summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorRobert James Kaes <rjkaes@users.sourceforge.net>2001-08-30 16:53:36 +0000
committerRobert James Kaes <rjkaes@users.sourceforge.net>2001-08-30 16:53:36 +0000
commitc825db8d562ab6b5c7e9249a90296748a98021ba (patch)
tree6c5e7004872eda7e56ebf1c3618dcbd45e9fcc6e
parentfc94a56f56821957da7cc4c8974dae87881bb514 (diff)
downloadtinyproxy-c825db8d562ab6b5c7e9249a90296748a98021ba.tar.gz
tinyproxy-c825db8d562ab6b5c7e9249a90296748a98021ba.zip
More fixes.
Diffstat (limited to '')
-rw-r--r--ChangeLog24
1 files changed, 24 insertions, 0 deletions
diff --git a/ChangeLog b/ChangeLog
index 953abe3..83967e0 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,27 @@
+2001-08-30 Robert James Kaes <rjkaes@flarenet.com>
+
+ * src/dnscache.c: Removed the insert_data() function and replaced
+ the "replace" logic in dns_insert() with a call to
+ ternary_replace(). This fixes the segmentation fault problem which
+ I introduced when I "tried" to fix the memory leak. Also changed
+ the LOCKing be around the entire dnscache() again since I still
+ need to work out how locking should be done when accessing the
+ ternary tree.
+
+ * src/ternary.c (ternary_insert_replace): Renamed ternary_insert()
+ to ternary_insert_replace() and added a extra argument. The
+ function can now "replace" data already in the tree without
+ causing a memory leak. Added two DEFINES to make coding easier:
+ ternary_insert() and ternary_replace() which both call this
+ function with the right arguments.
+
+ * src/utils.c: Removed xstrstr() since it was only used in once
+ place, and could be safely replaced with strstr. I can't even
+ remember why we had this function to begin with.
+
+ * src/reqs.c (compare_header): Removed the call to xstrstr() since
+ it's been removed from the source.
+
2001-08-28 Robert James Kaes <rjkaes@flarenet.com>
* src/log.c (log_message): Handle a debug enabled compile better.