mshow: unknown charset of extracted attachments

[ashiire]

When extracting attachments via mshow's -x or -O flags, any present charset information is lost. This is troublesome, as charset information is difficult to infer from file contents alone.

mshow does explicitly mention attachment charset in render mode if a filter is used to render the attachment, like so:
--- --- --- 3: text/plain size=235 charset="iso-8859-2" render="mshow-plaintext" ---
However, this seems like the wrong (and inconvenient) place to recover the information from.

I have two ideas that might help:

• (Add an option to) explicitly state charset information in list mode, if available.
• Add an option to automatically re-encode extracted attachments to UTF-8, same as in render mode.

I think that either one would be sufficient on its own, but both may be desirable.

[leahneukirchen]

I guess the second part wouldn't be too hard to add, as all the steps are implemented already.