Opened 4 years ago
Closed 3 years ago
#1486119 closed Bugs (fixed)
Date: header should be in user's local timezone, not server's timezone
| Reported by: | ptm | Owned by: | |
|---|---|---|---|
| Priority: | 8 | Milestone: | 0.4-stable |
| Component: | Core functionality | Version: | 0.3-stable |
| Severity: | trivial | Keywords: | timezone |
| Cc: | roundcube@…, jolexa@… |
Description
When about to send an email, RoundCube inserts the Date: header. The timezone used is the server's local time.
The timezone should match the user's preferences instead.
For example, if my server is in the US, but my user is in Australia, the header _should_ read:
Date: Sat, 12 Sep 2009 20:00:00 +1000 <- Australian time
_not_:
Date: Sat, 12 Sep 2009 03:00:00 -0700 <- USA time
Change History (6)
comment:1 Changed 4 years ago by ptm
- Cc roundcube@… added
- Priority changed from 5 to 8
comment:2 follow-up: ↓ 3 Changed 4 years ago by alec
comment:3 in reply to: ↑ 2 Changed 4 years ago by ptm
Replying to alec:
Is this recommended/required by some RFC? I think it's better to rely on server's time settings than client's.
I am not aware of any RFCs that deal with this specifically, however I can offer some better precedents:
- A multi-user unix host. The server resides in the USA, and the time is kept accurate by NTP. Users in California, USA log on and see the correct time. A user in Sydney, Australia, logs on too, and also sees the correct time... it has been adjusted for local time by setting the TZ environment variable.
The mail user agent for both these users will stamp the _correct_ time on outgoing mail... but in each individual user's timezone.
- Other webmail clients exhibit the desired behavior. Example: SquirrelMail.
So, yes, rely on the server's _time_, but also the user's desired _timezone_. Besides, the default will be the server's timezone anyway...
comment:5 Changed 3 years ago by darkside
- Cc jolexa@… added
comment:6 Changed 3 years ago by alec
- Resolution set to fixed
- Status changed from new to closed
Fixed in [2bf3cc6f].

Is this recommended/required by some RFC? I think it's better to rely on server's time settings than client's.