On 2/28/07, Joe Orton <[EMAIL PROTECTED]> wrote:
The caller could already pass in APR_BUFFERED in the mode parameter to apr_sdbm_open(), AFAICS.
Well, that doesn't help apr_dbm_* - which doesn't permit such flags to be passed. I only realized after I committed that we even had a bypass mechanism for apr_dbm_* which allows SDBM to be utilized - ugly!
I have very little trust in the buffered I/O code, people have been very fixing basic bugs in it all through 1.2.x, and there are a couple more reported in bugzilla already. So I wouldn't call this low-risk. http://issues.apache.org/bugzilla/show_bug.cgi?id=40963 could easily affect sdbm.
What would you suggest? The performance implications of not doing buffered reads just *kills* our server - so we need to do something and adding buffering lessens the load quite dramatically. In our situation, we couldn't care less about modifying the file - we only care to optimize the read-only case - and I believe that code is just fine and stable. Though I guess I'd prefer we fix the problems with buffering if they do exist. Would you be concerned if we added APR_BUFFERED to APR_DBM_DBMODE_RO for sdbm? -- justin
