Reasons for "Warning: While decoding this file Opera encountered errors."

When there's an error decoding the body of a message, Opera places "Warning: While decoding this file Opera encountered errors." in the body.

Since Opera isn't specific about what the error is, we'll build a list of known things that you can look for in the source of the message (right-click on message and choose "view all headers and message") to try and narrow it down.

The list is going to be short at first because when people file bug reports on decoding errors, they almost never provide an mbs file for the message, which means most of the errors can't be analyzed.

• You have a header line like this:

  Content-Type: text/plain; charset="utf-8

  where the quoted charset value isn't terminated properly (missing quote)

  Watch for missing end quotes in other headers too, especially the filename parameter in Content-Disposition, which can cause attachments to be missing.

• You have a body transfer encoding of:

  Content-Transfer-Encoding: quoted-printable

  where a line in the body source is longer than 76 characters (violating #5 in <http://tools.ietf.org/html/rfc2045#section-6.7>).

  You'll see this one with spam messages a lot, and xstandard.com notification emails (possibly sent by an asp page with the IIS SMTP service) for example (Edit: They recently fixed it though).

• You have a body transfer encoding of:

  Content-Transfer-Encoding: quoted-printable

  and the last character in the body data is a space. (Not sure if this is an Opera bug that it gives a warning or if it's a legit spec violiation.)

• You have 2 parts (attachments) in the message source that both have the same exact Content-ID header value.

  You'll mostly see this one from cell phone text messages sent to your email.

Besides above, here are some things that might cause a decoding error to be in a message:

• Anti-Virus email scanner.

• Low-Bandwidth mode when message body isn't fetch or was fetched incompletely in some way.

• Importing mbox file from another client where the other client does something goofy with how it generates mbs files.

Hi burntout426,

Eudora (bad people)!

I tried to import all my mail from Eudora after switching to Opera, about half of it failed with "Warning: While decoding this file Opera encountered errors."

Did a Google search without much luck... no-one seems to have solved it. So looked deeper, opened the Eudora *.mbx files as text documents and looked at the headers, finding inconsistencies. For the geeks here's a sample of the header and start of message:

...
MIME-Version: 1.0
Content-Type: multipart/alternative;
boundary="481282164-1729405018-1339125660=:66113"
X-TPG-Junk-Result: determined as NOT junk email, but would have with a high detection setting (BCC)

<x-html>
<table cellspacing="0" cellpadding="0" border="0" ><tr><td valign="top" style="font: inherit;"><DIV>Dear Rob,</DIV>
<DIV> </DIV>
<DIV>Thanks so much for your email which I really appreciated to read.</DIV>
...

Seems Eudora breaks the rules when it stores multipart messages in the mailbox - the header says it's in two or more parts e.g. text/rich-text/html separated by a random-number-string 481282164-1729405018-1339125660=:66113. But Eudora stores only the HTML part, not the other plain-text or rich-text parts. Also removes all the boundaries, the random-number strings. Perhaps to save space? And without changing the header - Eudora must be programmed to read it, but it's non-standard so Opera can't.


Solution:

In the text editor (notepad) I did a "find and replace", swapped out every instance of "multipart/alternative;" with "text/html; charset=utf-8".

That made the headers say all the messages were HTML (which they were, plain-text had been stripped off). It also left in the "boundary=random number" line, but that didn't seem to matter to Opera, luckily. When I IMPORTED all the mail it shows up in HTML format without any errors.

Hope this helps,

Cheers,
Rob