bug#58472: [PATCH] Make `message-unique-id' less prone to collisions

[Stefan Kangas]

This is a proposal to make `message-unique-id' less prone to collisions.

For the first 1-2 characters, it uses the return value `user-uid', which
on most single-user systems is a fixed value (for example, this gives us
"87" on almost any single-user Debian machine).[1]  It's also
unnecessarily leaky of potentially privacy sensitive information.

The next 8 characters are the current time, with some gymnastics to
emulate support for a fractional part to seconds.  This seems
unnecessary now that, AFAIU, `time-convert' can do that for us portably
(please correct me if I'm wrong, Paul).

I suggest that we instead base the left-hand side of the Message-ID on:

  1. (time-convert nil (expt 10 9))
  2. 2^N bits of pseudo random data (e.g. N=32)

We can then ignore the effective user id, while significantly decreasing
the risk of a Message-ID collision.[2]

Currently, we get values like:

    (message-unique-id)
    => "87o7uhi3at.fsf"            ; length 10

With the attached patch, we have instead:

    (message-unique-id)
    => "cnk29wgg1a4nvrpqcy.fsf"   ; length ~22

Note also that `message-number-base36' uses a Caesar cipher for some
reason:

    (message-number-base36 5 -1)
    => "u"    ; expect "5"
    (message-number-base36 (expt 36 3) -1)
    => "yzzz" ; expect "1000"

The patch fixes this also.

I don't know if this change should be in NEWS or not.

Footnotes:
[1]  Just for fun, you can search for 87 on
     https://en.wikipedia.org/wiki/Message-ID

[2]  See also: https://www.jwz.org/doc/mid.html

0001-Make-message-unique-id-less-prone-to-collisions.patch
Description: Text Data