Re: Purify report for fs-test

Branko ÄŒibej <brane@xbc.nu> writes:

> Right, here's the list of problems found by Purify when running fs-test:
>
> [W] UMR: Uninitialized memory read in svn_fs__bdb_open_uuids_table
> Reading 1 byte from 0x0013f091 (1 byte at 0x0013f091 uninitialized)
> Address 0x0013f091 points into a thread's stack
> Thread ID: 0x4a0
> Error location
> svn_fs__bdb_open_uuids_table [uuids-table.c:78]
> value.size = sizeof (buffer) - 1;
>
> apr_uuid_get (&uuid);
> => apr_uuid_format (buffer, &uuid);

valgrind on Linux doesn't report anything here. How does 0x0013f091
compare to buffer and &uuid? I note that APR's Windows UUID
implementation is completely different from the Unix one, could the
error be in inside apr_uuid_format? It's hard to see what could be
wrong with the line that calls apr_uuid_format.
                 
> BDB_ERR (uuids->put (uuids, 0, &key, &value,
> DB_APPEND | SVN_BDB_AUTO_COMMIT));
> svn_fs_create_berkeley [fs.c:520]
> svn_test__create_fs [fs-helpers.c:97]
> reopen_trivial_transaction [fs-test.c:232]
> do_test_num [svn_tests_main.c:153]
> main [svn_tests_main.c:289]
> mainCRTStartup [crtexe.obj]

-- 
Philip Martin
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org