From: Alex Musolino <alex@musolino.id.au>
Date: Tue, 19 Nov 2019 02:08:13 +0000 (+1030)
Subject: rio: fix bug causing short reads of wctl files
X-Git-Url: https://git.lizzy.rs/?a=commitdiff_plain;h=4d4107b38547c7f9f135cc22d3a29fcbad4a8c34;p=plan9front.git

rio: fix bug causing short reads of wctl files

Previously, reads of wctl files would return one byte less than
requested as the returned string must be null terminated.  Now we pass
the actual size of the allocated buffer to the handler, which is large
enough to accommodate a trailing partial rune and terminating null
byte.
---

diff --git a/sys/src/cmd/rio/xfid.c b/sys/src/cmd/rio/xfid.c
index 3985bea00..b8e4be84a 100644
--- a/sys/src/cmd/rio/xfid.c
+++ b/sys/src/cmd/rio/xfid.c
@@ -670,9 +670,9 @@ xfidread(Xfid *x)
 		}
 		c1 = crm.c1;
 		c2 = crm.c2;
-		t = emalloc(cnt+UTFmax+1);	/* room to unpack partial rune plus */
+		pair.ns = cnt+UTFmax+1;	/* room for partial rune and null byte */
+		t = emalloc(pair.ns);
 		pair.s = t;
-		pair.ns = cnt;
 		send(c1, &pair);
 		recv(c2, &pair);
 		fc.data = pair.s;