Is It possible to somehow wrap this in some test case?
When just looking at this hunk the patch looks wrong. The old length correction is ok for the code above it, but then immediately after that the data pointer is updated. (At that point also the buffersize must be adjusted!) Perhaps using a stringbuffer in this way is not really the appropriate operation? Bert Sent from Windows Mail From: astie...@apache.org Sent: Saturday, November 23, 2013 9:46 PM To: comm...@subversion.apache.org Author: astieger Date: Sat Nov 23 20:46:50 2013 New Revision: 1544878 URL: http://svn.apache.org/r1544878 Log: Ensure full distribution description is printed in svn --version -v * subversion/libsvn_subr/sysinfo.c (stringbuf_split_key): fix buffer length calculation Approved by: breser Modified: subversion/trunk/subversion/libsvn_subr/sysinfo.c Modified: subversion/trunk/subversion/libsvn_subr/sysinfo.c URL: http://svn.apache.org/viewvc/subversion/trunk/subversion/libsvn_subr/sysinfo.c?rev=1544878&r1=1544877&r2=1544878&view=diff ============================================================================== --- subversion/trunk/subversion/libsvn_subr/sysinfo.c (original) +++ subversion/trunk/subversion/libsvn_subr/sysinfo.c Sat Nov 23 20:46:50 2013 @@ -285,7 +285,7 @@ stringbuf_split_key(svn_stringbuf_t *buf key = buffer->data; end = strchr(key, delim); *end = '\0'; - buffer->len = 1 + end - key; + buffer->len -= 1 + end - key; buffer->data = end + 1; svn_stringbuf_strip_whitespace(buffer);
