With latest unstable, the unit tests have a failure with a single seed:

$ dist/build/unit/unit -t 'Checking that there are no empty lines in email
headers' --test-seed=1910310936
There are a total of 34 primitive patches.
There are a total of 384 patches.
"aaaaaaaaaaaaaaaaaaaaa"
"aa\ACKa\132aaa\DC4a\200aaaaaaaaaaa\153a "
Checking that there are no empty lines in email headers: [Failed]
Falsifiable with seed 1910310936, after 65 tests. Reason: Falsifiable

(build them by passing -ftest to cabal configure/cabal install).

Reinier: would you be a good person to investigate this, or are things still 
tricky   on your end?  Thanks :-)

On Tue, 22 Dec 2009, Eric Kow wrote:

>
> Eric Kow <kowey@darcs.net> added the comment:
>
> Reinier: would you be a good person to investigate this, or are things still
> tricky   on your end?  Thanks :-)

FWIW, I don't think it's particularly urgent, I just thought I should 
record it. I suspect it's been around for ages, just not very likely to be 
hit by random testing.

Ganesh

OK, I'll bump this to the next release on the grounds that we have enough work
on our hands.

The following patch updated issue issue1716 with status=resolved;resolvedin=2.6.0 HEAD

* resolve issue1716: allow mail header lines of all whitespace in test 
Ignore-this: 24a12e1cf2631b5363636cf32cf5e8da
  
Reading RFC822, it seems all-whitespace lines are allowed in the header section
when you've got trailing space that doesn't fit on the line before.

The following patch updated issue issue1716 with status=resolved;resolvedin=2.5.0 CURRENT

* resolve issue1716: allow mail header lines of all whitespace in test 
Ignore-this: 24a12e1cf2631b5363636cf32cf5e8da
  
Reading RFC822, it seems all-whitespace lines are allowed in the header section
when you've got trailing space that doesn't fit on the line before.