On Tue, Oct 30, 2007 at 10:31:41AM +0000, Johannes Schindelin wrote: > Hi, > > On Tue, 30 Oct 2007, Matti Aarnio wrote: > > > The git-send-email does send posts without any sort of MIME labeling: > > > > From: / To: removed > > > > Subject: [PATCH 0/2] Blackfin I2C/TWI driver updates > > Date: Tue, 30 Oct 2007 17:33:15 +0800 > > Message-Id: <1193736797-9005-1-git-send-email-bryan.wu@xxxxxxxxxx> > > X-Mailer: git-send-email 1.5.3.4 > > Precedence: bulk > > > > > > which per MIME rules means that the message in question is equivalent > > to one with header labels: > > > > MIME-Version: 1.0 > > Content-Type: text/plain; charset=US-ASCII > > Content-Transfer-Encoding: 7bit > > AFAICT MIME headers are only added when needed. (But that might only > apply to format-patch; however, if you signed off with your name, all > should be well.) I wish that were true.. We (VGER's Postmasters that is) would not see so much rejections from all over the places... Would it motivate you if we sent all GIT caused ones to you ? The MIME headers should always be added. The charset= -value can be debated about. If the headers are not there, and VGER is sending email to system which does not report capability of receiving 8BITMIME, then the message will be converted to QUOTED-PRINTABLE (and headers be added) Worst that can happen when the message is mis-labelled as US-ASCII and still contains 8-bit chars is not outright rejection, but stripping the 8th-bit off.. > > Now if the git-send-email would add following three lines in all > > outgoing email headers, things would be 99% correct for a long time.. > > > > MIME-Version: 1.0 > > Content-Type: text/plain; charset=ISO-8859-15 > > Content-Transfer-Encoding: 8BIT > > No. > > Not at all. > > ISO-8859-15 is just as wrong as ASCII. You just forget about the majority > of the population on this earth. That is just as arrogant and snobbish as > the people who thought that ASCII would be good enough for everyone. I wish the world was perfect and everybody would use single unified charset.. which obviously is not true. However as the kernel sources DO NOT SPECIFY COMMENT CHARSET, using ISO-8859-15 is LESS WRONG than picking UTF-8. The reasoning here is that getting the email thru is more important than absolutely correct labels. If you wish, you may scan the message content at first to determine if it does indeed include characters outside the US-ASCII range. You may even determine if the charset in question is likely UTF-8 instead of one of the Latins. Whatever you do, do put the MIME labels in place. > Ciao, > Dscho /Matti Aarnio -- one of <postmaster @ vger.kernel.org> - To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
