robert burrell donkin wrote: > > On 3/6/07, Stefano Bagnara <[EMAIL PROTECTED]> wrote: > > robert burrell donkin ha scritto: > > > i'm a torque expert so here are a couple of questions > about the schema... > > > > That's cool ;-) > > Well, Joachim is the man for your answers, but I try with > my knowledge > > about... > > > > > 1. there don't seem to be any indexes. is this intentional? does > > > torque automatically add indexes? > > > > It is not intentional: it was the first easy step to make > it working. > > i'll try some indexes locally and see if they make a > noticeable difference > > > > 2. the header is a VARCHAR length 1024. this seems a > little short to > > > me. is there a good reason for this choice? > > > > 1024 == A random number for the first tests > > The rationale is: IIRC the RFC does not require limits on the header > > size, so we are stuck to a arbitrary number that will not > fit any message. > > Maybe 1024 will fit 99% of messages, but maybe 2-3K will > work better: > > those figures accord with my experience > > > messages with 1MB headers will need a special handling anyway > > true > > ATM a message with a header that is too big is simply rejected. even > if only 1% are effected then perhaps we shouldn't be just rejecting > these mail since they are RFC compliant. any ideas about better ways > to handle this case? > > - robert
Why not use the maximum VARCHAR size supported by the underlying DBMS as an overridable default limit? When the limit is exceeded I cannot see a better alternative than rejecting the message with "452 Requested action not taken: insufficient system storage". The "bounce" seems perfectly reasonable behaviour as every system ultimately has hard limits. We do need to validate that we can withstand a flood of such huge headers. Cheers -- Steve --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
