Re: Working with public-inbox.org [Was: [PATCH] rev-parse: respect core.hooksPath in --git-path]

3 messages, 3 authors, 2017-02-14 · open the first message on its own page

Re: Working with public-inbox.org [Was: [PATCH] rev-parse: respect core.hooksPath in --git-path]

From: Junio C Hamano <hidden>
Date: 2017-02-14 04:41:59

Arif Khokar [off-list ref] writes:
One concern I have regarding this idea is whether or not SMTP servers
typically replace a Message-Id header set by the client.
The clients are supposed to give Message-IDs, but because some
clients fail to do so, SMTP server implementations are allowed to
add an ID to avoid leaving a message nameless (IIRC, 6.3 in
RFC2821).  So "replace" would be in violation.

But some parts of the world ignore RFCs, so...

Re: Working with public-inbox.org [Was: [PATCH] rev-parse: respect core.hooksPath in --git-path]

From: Arif Khokar <hidden>
Date: 2017-02-14 05:09:57

On 02/13/2017 11:41 PM, Junio C Hamano wrote:
Arif Khokar [off-list ref] writes:
quoted
One concern I have regarding this idea is whether or not SMTP servers
typically replace a Message-Id header set by the client.
The clients are supposed to give Message-IDs, but because some
clients fail to do so, SMTP server implementations are allowed to
add an ID to avoid leaving a message nameless (IIRC, 6.3 in
RFC2821).  So "replace" would be in violation.

But some parts of the world ignore RFCs, so...
Based on my testing, gmail and comcast (and my work email) will preserve the Message-Id header set by the client, but hotmail.com/live.com/outlook.com will replace it with their generated value.

Based on a small sample of email addresses of those who post to this list, it appears that most people are using their own MTA to send email, or are using gmail, so they probably wouldn't be affected by the issue.

Re: Working with public-inbox.org [Was: [PATCH] rev-parse: respect core.hooksPath in --git-path]

From: Jeff King <hidden>
Date: 2017-02-14 05:14:10

On Mon, Feb 13, 2017 at 08:41:51PM -0800, Junio C Hamano wrote:
Arif Khokar [off-list ref] writes:
quoted
One concern I have regarding this idea is whether or not SMTP servers
typically replace a Message-Id header set by the client.
The clients are supposed to give Message-IDs, but because some
clients fail to do so, SMTP server implementations are allowed to
add an ID to avoid leaving a message nameless (IIRC, 6.3 in
RFC2821).  So "replace" would be in violation.

But some parts of the world ignore RFCs, so...
I know there are some terrible servers out there, but I think we can
discount any such server as horribly broken. Rewriting message-ids would
cause threading problems any time the sender referred to their own
messages. So "format-patch --thread" would fail to work, and even
replying to your own message from your "sent" folder would fail.

-Peff
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help