thanks for the suggestions. i have tried all of these suggestions and we're working on this.
is there a function in SQLite that i can call that will (in effect) verify the page cache? what i'd like to do is pepper the code that is calling into SQLite w/ verify_page_cache to attempt to figure out _where_ the corruption is occurring. btw - i expanded out the failing for loop and it appears that pNext is the culprit. thanks tom -----Original Message----- From: sqlite-users-boun...@sqlite.org [mailto:sqlite-users-boun...@sqlite.org] On Behalf Of Dan Sent: Monday, April 06, 2009 9:57 PM To: General Discussion of SQLite Database Subject: Re: [sqlite] heap corruption? On Apr 7, 2009, at 10:45 AM, Tom Broadbent wrote: > hello - > > i'm using SQLite in an embedded application (WinCE) and i'm running > into what appears to a heap corruption issue. the access violation > happens at the following line (in pcache1Fetch): > > for(pPage=pCache->apHash[h]; pPage&&pPage->iKey!=iKey; pPage=pPage- > >pNext); > > mostly i'm curious what others have experienced relating to heap > corruption and SQLite. the bad address that is causing the AV > appears to be UNICODE (ascii-range bytes, 0x00, ascii-range bytes, > 0x00). i realize this isn't much to work w/ but mostly i'm > interested in a) others' experiences w/ this type of problem, b) > hints to find the cause of the heap corruption, and c) tools that > work w/ WinCE for detecting heap corruption (i've looked into > CodeSnitch a bit). > > i'm running a vanilla version 3.6.6.2 w/ SQLITE_OS_WINCE, > SQLITE_OMIT_LOCALTIME. Compiling with both SQLITE_DEBUG and SQLITE_MEMDEBUG adds a bunch of checks to each call to malloc() and free() that SQLite makes. If SQLite is corrupting the heap itself it is likely an assert() will fail during one of these checks. The stack trace might make it clearer what is going on. Also, you could try using the SQLITE_CONFIG_HEAP feature to configure SQLite with its own private heap space by calling sqlite3_config() before any other API: static u8 heap[1024*1024]; sqlite3_config(heap, 1024*1024, 32); If it is not SQLite corrupting the heap (it could be other parts of the app), then the crash will likely occur in some other sub-system when SQLite is configured this way. Dan. > thanks > tom > > ______________________________________________________________________ > This email has been scanned by the MessageLabs Email Security System. > For more information please visit http://www.messagelabs.com/email > ______________________________________________________________________ > _______________________________________________ > sqlite-users mailing list > sqlite-users@sqlite.org > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users ______________________________________________________________________ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email ______________________________________________________________________ ______________________________________________________________________ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email ______________________________________________________________________ _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
