On 04/01/2013 01:15 PM, Bk wrote:
I have updated make file with "TCC = armv7l-timesys-linux-gnueabi-gcc -g
-DSQLITE_OS_UNIX=1 -DSQLITE_DISABLE_LFS=1 -I. -I${TOP}/src
-I${TOP}/ext/rtree"
The test still fails with DSQLITE_DISABLE_LFS=1 with same error
sysfault-2.setup... Ok
ftruncate: : File too large
ftruncate: : File too large
ftruncate: : File too large
ftruncate: : File too large
sysfault-2.1-vfsfault-transient.1...
Expected: [0 ok]
Got: [1 {nfail=1 rc=1 result=disk I/O error}]
Below is the debugger screen with value of "sizeof(off_t)" . The values
seems different 4 in ts_ftruncate and 8 in robust_ftruncate.
That'll be the reason the test is failing. For whatever reason the
test_syscall.c file is being compiled with a different definition
of off_t (and probably other stuff) than os_unix.c. Find out why
and you'll know how to fix the test.
It will be (I imagine) something to do with the pre-processor magic
at the top of os_unix.c. Perhaps something included there that isn't
included in test_syscall.c
I would have thought it would be the SQLITE_DISABLE_LFS though. You
did a clean build with the new Makefile?
Dan.
_______________________________________________
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users