On Wednesday, 12 June 2013 20:38:48 CEST, Thomas Lübking wrote:
It's a matter of what information is stored locally.
Those clients will either talk to akonadi or open a local database
where they cache the mail headers, ie. proxy the imap server with ALL
headers present locally and all sorting is done locally as well.
Trojita relies on the server to do this and only presents the really
required data, leading to eg. no sort abilities for gmail users at
all.
This is a good explanation of why it's done this way at this moment. Yes, it
could be improved (perform a client-side filtering if the data is already
there, etc), but that's a lot of code which only gets triggered under very
specific circumstances -- i.e. a huge potential for breakage in unexpected or
unusual situations.
Only Oracle seems to support CONTEXT=SORT atm [1] (what is a shame twice!)
I find it weird that they do CONTEXT=SORT but not ESORT, but the CAPABILITY
listings seem to be the supportive evidence -- cool. Anybody got a testing
account? I asked for one last year, but never heard back from them.
This cannot be handled by either (domain)filters or a single setting.
Filters are to complex and much of a task and it's also completely
unpredictable what domain will contain the wanted filter or whether
todays url on that domain is nice and tomorrows is evil.
For similar reasons, whitelisting senders won't work - not only would
you have to setup such list, it's also completely worthless due to the
... errr .. weak credibility of that information.
The "domain" I meant was the domain where the images are hosted. In my opinion,
sender addresses are indeed extremely unreliable (and easily faked), but it's only the
trustwortihness of the target image URL which matters.
Presumably, the original poster's setup is that his their colleagues are
sending HTML mail with images referring to remote screenshots, or stuff like
that. Are these screenshots/images/whatever hosted on a single domain?
A single setting is a complete no-go - you'd trap the innocent fools
into downloading mail address verificators.
The question is always "do i trust this mail?" and if so, one could
push a button to download all external resources, but outside all
local environments this BrainFilter™ step is inevitable - it's avery
hostile world out there.
I ten to agree with this. Perhaps the settings could be made persistent for
each already-shown mail -- when the images were loaded once, a special flag
could be certainly set, and when Trojita sees the flag next time, the images
could be loaded automatically. Beucase the IMAP flags cannot be specified by
the sender, I don't see any security problem with this. It would, ohwever, not
elliminate the need to click that button at least once for each mail...
Cheers,
Jan
--
Trojitá, a fast Qt IMAP e-mail client -- http://trojita.flaska.net/
