Inline crate tag-generator to crate imap-types
Closed this issue · 8 comments
Hm... I can't remember why we said we want to have it it in imap-types
... Could as well be integrated into imap-next
? Any memories? :D
imap-next
is not the origin of typeTag
and is not using theTagGenerator
itself- People might want to use
TagGenerator
even if they don't useimap-next
- An own crate would be optimal but it would cause too much overhead
- Given that we only have the crates
imap-next
,imap-codec
andimap-types
, the latter is the most appropriate place IMO
imap-next
is not the origin of typeTag
and is not using theTagGenerator
itself
I think the question could be different: Do people who use imap-next
want to use tag-generator
, too? imap-client
does, for example. imap-next
kinda asks to implement some procedure. Up until this point, we don't care about unique tags so much because it's too low-level.
- People might want to use
TagGenerator
even if they don't useimap-next
Yeah maybe... but I guess this argument is true for everything...
- An own crate would be optimal but it would cause too much overhead
Yeah. I don't think it's worth it...
- Given that we only have the crates
imap-next
,imap-codec
andimap-types
, the latter is the most appropriate place IMO
I would exclude imap-codec
. Not sure about imap-next
or imap-types
. I feel that in imap-types
it would be kind of "feature creep". Not so much in imap-next
, though, because it's functionality that people using imap-next
probably want.
Also, having it in imap-types
is kind of providing a "blessed" method to create tags. Maybe we want this...
I think the question could be different: Do people who
use
imap-nextwant to use
tag-generator, too?
imap-clientdoes, for example.
imap-next` kinda asks to implement some procedure.
If someone implements a small client or script using the Fragmentizer
, they also want to generate Tag
s. Exposing the TagGenerator
to Python might also be an option. In both cases imap-codec
or imap-types
would be better.
I would exclude imap-codec.
imap-codec
is already opinionated how messages should be encoded. TagGenerator
fits to that.
Okay, let's put it into imap-codec
then? 👯 I think for me it's most important to not let it into imap-types
because I really want imap-types
to be a collection of types only.
To maximize the confusion, here another argument: imap-next
does re-export imap-types
, but not imap-codec
.
I want to fight (for no good reason) for not having it in imap-types
. Let's discuss with a donut soon.