search

fmbiete / Z-Push-contrib

Z-Push fork with changes that I will try to contrib
GNU Affero General Public License v3.0
134 stars 62 forks source link

Character encoding messed up #156

Open fortiko opened 9 years ago

fortiko commented 9 years ago

Just for the record, I sent myself a "share" email on iOS 8.2, and received the following preview in my Inbox (up-to-date git commit c7a73b39dea4cdf78c5ba94364851af6d3120ba0):

asian

the email only contains a URL, and when I just repeated the experiment, I do get a preview without asian letters, but with no content either:

no-preview

If I access the same Inbox via IMAP, I get the correct preview:

imap

2 issues:

1) Preview of this specific "share" messages in Z-Push is not working: is this a problem of the outgoing format that iOS sends, or a parsing problem on Z-Push's side? 2) Asian character jumble-up: why? I have kept the corresponding messages for now, with all headers intact. Does it make sense to send them somewhere for closer inspection? I do not see any significant difference in them, to be honest.

fortiko commented 9 years ago

Great, now it fixed itself for some reason: the 2 messages from a couple of minutes ago have a preview now, and the asian characters have also disappeared in favour of the preview:

correct

Even though it was only "temporary" (for about 14 hours), this could be indicating something more important in terms of character encoding. Imagine you cannot read or access an important email, because all you see is a mountain of asian characters…

Any ideas?

fmbiete commented 9 years ago

Hi @fortiko, if you saved your log server in WBXML level, compare the wrong and the ok GetMessage operation. Also look if the message was sent alone or with other messages to the phone.

fortiko commented 9 years ago

It is not about the message, I have now observed that the same messages flip-flops back and forth between "Asian" characters and Latin characters. Now I am trying to reproduce the issue and save the WBXML log to figure out what causes it… Heisenberg indeed!