On 03/15/2023 06:57 AM, Oswald Buddenhagen wrote:
> On Tue, Mar 14, 2023 at 08:37:27PM +0100, H wrote:
>> I have never used mutt - do I install it on one of my workstations and then 
>> access theĀ  mail backup directory on the server over whatever suitable 
>> connection?
>>
> via imap. possible, but i wouldn't do that.
>
>> Or, do I install it on my server and run it there after ssh-ing into the 
>> server?
>>
> also possible, and if the server's mail store is directly supported by mutt, 
> you could access it that way - at the risk of messing it up. an imap 
> connection to localhost would be safer.
>
> but the way i'd recommend is to install it on the machine with the primary 
> backup and use it to access the isync'd maildir directly.
> isync was originally written by the same guy as mutt, to be used together. as 
> i do to this day.
>
>> When you say ",U=xxx" infixes does this include the entire ",Uxxx:y,S/RS" 
>> part of the filename?
>>
> no. in perl terms, /,U=\d+/
>
>> If I understand you correctly I need to remove this infix part from each 
>> file in the older backup directories /before/ folding them into the 
>> "up-to-date clone"? This means that the message files in the "up-to-date 
>> clone" will still have the infix part in the filename whereas the message 
>> files pushed from older mail backups into that directory will not.
>>
> correct
>
>> Of what importance is this infix part of the filename?
>>
> these are unique message ids added by isync. they are mailbox-specific, so 
> you need to remove them from moved in "foreign" messages.
>
>
>
> _______________________________________________
> isync-devel mailing list
> isync-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/isync-devel

Thank you for your explanations. Being not familiar with Maildir format etc., I 
did some digging and now believe I have some additional understanding of the 
",U=xxx" part, the identifier chosen by isync to make sure that the message has 
a unique identifier.

The beginning of the filename contains the date and time it was downloaded 
together with the hostname of the computer to which it was downloaded followed 
by an underscore and another identifier (possibly Q, the delivery number? as 
described in https://en.wikipedia.org/wiki/Maildir). The last part ":2,S/RS" is 
the status of the message and obviously needs to be preserved. The filename 
itself though does not seem to have any particular meaning besides needing to 
being unique to the message.

I did download and install mutt to the server where the backup(s) reside and 
quickly realized the learning curve is somewhat steep... I have not yet 
connected to localhost for archive access though. The same Wikipedia page as 
above also lists names of some other MUAs that I may look into. If anyone has 
additional suggestions I would gratefully accept those - I have been using 
Thunderbird to manage my several email accounts.

I also realized that I, in addition, have several zipped Thunderbird mail 
archives that I also ought to "dump" into this to-be consolidated mail archive. 
It is my understanding that Thunderbird (I have been using version 59) uses the 
mboxrd format. I think I have seen an import/export utility that I hope might 
be able to export into Maildir format which I can then "dump" into my archive 
for subsequent de-duplication.

I have set aside time two weeks from now to do this.

The other message from Evgeniy suggests that some messages might not be 
de-duplicated due to having different identifiers in the email header. However, 
you wrote that the 'D~=' sequence in mutt would de-duplicate based on 
message-ID which you said "should be good enough". Could you shed some further 
light on this last statement so I know what the limitations are?

Thank you.




_______________________________________________
isync-devel mailing list
isync-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/isync-devel

Reply via email to