On Thu, 2008-04-10 at 17:41 +0200, Jérôme Haguet wrote: > About the attached patch, I did not test it ... > Is it at least the kind of thing you are expecting ?
Well, this patch only addresses the Content-Type of the email that may
be sent out. How about the encoding of the messages that will be entered
in the bug-tracking system ?
> + %mail->{'Content-Type'} = 'text/plain; charset="utf-8"';
Isn't this an "assumption" we may be incorrectly making ? How do we know
the user entered a message in UTF-8 ?
To me a general patch would be one that accounts for both Windows and
Linux. For example, as Thorsten discovered in the past, using the "chcp"
command to discover the encoding in Windows. And perhaps discovering the
locale in Linux. And since, apparrently, Subversion throws away the
environment when committing, does that mean we need to discover the
locale (perhaps both in Linux and using chcp in Windows) at glue
installation time and set it as a configuration variable in the glue
(and pass it to the daemon) ? Or can we make absolutely sure that the
glue will read everything in UTF-8 so that setting UTF-8 as the
Content-type will no longer be "an assumption" ?
In any case, all these issues need to be considered when building a
general patch.
signature.asc
Description: This is a digitally signed message part
_______________________________________________ scmbug-users mailing list [email protected] http://lists.mkgnu.net/cgi-bin/mailman/listinfo/scmbug-users
