Le mardi 12 février 2013 17:17:24, Santiago Vila a écrit : > El 12/02/13 16:25, Thomas Preud'homme escribió: > > That said, I'd really like to understand what's the objective with this > > check since it's only performed at request time and no longer after. Is > > this for memory consumption, spam, something else? How is this goal > > achieved by enforcing a check at request time only? > > I don't know for sure as I didn't write that check. > > What you say about spam is one possible reason to keep some kind of size > check, as the -request address may be abused to spam people.
The check don't add much protection to the subscription system. Only the mailing list system receive -request so the spam at this level is quite useless. On the other hand, if one has tweaked its spamming system to try to subscribe, it can certainly also do a correct subscription before sending spam to the whole mailing list. Also, spam are often not so big and could fit in 4096 bytes. It would much more effective to detect URLs. > > The meaning of the check, based on what it actually does, is something > like "messages longer than this length are obviously/probably wrong and > should not be processed automatically". > > If you have evidence (and I think you have) that nowadays it is "not so > obvious", I'm happy to increase the size, yes. > > To summarize on this side: Do you think 16384 would be enough? Yes, 16384 would be enough. Increasing it to this value would solve my problem (I already did on my own setup). Best regards, Thomas
signature.asc
Description: This is a digitally signed message part.
