I'm running Slackware64 14.2 and using mail (mailx) as an email client from the command line. Occasionally, I get a message body of [Binary content], which generally means there is some non-ASCII character in the message somewhere. I can read the message find in a text editor with "offending" characters rendered with their correct accent-marks, etc.

Is there something I can set for mailx to go ahead and show these characters? Otherwise, it's a bit of a pain to have to save the message to a file and read it with an editor later.

mfoley --

The dos2unix command may do what you need ( produce pure-d 7-bit ASCII for mailx ).

The dos2unix and unix2dos commands are available on SBo: misc/dos2unix

Check `man dos2unix` for options once it is installed from SBo.

HTH

-- kjh

The problem isn't that I can read the message at all. If I save the message to disk, then use an editor to open it, I can see the message just fine, including special characters. But, I want to be able to see the message body from within mailx when I'm reading messages interactively.

One of the problem characters causing the [Binary content] response is the copyright character: ©. Which is hex A9.

I've tried setting 'print-all-chars' in my .mailrc file. I've tried 'set ttycharset=utf8'. Neither of these seem to work.

Maybe that copyright/A9 character is not utf-8. Perhaps a different charset?

mfoley --

Oops.

Yes, I see ...

I thought you needed to 'sanitize' a text file for mailx send mode which I do like this when messing with pesky Windows files:

Not sure about mailx read mode ... maybe mutt can handle your high-ASCII chars ?

Sorry.

-- kjh

p.s. Maybe if you export a different value for PAGER, it would work for you ?

man mailx # search for PAGER

The less command displays high-ASCII as Hex-Codes for me using my ancient LESS variable:

OK, I've figured out a solution from the sender side. Some of the messages have control characters; codeset isn't the issue. The messages have BEL characters, carriage-returns, ESCape, etc. Before piping one type of message (sbopkg update output) into mail I do:
This sbopkg output has ESC sequences and CRs. I first translate the newlines to "@@" to preserve them, then deleted all control characters and translate the "@@" back to newlines.

Another one is:
This is to remove CRs, the copyright symbol (although that might actually be OK) and the BEL character. This is to cleanup the output of a todo script.

In the end, my problem was not the codeset but control characters that mailx chokes on.

kjhambrick, I tried setting PAGER to less, but it didn't work for me, even though when I use less on the actual text I'm going to email it does show the control characters as e.g. ^G, highlighted. Nevertheless, I still get the "[binary content]" message in mail. This fellow was able to get less working: https://unix.stackexchange.com/quest...-in-cron-mailx, but I was not. Perhaps I'm not specifying it correctly in .mailrc:

I was also having issues with the output with mailx attaching a .BIN file when I wanted the output to be in the body of the email.
I have been using "tr -d \r" to fix this issue, which worked on basically all scripts I was running, until one day even with this in place I was still getting a .BIN file.
So I simply ran the script and piped this to 'cat -vet', so script|cat -vet and this showed me what characters were in the mix. The \r in the tr command sorts out any ^M characters, but this time is was getting some ^G characters, which looking up the ascii character set (https://docstore.mik.ua/orelly/unix3...ut/appa_01.htm) the octal for this is 007. So to fix the output to my script to work with mailx I simply use: script | tr -d "\007\r" |mailx -s "Test" somebody@nowhere.com

So 'cat -vet' is your friend to assist you with what characters are in place for you to apply the necessary octal value to delete.