I think you can't deny MailAddress represent the same concept in both mailet and protocols/smtp .
Thus, it seems logical to have one object to represent it. I agree the way I proposed to share this code might not be optimal. Maybe some common dependency might be better. Note : MailetMailAddressAdapter converts org.apache.james.protocols.smtp.MailAddress into org.apache.mailet.MailAddress. Both MailAddress classes are copied and paste from one another and have started to diverge. One is tested not the other. Le 02/05/2016 à 16:22, Matthieu Baechler a écrit : > > > On 04/29/2016 08:32 AM, Benoit Tellier wrote: >> Moreover, we should take advantage of this task to factorize code : >> >> We have two MailAddress classes : >> >> org.apache.mailet.MailAddress >> org.apache.james.protocols.smtp.MailAddress; >> >> These two class have almost the same content (copy and paste ?), and >> tests are only written for the mailet one. >> >> In my opinion, we must make smtp protocol depend on mailet api to reuse >> MailAddress object across our projects and fight code duplication. >> > IMO you overestimate the value of sharing code. > > Why SMTP should depend on mailet from a business point of view ? > --------------------------------------------------------------------- To unsubscribe, e-mail: server-dev-unsubscr...@james.apache.org For additional commands, e-mail: server-dev-h...@james.apache.org