fix: recognize self addresses in various places (instead of just the "primary") - #8533
fix: recognize self addresses in various places (instead of just the "primary")#8533hpk42 wants to merge 5 commits into
Conversation
2da385b to
11a09f7
Compare
| let self_name2 = stock_str::self_msg(context); | ||
|
|
||
| if self_addr.contains(query) | ||
| if self_addrs.iter().any(|a| a.contains(query)) |
There was a problem hiding this comment.
This change and the test in contacts_tests looks good.
d3b1853 to
8e8f019
Compare
8e8f019 to
55d30b1
Compare
| } else { | ||
| let mut new_members: BTreeSet<ContactId>; | ||
| // True if a Delta Chat client has explicitly and really added our primary address to an |
There was a problem hiding this comment.
this code block is untested, and writing a test seems involved (i tried).
I wonder if the original reason for this block has to do with pre-v2 versions?
You can also reach the block today i guess with someone sending from a stale member list (60 days no messages), so without member-timestamps, but it requires more conditions to cause this block to execute. And if we want to accept an explicit addition, we should still prefer later arriving member-timestamps and i am nto sure that's happening. It's all not too important i think, but if you have an easy answer, let me know. I just didn't bother with the test because the self-addr fix is small enough and straightfoward.
| let from = context.get_primary_self_addr().await?; | ||
| let lowercase_from = from.to_lowercase(); | ||
|
|
||
| recipients.retain(|x| x.to_lowercase() != lowercase_from); | ||
|
|
There was a problem hiding this comment.
This looks like the self recipient may be added twice now under some circumstances? I just noticed that we don't have any tests at all for recipients, will write one
Edit: Nevermind, we do have some tests for it
There was a problem hiding this comment.
What we did not have is a test for not adding the self address twice in a group, or for not adding it at all when bcc_self is off. I wrote a test, and directly pushed it since @hpk42 said he won't engage much with DC development this week
There was a problem hiding this comment.
The self recipient may now be added twice if one of the group members' transports uses the same address as the self (primary) self address. This is such a rare case (and an already-broken setup) that it seems fine. For all other cases, the if id != ContactId::SELF checks in mimefactory.rs are most likely enough (confirmed by the test I just added).
3 participants
No description provided.