On Sat, 19 Jun 2004, Christian Balzer wrote:
> As one can gather from my .sig and whois, I'm hailing from Japan.
> The standard encoding for mails with Japanese characters is
> ISO-2022-JP which equals JIS. A typical example follows, the actual
> escape character is denoted by ESC. ;)
> ---
> ESC$BEl5~ETK-Eg6hCSB^ESC(B2-51-14ESC$BHtfF%S%kESC(B7ESC$B3,ESC(B
> ---
> The (faked) address of a spammer, btw. :)
>
> Exim 3 would throw a hissing fit when encountering ESC characters,
> 4.x seems to be more tolerable but it still doesn't work. The above
> string in a bounce message file results in:
> ---
> 2004-06-19 16:43:00 1BbaV9-0005ty-5T Failed to expand string from
> bounce_message_file or warn_message_file (intro): unknown variable name "BEl5"
> ---
> and the whole section reverts to the default text.
If you put a string like that into a bounce message file that is
expanded, you must escape the dollar signs. Or surround the whole
sequence with \N ... \N which disables expansion for what is in between.
> An obvious solution would be to ignore unknown $xyz sequences
On general principles, I do not like that. A simple typo then goes
unnoticed. Far too much of a gotcha.
> Along the same lines I would suggest adding a quota_warn_message_file
> option,
Noted.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
Get the Exim 4 book: http://www.uit.co.uk/exim-book
