---------------------
PatchSet 10347 
Date: 2007/12/29 14:22:53
Author: adri
Branch: s27_adri
Tag: (none) 
Log:
Add more items to my TODO list.

Members: 
	ADRIAN_TODO:1.1.2.18->1.1.2.19 

Index: squid/ADRIAN_TODO
===================================================================
RCS file: /cvsroot/squid-sf//squid/Attic/ADRIAN_TODO,v
retrieving revision 1.1.2.18
retrieving revision 1.1.2.19
diff -u -r1.1.2.18 -r1.1.2.19
--- squid/ADRIAN_TODO	29 Dec 2007 05:36:13 -0000	1.1.2.18
+++ squid/ADRIAN_TODO	29 Dec 2007 14:22:53 -0000	1.1.2.19
@@ -111,17 +111,30 @@
 
   (in any case, bring sanity to the disk code!)
 
+* add a "header size" swap meta type (if one doesn't exist already) and then
+  modify the store client code to read in -that- much data (in multiple reads if needed)
+  before parsing the headers. Yes, at this point we can handle >4k reply status+headers..
+
+Stuff to do before fleshing out stmemAppendRef()
+
 * Audit all the storeClientRef() users (again!) and see if they're able to handle buffers
   larger than SM_PAGE_SIZE. The stmem stuff can now handle arbitrary-sized buffers and
-  these should be read/written as much as possible.
+  these should be read/written as much as possible. clientSendMoreData(), for example,
+  has an assert() to force this.
 
-* .. then, add a "header size" swap meta type (if one doesn't exist already) and then
-  modify the store client code to read in -that- much data (in multiple reads if needed)
-  before parsing the headers. Yes, at this point we can handle >4k reply status+headers..
+* Flesh out the region referencing and quickly convert everything doing buf_t manipulations
+  directly (so libbuf/String, src/store.c::storeAppendRef(), all of src/stmem.c,
+  parts of src/http.c) are using the region stuff instead.
+
+  - stuff like the src/http.c routines that use a const char *buf internally could probably
+    benefit from being turned into array offset manipulations. It might hurt performance
+    a little but we can always convert back the critical performing code into pointer
+    arithmetic where required. This pass is about tidying things up, remember..
+
+And urgently, now that http.c is using buf_fill()!
 
 * Hm, what else? Ah, buf_fill() does a read() where it should do an FD_READ_METHOD
   and update the comm counters. Look at moving buf_fill() into src/ so it can access
   the FD_READ_METHOD and associated gunk (don't make buf depend on comm and comm depend
   on buf; have a "netbuf" library which depends on both. No cyclic dependencies!)
 
-