[pmg-devel] [RFC pmg-api] trim Message-ID when parsing E-mail

Stoiko Ivanov s.ivanov at proxmox.com
Thu Apr 17 16:31:03 CEST 2025


Thanks for the patch!

generally looks ok - and could be merged - some nits/questions inline (can
fixup most upon applying):

On Mon, 14 Apr 2025 17:36:56 +0200
Hannes Duerr <h.duerr at proxmox.com> wrote:

> when we currently parse emails we do not remove trailing newlines from
> the message-id. The consequence of this is that if you use the rule
> system macro __MSGID__, there is also a newline at the end of the
> string. This in turn leads to problems if you create a rule and want to
> add something after the message ID.
As we talked off-list about this I know the context - but mentioning the
case of - modifying the message-id, after stripping <> and the domain to
not leak information  would help people in the future remembering why this
was added.

> 
> Signed-off-by: Hannes Duerr <h.duerr at proxmox.com>
> ---
> 
> Notes:
>     I'm not 100% sure whether we can simply trim the newline without any
>     problems. I have searched through the code a bit, but I have not come
>     across any problematic areas. We use the msgid for the log messages, but
I was surprised that the newline might be there at all - but the docs of
MIME::Head (https://metacpan.org/pod/MIME::Head) say so:
NOTE: The header(s) returned may end with a newline. If you don't want this, then chomp the return value.

Else I looked through our code and also think that this is read once, and
then used in a log-message and for adding it as template variable
https://git.proxmox.com/?p=pmg-api.git;a=blob;f=src/bin/pmg-smtp-filter;h=32bad7b881733147e5defe64a590dc709669dd1a;hb=HEAD#l184


>     it is in the last position there and a newline is set. This means that
>     the log message does not change
>     If anyone has any objections or other ideas on how to tackle the
>     problem, I'd be interested to hear them
How did you test this - I gave it a quick spin with a modify field object
(adding a X-my-msgid header) and it looked ok - but out of curiosity?

> 
>  src/PMG/MailQueue.pm | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/src/PMG/MailQueue.pm b/src/PMG/MailQueue.pm
> index 4e37cb9..2af95bd 100644
> --- a/src/PMG/MailQueue.pm
> +++ b/src/PMG/MailQueue.pm
> @@ -4,6 +4,7 @@ use strict;
>  use warnings;
>  
>  use PVE::SafeSyslog;
> +use PVE::Tools qw (trim);
>  use MIME::Parser;
>  use IO::File;
>  use Encode;
> @@ -394,7 +395,7 @@ sub parse_mail {
>      PMG::MIMEUtils::fixup_multipart($entity);
>  
>      if ((my $idcount = $entity->head->count ('Message-Id')) > 0) {
> -	$self->msgid ($entity->head->get ('Message-Id', $idcount - 1));
> +	$self->msgid (trim($entity->head->get ('Message-Id', $idcount - 1)));
nit: when touching this you could remove the space `msgid(...)` vs 
`msgid (...)`

>      }
>  
>      # fixme: add parse_time to statistic database





More information about the pmg-devel mailing list