvbuff_t *volatile payload_tvb = NULL;