incorrect display of headers related to folding

[vinc17]

Under some conditions (when a long header line ends with a space?), Mutt regards a header line as being part of the previous one, e.g. instead of displaying:

Subject: 1234567890 1234567890 1234567890 1234567890 1234567890 1234567890 test
To: a@b.invalid

it displays:

Subject: 1234567890 1234567890 1234567890 1234567890 1234567890 1234567890 test
        To: a@b.invalid

Note: this is only a display problem.

[vinc17]

testcase: mailbox file

[pdmef]

I cannot reproduce this with -nF /dev/null. Is there a display_filter involved?

[vinc17]

Replying to pdmef:

I cannot reproduce this with -nF /dev/null. Is there a display_filter involved?

No display_filter involved. And the bug still occurs with:

./mutt -nF /dev/null -f folding-bug

under Debian and unpatched Mutt's tip.

[vinc17]

And in case this could matter, the bug still occurs with LC_ALL=C TERM=vt100. I forgot to say, the terminal must have 80 columns.

[pdmef]

I can now reproduce it with 80 cols.

The problem is that mutt_write_one_header() wraps the header at the trailing space and inserts a tab for folding (see #2995). To me it seems that it fails to notice that there's no further content and thus doesn't append \n to finish this header.

I'm marking milestone as 1.6 so that once we touch that function for #2995 we should also fix this issue.