open
https://gitlab.synchro.net/main/sbbs/-/work_items/1357
The RFC 2047 decoder in `exec/load/mimehdr.js` loses the white space
between an encoded word and plain text that follows it.
**Reproduced** (current master):
```
mimehdr.to_cp437("Re: =?iso-8859-1?q?caf=E9?= is open")
-> "Re: caf<e-acute>is open"
mimehdr.to_cp437("=?utf-8?q?caf=C3=A9?= menu.pdf")
-> "caf<e-acute>menu.pdf"
```
**Cause.** In `decode()` the regular expression is
```
/(\=\?[a-zA-Z0-9-]+\?.\?[^ ?]*\?\=|\S+\s*)/g
```
The plain-word alternative (`\S+\s*`) keeps its trailing white space,
but the encoded-word alternative does not, and the next `exec()` then
skips that white space. RFC 2047 says white space is dropped only
BETWEEN two adjacent encoded words; white space between an encoded
word and ordinary text is significant.
**Suggested fix:** keep the white space after an encoded word unless
the next token is another encoded word.
-- *Authored by Claude (Claude Code), on behalf of @rswindell*
--- SBBSecho 3.38-Linux
* Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)