Hi Sean-- Thanks for the review!
On Sat 2019-09-28 08:53:38 -0700, Sean Whitton wrote: > I read through the script and I'm a bit apprehensive about the > complexity involved in talking to the IMAP server, because it renders > imap-dl significantly more complicated than anything else in > mailscripts: > >> + # these objects are weirdly structured. i don't know why >> + # these trailing close-parens show up. so this is very >> + # ad-hoc and nonsense Yeah, this is a bit of a shame. I would welcome someone with more IMAP knowledge than me making it clearer what's happening here or suggesting a more rigorous handling of the returned tuples. > In general, it seems to me that mailscripts should fairly strongly > favour simplicity over features, so that the scripts in mailscripts can > be useful fixed points around which to build more complex (and perhaps > more fragile) systems. > > Of course, imap-dl *is* simple in terms of its features, so I hope we > can mitigate the risks of this IMAP complexity somehow. right, i'm aiming for simple features. that means masking some of the complexity. > Would you consider writing an integration test suite for imap-dl? I > would like it to be possible for us to be completely confident that > imap-dl won't ever lose any user messages. I would love more generally to have an integration test suite that can work for multiple IMAP clients (i want a test suite for thunderbird+enigmail too, for example!) To do this, we need an imap server that can receive and deliver messages, though, right? I've sketched out a design for this over at https://salsa.debian.org/debian/test-daemons -- it's a lot of work! But it seems like it could pay off in the future, too. Once it's done, i'd be happy to add an integration test that uses imap-dl this way. Do you think we need this to be done in order for mailscripts to adopt imap-dl? that would make me kind of sad, though of course i can continue to use imap-dl without its adoption by mailscripts. > I'm not keen on introducing imap-dl as a getmail replacement. > Hopefully, in the future, the majority of imap-dl's users won't be > people who ever used getmail. It would be good for the docs to > introduce imap-dl without reference to getmail, and then explain later > how it can be a getmail replacement. That makes sense to me. do you want to propose a patch to the documentation? --dkg
signature.asc
Description: PGP signature
