¿ªÔÆÌåÓý

ctrl + shift + ? for shortcuts
© 2025 Groups.io

Possible bug in "Send P Msgs to more than one BBS"


 

I was having a problem with mail routing the wrong direction and I think I've tracked it down to the "Send P Msgs to more than one BBS" feature for HA forwarding.

Steps to recreate:

Create a first BBS that sends HR personals to WW.
Create a second BBS with a name later alphabetically that sends HR personals to #77.gbr.euro
Activate Send P Msgs to more than one BBS

Send a message to MM0UHR@MM0UHR.#77.gbr.euro.

Expected behaviour is that the mail will go to the second BBS. In practice it goes to the first.

The problem seems to be in MailRouting.c lines 1476-1513. It searches through BBSs until it gets any match (which the first BBS will match with WW) and sends to every BBS that has a match at that specific match level (excluding the better match second BBS). Having completed the send loop it quits the search loop and the more accurate BBS is never found.

I think it should complete the search loop entirely and then start the send loop.

73

Chris


 

¿ªÔÆÌåÓý

Thanks, Chris.

A bit further down the code it does just that - comment is


?? ???? // We should choose the BBS with most matching elements (ie match on #23.GBR better that GBR)
?? ???? // If SendPtoMultiple is set I think we send to any with same match level

?? ???? // So if SendPtoMultiple is set I think I need to find the best depth then send to all with the same depth

But as you say it never gets that far. I think the code needs reordering. I'll look into that.

73,
John


On 28/06/2024 09:14, Chris - MM0UHR wrote:

I was having a problem with mail routing the wrong direction and I think I've tracked it down to the "Send P Msgs to more than one BBS" feature for HA forwarding.

Steps to recreate:

Create a first BBS that sends HR personals to WW.
Create a second BBS with a name later alphabetically that sends HR personals to #77.gbr.euro
Activate Send P Msgs to more than one BBS

Send a message to MM0UHR@MM0UHR.#77.gbr.euro.

Expected behaviour is that the mail will go to the second BBS. In practice it goes to the first.

The problem seems to be in MailRouting.c lines 1476-1513. It searches through BBSs until it gets any match (which the first BBS will match with WW) and sends to every BBS that has a match at that specific match level (excluding the better match second BBS). Having completed the send loop it quits the search loop and the more accurate BBS is never found.

I think it should complete the search loop entirely and then start the send loop.

73

Chris


 

¿ªÔÆÌåÓý

The latest code in my Beta area and git should fix this.

73,
John


On 28/06/2024 09:55, John G8BPQ wrote:

Thanks, Chris.

A bit further down the code it does just that - comment is


?? ???? // We should choose the BBS with most matching elements (ie match on #23.GBR better that GBR)
?? ???? // If SendPtoMultiple is set I think we send to any with same match level

?? ???? // So if SendPtoMultiple is set I think I need to find the best depth then send to all with the same depth

But as you say it never gets that far. I think the code needs reordering. I'll look into that.

73,
John


On 28/06/2024 09:14, Chris - MM0UHR wrote:
I was having a problem with mail routing the wrong direction and I think I've tracked it down to the "Send P Msgs to more than one BBS" feature for HA forwarding.

Steps to recreate:

Create a first BBS that sends HR personals to WW.
Create a second BBS with a name later alphabetically that sends HR personals to #77.gbr.euro
Activate Send P Msgs to more than one BBS

Send a message to MM0UHR@MM0UHR.#77.gbr.euro.

Expected behaviour is that the mail will go to the second BBS. In practice it goes to the first.

The problem seems to be in MailRouting.c lines 1476-1513. It searches through BBSs until it gets any match (which the first BBS will match with WW) and sends to every BBS that has a match at that specific match level (excluding the better match second BBS). Having completed the send loop it quits the search loop and the more accurate BBS is never found.

I think it should complete the search loop entirely and then start the send loop.

73

Chris