The utf8-length patch appeared in version 1.03-22 systematically uses
UTF8.length to compute string lengths, which will produce wrong
results if the header uses another encoding, such as iso-latin-1. The
attached patch proposes to fall back to plain String.length if the
given string is
not a well formed utf-8 string.

This issue has already been reported on Ubuntu's bts
(https://bugs.launchpad.net/ubuntu/+source/headache/+bug/1083216) as
I'm using Xubuntu 14.10, but it seems that noone there is interested
in fixing it (issue was opened on 2012-11-26).

Le 2015-02-20 11:39, Virgile Prevosto a écrit :

Who uses iso-latin-1 anyway? :-P

I'll try to look if it is possible to include this for Jessie (I have
some doubts). Otherwise, it will wait for Jessie's release. (Which
means
if you get lucky, you'll have it in Ubuntu within 2 years! :D).
no one there doing OCaml stuff (and speaking on Ubuntu, globally). It
is not supported (except the compiler itself). So, of course, no one
will ever reply.

Regards,

2015-02-20 14:42 GMT+01:00 Mehdi Dogguy <mehdi@dogguy.org>:

Well, back at the time of the Ubuntu report, we still had some, but I
have to admit that I needed to forge one in order to check that the
patch still apply nowadays ;-)

Many thanks for the prompt reply in any case. I hope I won't me
confronted again to a latin-1 header in the mean time ;-)

I guess I still kind of hoped that the issue would be propagated to
Debian at some point, but I can understand that it wasn't important
enough to be handled by Ubuntu's bug managers

Best regards,

Sure. No problem. Do not hesitate to submit bugs directly to Debian
next time.  The only issue would be to reproduce the problem on a
Debian machine, which is generally easy with OCaml packages (well, as
long as there is not fishy binding stuff involved).

Kind regards,